Sigbovik 2011#
A Record of the Proceedings of SIGBOVIK 2011
April 1st, 2011 Carnegie Mellon University Pittsburgh, PA 15213
{width="1.3194444444444444in"
height="1.3055555555555556in"}
- Association for Computational Heresy
:
Advancing computing as Tomfoolery & Distraction
SIGBOVIK
A Record of the Proceedings of SIGBOVIK 2011
ISSN 2155-0166
April 1, 2011
Copyright is maintained by the individual authors, though obviously
this all gets posted to the internet and stuff, because it's 2011.Permission to make digital or hard copies of portions of this work for
personal use is granted; permission to make digital or hard copies of
portions of this work for classroom use is also granted, but seems
ill-advised. Abstracting with credit is permitted, abstracting with
credit cards seems difficult.Additional copies of this work may be ordered from Lulu, see
http://sigbovik.org/ for details.
ii
A. massage... damn you autocorrect... message from the organizing
> committee:
The Association for Computation Heresy Special Interest Group (ACH
SIGBOVIK) on Harry Q. Bovik is thoroughly thrilled to present this,
the First Annual Fifth Anniversary Celebratory Intercalary Workshop
about Symposium on Robot Dance Party of Conference in Celebration of
Harry Q. Bovik\'s (26)th birthday, or the Fifth Annual SIGBOVIK for
shorts, in case of warmer weather. SIGBOVIK continues to evolve,
bringing promises both new and old in, respectively, the new exciting
directions our research has taken and the tried-and-true classic
topics.While this year\'s SIGBOVIK has fewer submissions than previous years,
the innovation and quality of research has skyrocketed: we are proud
to be publishing several papers which showcase long involved processes
of collaborative and experimentative research, other papers which
exemplify the value of feedback and criticism in academic pursuits,
and even up to one or more paper(s) written entirely by an Artificial
Stupidity Engine (itself known to be one of Harry Bovik\'s favoured
fields).The tracks this year have been hand-selected by our most talented
track-layers, with only minimal injury to body or spirit, to most
effectively contribute to the study of Conference Theory. We are
pleased to be able this year to publish a special track, Future Work,
as an indicator of some of the many future research directions of this
esteemed conference. We hope you will find it, as well as the more
regularly scheduled tracks, deeply enlightening.Also of note is that this year, the conference organizers have decided
to take a stand on the recently growing issues of furniture
discrimination, and have each selected a special piece of furniture to
honour by taking on as their name, rather than relying on the
prejudiced default of \"Chair\". These efforts are reflected below in
the list of organizers.With thanks for your continued participation,
The SIGBOVIK 2011 Organizing Committee:
That Recliner in the 412 Lab Motivational Posting
Ottoman Emperor Mayor of Publi City
Footstool Unsigned Long Internet
End Table Treasure Keeper
Knick-NAK Giver of Guidants
Couchy Sequins Morel Support
Floor Lamprey Glassblower
Beanblag Nutritional Expert
iii
Table of Contents
Track 1: Party Time
• Crowd-Sourced Party Planning . . . . . . . . . . . . . . . . . . . .
. . . . . . . . 3Nels E. Beckman
- The Holiday Coverage Problem: the ultimate approach to
deadline avoidance . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . 7 Tempus Fidget and Ukidding NowheyTrack 2: Novel Algorithms
• An Objection to "The Box and Circles Plot" . . . . . . . . . . . . .
. . . . . 17 Prudy Coldfish
- Feasibility Analysis of theorem prover, extracted using Coq's
code at the case, the triangle equalities . . . . . . . . . . . . . .
. . . . . . . . 21 Sourbot Tobruos• What words ought to exist? . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . 27 Dr. Tom Murphy VII Ph.D.• There's a Hole in the Bucket: DNA Memory Leaks . . . . . . . . . . .
. 43 G. Biolo, Ph.D.
iv
Track 3: Sleazeball-Computer Interaction
• An Objection to "An Objection to "The Box and Circles Plot"" . . .
47 Long T. Jongwood• Good Friends Project: A Little Social Experiment in Stalking . . . .
49 Brian C. Becker, Pyry Matikainen, Heather Jones,Prasanna Velgapudi, Michael Furlong, Brina Goyette, Heather Justice,
Umashankar Nagarajan, and Joydeep Biswas• Who is the biggest douche in Skymall? . . . . . . . . . . . . . . .
. . . . . . . 53 Dr. Tom Murphy VII Ph.D.
- A Materials Based Approach to Hardware Digital Rights
Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . 59 Matt Mets, Matt Griffin, and Marty
McGuire
- The "Who's Yo Daddy?" Problem, Or: Solvin' Concurren'
Axcess Issues wit Computers, Yo . . . . . . . . . . . . . . . . . . .
. . . . . . . 65 Dr. Donna M. Malayeri, PhD,Heather Miller, PhD©, Esq., PPPoE, P2P, and Herr Doctor Doktor
Klaus HallerTrack 4: Military-Industrial Complexity Theory
- An Objection to "An Objection to "An Objection to "The Box and
Circles Plot""" . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . 71 Gen. Reginald F. Stout, (Mrs.)• Going to the Mall Is Like, Totally NP-Complete . . . . . . . . . . .
. . . . 73 Brittany, Bethany, and Tiffany• Optimal Image Compression . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . 75 James McCann and Notreally A. Coauthor• On unlexable programming languages . . . . . . . . . . . . . . . . .
. . . . . . 79 Robert J. Simmons
v
Track 5: Developers, Developers, Developers, Developers
- An Objection To "An Objection to "An Objection to "An Objection to
"The Box and Circles Plot"""" . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . 85• MLA-Style Programming . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . 87 Richard Ha• Theoremin: Proof by Handwaving . . . . . . . . . . . . . . . . . . .
. . . . . . . 93 Ben Blum
- Quantitative Comparison of Typesetting Software: TeXShop
vs. TechShop . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . 95 Matthew Sarnoff• Med School, CS Grad School, Both, or Neither? . . . . . . . . . . .
. . . . 99 Brian HirshmanTrack 6: Future Work
• Classico Logic . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . 105• Hygienic Image Macros . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . .107• Kanye\'s Lost Cosmological Theorem . . . . . . . . . . . . . . . . .
. . . . . . 113• Headdesk Design Patterns . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . 151•
TBD.................................................157
Taus Brock-Nannestad and Gian Perrone• Theorems for A Pound of Flesh . . . . . . . . . . . . . . . . . . .
. . . . . . . . 161• Ringard\'s Paradox . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . 163• Parametric Polyamory . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . 165• Highest-order Logic . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . 173• Libeled Deduction . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . 175• Antialiasing Pointers . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . 181• The Lambda Timecube . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . 183• GPS Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . 195
vi
SIGBOVIK 2011: TRACK ONE
Party Time
- Crowd-Sourced Party Planning . . . . . . 3
Nels E. Beckman
Keywords: party, par-tay, shindig, kegger, hootenanny, mixer, rave,
box social, rager, ice-cream social, sockhop, gathering, festival
- The Holiday Coverage Problem: the ultimate approach to
> deadline
avoidance . . . . . . . . . . . . . . . . . . . . . . . . 7
Tempus Fidget, Ukidding Nowhey
Keywords: calendrics, diversity, religion, holiday, deadline,
coverage analysis
1
2
Crowd-Sourced Party Planning
Nels E. Beckman
Google Inc.
Abstract
This paper presents, howmuchbeer.com, a crowd-sourced approach to
planning your next rager.
Categories and Subject Descriptors K.4.2 [Computing Milieux]:
Social Issues
General Terms algorithms, economics, experimentation, human factors,
measurement
Keywords party, par-tay, shindig, kegger, hootenanny, mixer, rave, box
social, rager, ice-cream social, sockhop, gathering, festi-val
- Introduction
For centuries man has struggled with party planning. Planning a
successful party is just not an easy task. There are enumerable things
to do: one must buy food, buy drinks, make a playlist, invite friends to
the party, prevent enemies from hearing about the party, clean the
house, hang up cool posters, make finger snacks, rearrange the
furniture, buy toilet paper and give your pets hair cuts. There are also
quite a number of variables. Of the people you invited, how many will
actually show up? Of the people you didn't invite, how many will
actually show up? How much food will each person eat? Will it be a crazy
party, or a chill gathering? Will anyone throw up? Will anyone else be
having a party on the same night? And of course, how much beer will each
person drink?
In this paper, we present a bold new solution to that last ques-tion. We
will show how, by crowd-sourcing your party planning, you can slightly
decrease the effort it takes to throw a party. This crowd-sourcing
engine is known as "How Much Beer?.com."
- Approach
Our web site1 is central repository for determining how much beer to
buy for a particular party. The site has been designed with two use
cases in mind.
2.1 Use Case 1: Before the Party
Before a given party, the party planners would like to know how much
beer they should buy. When entering the web site, they are greeted with
a phrase for which they must fill in the blanks:
- howmuchbeer.com
Permission to make digital or hard copies of all or part of this work
for personal or classroom use is granted without fee provided that
copies are not made or distributed for profit or commercial advantage
and that copies bear this notice and the full citation on the first
page. To copy otherwise, to republish, to post on servers or to
redistribute to lists, requires prior specific permission and/or a fee.
SIGBOVIK '11 April 1, Pittsburgh, PA, USA-A-OK Copyright c 2011 ACH
Copyrights are totally radical.. . . $10.00
I'm going to have BLANK people, and I think it's going to be a BLANK
party.
In the first blank, users enter the number of guest they expect to
attend. In the second blank, user choose how crazy their party will be
based. There are three choices: "Chill," for relaxed parties featuring
high-brow conversation and organic snacks, "Normal," for regular parties
with little to no chance of skinny dipping, and "Wild," for parties with
two or more visits from the police. After making these choices, the
results are presented. The results tell the user how much beer they
should buy, in terms of cases (24 12-ounce bottles), and six-packs (6
12-ounce) bottles. In the future we plan to allow users to customize the
resulting list in case they would like to buy their beer in, for
example, kegs or 40-ounce bottles.
The results are computed by taking all records of previous parties
(described below), calculating an average (mean) ounces per attendee for
that particular party craziness level, and multiplying this average by
the number of attendees expected. Outliers are removed. When the results
are presented, in addition to the mean, users are presented with results
for one standard deviation above and below average. This will allow
party planners who want to be on the safe side, and cheap bastards
(respectively) to buy an appropriate amount of beer.
Finally, an associated Android mobile phone application2 al-lows party
planners who have already gotten to the beer store the ability to
determine their beer needs.
2.2 Use Case 2: After the Party
After party-goers have thrown their party, we encourage them to return
to our site and to share how much their party-goers actually drank. With
this information, the quality of our site will gradually increase over
time. When users fill in information about their party, the results are
a little different. Again, they fill out a sentence with blanks:
It was a BLANK party. We had BLANK people show up, and we drank BLANK
BLANK of beer!
The first blank again has users choose from one of three crazi-ness
levels, "Chill," "Normal," and "Wild." The second blank is for the
number of attendees and the third blank is for the quantity of beer
drunk. The fourth blank is for type of beer containers used. Users can
enter their beer consumed in ounces, gallons, bottles, kegs, and even
"power hours."
When this information is submitted, it is stored in our database for
later retrieval. Spam detection is performed in order to reject bad
data. (For example, beer industry executives who would like Americans to
drink more beer.) When submitting their information, users have the
option of logging in. Because of the accountability of a logged in user,
their entries are rated more highly. There is even an adminitrator mode
where authoritative articles on the subject of
3
beer consumption can be added to the set of records. These entries
are even more highly rated.
As of this date, data from 15 parties has been gathered, although
sadly, 11 of those were my own. Why doesn't anyone one to come
to my parties?
- Discussion and Future Work
While our site has already achieved our goals, fast becoming the
home for beer-centric party data, there are some unresolved issues,
and many potential avenues for future research.
First, it should be noted that not all beers are created equal!
One mustn't assume that a lowly session beer is equivalent to,
say an Arrogant Bastard or a Belgian Tripel. In other words, the
percentage of alcohol by volume is quite important. In the future,
we plan to collect this data and recalculate our results based on
actual value of alchohol. In our initial phases, we wanted to keep
the interface as simple as possible.
Along the same lines, readers may be wondering about other
types of alcoholic drinks. Don't people throwing a party tend to
have other types of drinks, say beer and wine, and won't the quan-
tity of those drinks surely affect how much beer people consume?
Well perhaps, but that's kind of the beauty of averages. If enough
people have parties with other kinds of alcohol, but faithfully record
the amount of beer consumed per person, we will still maintain an
accurate picture.
In the future we do, however, plan to support better curve-fitting
than a simple mean. Specifically, if a users selects a point in the
beer data space that is close to an existing point, or in between two
existing points, it might make more sense to extrapolate linearly
from those existing points, rather than taking an average over all of
the points. This approaches will be explored in good time.
The most important thing we want to get across is that beer is
good and should be drunk in an evidence-based manner,
- Related Work
I'm not going to pretend I invented crowd-sourcing or anything [1].
Acknowledgments
The author was funded by several grants including, NSF Grant
A5F88, entitled "Investigating the Potency of Eisbier," a grant#
from the Great Lakes Fund and student stipend from the Friends
of Dogfish.
References
- M. Vukovic, S. Kumara, and O. Greenshpan. Ubiquitous crowd-sourcing.
In Proceedings of the 12th ACM international confer-ence adjunct
papers on Ubiquitous computing, Ubicomp '10, pages 523--526, New
York, NY, USA, 2010. ACM. ISBN 978-1-4503-0283-8. doi:
http://doi.acm.org/10.1145/1864431.1864504. URL
http://doi.acm.org/10.1145/1864431.1864504.
4
sigbovik
{width="1.1368055555555556in"
height="1.1368055555555556in"}
ach 2011
Plenary Program Committee Con fi dential Paper Reviews
Paper 10: Crowd-Sourced Party Planning
PC MEMBER: Chris Martens
OVERALL RATING: 1 (weak accept)
REVIEWER'S CONFIDENCE: 1 (low)
The author has invented an impressive piece of social technology that
infers the appropriate quantity of beer to obtain for a party, shindig,
hootenany, rave, or other festive event. The re - viewer, excited,
immediately wanted put the beer robot to use. The immediacy of the
desire for gratification was great, so she had no time to invite other
people and prompty fi lled in the fi rst blank with "1" (people
attending the party); and of course it was going to be a WILD party. The
beer robot told her to buy two six packs.
The next morning, the reviewer's confidence in the unquestioned wisdom
of the beer robot was shaken. The Facebook and Twitter posts from her
account looked, at best, hazily familiar.
We therefore must recommend that howmuchbeer.com be used with the buddy
system. Perhaps the validity of the author's results could be checked by
way of a formalization in Drunken Logic (see SIGBOVIK 2007).
Also, the reviewer's confidence is "low" because omg she was not invited
to all eleven of the author's parties! :( :( :(
But whatevs it's still pretty awesome
5
6
The Holiday Coverage Problem
the ultimate approach to deadline avoidance
Tempus Fidget
Ukidding Nowhey
In its ongoing quest to accommodate actual, potential, and imagined
diversity of all kinds, a Certain Major University annually distributes
a list of religious and ethnic holidays with a request to minimize
scheduling of required assignments during these holidays. We explore the
implications of this request and, in particular, the question of whether
there is a degree of diversity accommodation that achieves the ultimate
in deadline avoidance, a calendar in which all due dates are excluded.
1. INTRODUCTION
Universities are diverse, multi-ethnic, multi-cultural, multi-faith
communities that strive to empower all students to achieve their full
potential within the fabric of their personal belief systems. To that
end, a Certain Major University strives to avoid interference between
academic activities and the obligations and holidays of these
multi-ethnic, multi-cultural, and multi-faith communities.
Annually a Certain Major University sends an 8-page memorandum to the
faculty that begins,
"As you are aware, a Certain Major University is a very diverse
community in which many different religions and ethnic groups are
represented. . . . I urge each of you to make an effort to minimize
the scheduling of required assignments and/or events during these
religious observances. Faculty members should invite students to bring
scheduling conflicts to their attention so that a reasonable
accommodation may be made. University policy dictates that when a
student has a conflict between a religious holiday and a graded
assignment, he or she should contact the faculty member directly in an
effort to work out a resolution. . . . I appreciate your assistance in
this important matter." [1]
This is followed by a 6-7 page list of holidays for the current year,
including Christian, Jewish, Hindu, Islam,
Gimme that old time religion
Gimme that old time religion
Gimme that old time religion
It's good enough for me [2]
The obvious question a reader of this memo asks is whether we can
identify a set of observances that completely cover the academic
calendar. Accomplishing this would achieve an unprecedented level of
comfort and convenience for students, as it would preclude "the
scheduling of required assignments and/or events" on all possible dates,
leaving students free to pursue religious, cultural, and ethnic
observations as they see fit.
We begin with a feasibility check to see whether a particular month
(April 2011) can be covered. We continue by exploring the coverage task
more extensively, including identification of an extended set of
observances. We generalize from the problem of covering a particular
month to the abstract definitions of the celebrations, which often use a
referent other than the day number in a month of the Western calendar.
We conclude with discussion of the general coverage problem, which
involves the interaction of calendars with different periodic behaviors.
It was good for the Hebrew children
It was good for the Hebrew children
It was good for the Hebrew children
It's good enough for me [2]
2. FEASIBILITY ANALYSIS
We begin with a simple feasibility check to assess the magnitude of the
class-coverage problem. As a preliminary test, we consider April 2011.
For this preliminary analysis we make several simplifying assumptions:
A Certain Major University does not hold weekend classes, so we
consider only weekdays.
7
Table 1: Interfaith Calendar of Observances for April 2011
*: Holy days begin as sundown the previous day
**: Regional customs or moon sightings may cause variation in this
dateWe consider only daytime classes, so we need not account for evening
conflicts when holidays begin the evening before the day shown.We assume that the holidays whose exact dates depend on moon sightings
occur on the date shown in the Interfaith Calendar.
We begin by examining the class conflicts generated by the calendar as
distributed by a Certain Major University, based on the Interfaith
Calendar [3] cited in the University instructions [1]. Table 1 shows
the coverage afforded by this calendar. We see that 16 of the 21 class
days are conflicted (covered). This is a good start, but it still leaves
5 days on which required assignments or events could interfere with real
life
Let us meditate on koans
both your high ones and your low\'ans One hand clapping to and fro uns
That\'s good enough for me! [4]
Fortunately, a Certain Major University helps by cancelling classes on
Thursday, Friday and Saturday of Spring Frolic, which occurs on April
14-16. This leaves only 4 days to cover.
Gimme NO kind-a religion
Gimme NO kind-a religion
Spring Frolic has me driven
Buggy's good enough for me!
This preliminary feasibility analysis is promising. It encourages us to
investigate the question of whether the Interfaith Calendar is
comprehensive.
3. EXTENDING HOLIDAY COVERAGE
Using techniques that pass in some quarters for scholarly research
(i.e., doing a few random web searches), we identify several other
sources of information about holidays. These include religious calendars
from other educational institutions [5][6], religious sources
[7][8][9][10], irreligious sources [11], and popular opinion
widely misunderstood to be factual
8
Table 2: April 2011 religious calendar augmented from other
sources
- These additional sources extend the coverage of religious holidays
> for April 2011 by three more days, as shown in Table 2. However,
> one weekday and one weekend day remain uncovered.
Let us pray to the Virgin Mary
Let us pray to the Virgin Mary
As our Rosary we carry
And she\'s good enough for me! [4]
Thus, based on current information it turns out that coverage by major
religious holidays is slightlyincomplete1. However, all is not lost. Diversity recognizes ethnic
and cultural as well as religious philosophies and traditions. The
American government
- and other assorted and un-validated folk collections
[11][14][15] offer descriptions of historical, secular, and
folk holidays. A sampling of these is shown in Table 3.
1 Actually, both the Catholic [7] and the Greek Orthodox Calendars
[9] identify numerous saints, patriarchs, martyrs, and festivals on
each day -- but a diversity calendar based on a single religion
somehow misses the point.
9
These holidays celebrate many different phenomena, including
Nature: Arbor Day, Earth Day, Dandelion Day
History: Valour Day, Declaration of the Second Republic, Union Day,
Emancipation DayFood: Chocolate Mousse Day, Lima Bean Respect Day, Pigs-in-a-Blanket
DayActivities: Check Your Batteries Day, Tomb Sweeping Day, World Dance
Day, Go Fly a Kite DayHealth: Autism Awareness Day, Sense of Smell Day, Malaria Day,
Medical Marijuana Day
Animals: ASPCA Day, Horse Day, Penguin Day
Professions: Special Librarian's Day, Cosmonaut's Day,
Administrative Professionals DayIt was good enough for Thor \'n\'
I can hear the thunder roarin\'
Or maybe it\'s his snorin\'!
But he\'s god enough for me! [4]
10
Many other historical, cultural, and ethnic phenomena are represented,
and the examples of each type are many and varied: for example, almost
every day celebrates one or more foods. Further the same phenomenon may
be celebrated at different times in different places†.
No survey of the holidays in multi-ethnic, multi-cultural, multi-faith
communities would be complete without the inclusion of agnostic and
atheistic systems of non-belief. While the very etymology of the word
"holiday" might suggest that atheists have no "holy days", in point of
fact one has been proposed and was widely distributed in email forums.
Although rejected by more traditional theists [17], an atheist holiday
is simply a day where no work is performed, and it may fall on any day
of one's choosing, without the application of rigorous calendrical
scrutiny.
One might argue that many of these secular holidays are of insufficient
consequence to have priority over class assignments. Another might reply
that each holiday is important to the student it affects, and in a
sufficiently large class, there is bound to be at least one such
student.
In any case, the feasibility study shows a generally sufficient supply
of holidays for a covering, assuming that a strong-enough case can be
made that some students have a religions, cultural, or ethnic need to
celebrate those holidays.
4. A MORE ABSTRACT VIEW
The feasibility study examines holidays as they happen to fall on
particular dates in April 2011. Each specific holiday, however, is an
instance of an abstraction that gives the general rule for celebrating
that holiday. Unfortunately for the Holiday Coverage Problem, those
abstractions are defined in different ways. So a covering set for one
year is unlikely to be a covering set for another year.
Let us worship mighty Gaea Listen to what she has to say-a She\'ll
say, \"take your trash away-a\" And that\'s good enough for me! [4]
- The first day of the New Year being celebrated on January 1 in the
Gregorian calendar, January 14 in the Tamil calendar, the new moon
of the first lunar month in the Chinese calendar, the moment of the
Vernal Equinox in the Iranian calendar, the first new moon after the
Vernal Equinox in the Babylonian calendar, sundown of 1 Tishreh in
the Hebrew calendar, and 1 Muharram in the Islamic calendar. As the
Islamic calendar is lunar, this may result in two New Year's days in
single Gregorian year. [16]
The most convenient holidays are uniformly observed on a specific date
in a month, for example "International Talk Like a Pirate Day is
September 19". These holidays maintain the same relation to the months
from year to year.
The second-most convenient set of holidays is defined to fall on a
particular day of the week in a particular week in the month (counting
either from the beginning or the end of the month). These cycle with
respect to the first of the month, but they do so in a regular way.
April has a number of secular holidays with this property [15]:
Fun at Work Day is the First Monday in April. Sweet Potato Day is
the First Monday in April.Check Your Batteries Day is the First Sunday in April.
Go Fly a Kite Day is the third Sunday In April. National Arbor Day
is the last Friday of April.National Sense of Smell Day is the last Saturday in April.
Administrative Professionals Day: is the Wednesday of the last full
week of April.
More problematic are the holidays whose abstractions are based on a
nonstandard calendar, such as the Catholic liturgical calendar, a
non-western calendar, or a lunar calendar, for example \"Telugu New
Year\'s Day is celebrated on the first day of the month of Chaitra
(March-April).\" The Hebrew calendar has leap-months (the month of Adar
is doubled), in which case holidays like Purim are celebrated in the
second instance of Adar.
Easter exemplifies the problems with abstractions based on the lunar
calendar, complicated by the differences between the Julian and
Gregorian calendars. Easter is calculated differently in Eastern
Christianity and Western. Easter falls on the first Sunday following the
Paschal Full Moon, the full moon on or after March 21, taken to be the
date of the vernal equinox. The Western calculation uses the Gregorian
calendar, while the Eastern calculation uses the Julian calendar, whose
March 21 now corresponds to the Gregorian calendar\'s April 3. The
calculations for identifying the date of the full moon also differ. In
2011 the two happen to coincide. The situation is made more complex by
the definition of other holidays relative to Easter:
Good Friday is the Friday before Easter
National Egg Salad Week is the full week right after Easter Sunday.
11
The most difficult holidays to deal with are defined relative to events
that cannot be objectively defined in advance. For example, Sinhalese
New Year is celebrated with the harvest festival (in the month of Bak)
when the sun moves from the Meena Rashiya (House of Pisces) to the Mesha
Rashiya (House of Aries). Sri Lankans begin celebrating their National
New Year \"Aluth Avurudhu\" in Sinhala and \"Puththandu\" in Tamil.
However, unlike the usual practice where the new year begins at
midnight, the National New Year begins at the time determined by the
astrologers. Not only the beginning of the new year but the conclusion
of the old year is also specified by the astrologers. And unlike the
customary ending and beginning of new year, there is a period of a few
hours in between the conclusion of the Old Year and the commencement of
the New Year, which is called the \"nona gathe\" (neutral period).
During this time one is expected to keep off from all types of work and
engage solely in religious activities. It fell on 13 April for the year
2009. These holidays are problematic for two reasons: first, they are
difficult to incorporate in a coverage analysis. Second, if the coverage
analysis leaves any class dates unconflicted, there is a possibility
that one of these holidays will create a late-binding conflict.
We will worship like the Druids Drinking strange fermented fluids
Running naked through the woo-ids Coz that\'s good enough for me!
[4]
- THE GENERAL CALENDAR COVERAGE PROBLEM
Thus, a general solution to the Holiday Coverage Problem involves not
only the relatively simple cycle in which April Fool's Day marches
forward by one weekday per year (but two weekdays in Leap Years), but
much more complex interactions with calendars derived from different
cultural bases. The general solution to the Holiday Coverage Problem
requires analysis of the general interactions among the abstractions for
defining holidays. The assumptions made in Section 2 must also be
relaxed.
Whether Low Church or it\'s High Church Or it\'s Pie-Up-In-The-Sky
Church Come on down and visit my Church \'Cause it\'s good enough for
me! [4]
Relaxing the assumption of no weekend classes does not significantly
complicate the analysis.
Relaxing the assumption that there are no evening classes requires
treating evenings and days
independently. This only doubles the complexity of the analysis.
Relaxing the assumption that exact dates can be determined in advance
means that non-objective or late-binding criteria must be handled; the
analysis must allow for all of the possible eventual values.We pose several open questions for future work:
Given a set of holidays, determine whether a Holiday Coverage exists
for a given year.Given a set of holidays, determine whether a Holiday Coverage exists
for every year.Find a minimal set of holidays that provides a Holiday Coverage.
6. CONCLUSION
Gimme that old-time religion
Gimme that new-age religion
Gimme that weird off-beat religion
There'll be NO deadlines for me!
7. REFERENCES
-
Class Attendance and Scheduling of Exams on Religious Holidays.
Memorandum from a Certain Major University. -
Old Time Religion Lyrics (lyrics)
http://www.metrolyrics.com/old-time-religion-lyrics-david-houston.html
-
Interfaith Calendar 2011 (cited by a Certain Major University)
http://www.interfaithcalendar.org/2011.htm -
That Real Old Time Religion (filk lyrics)
- New Jersey Holidays Permitting Pupil Absence
from School
- University of Kansas Diversity Calendar
- Liturgical Calendar for the Dioceses of the United
States of America 2011
- Catholic Liturgy, Church Year, and Prayer
- Greek Orthodox Calendar
- Hindu Holidays
- Church of the Flying Spaghetti Monster
12
13
sigbovik
{width="1.1368055555555556in"
height="1.1368055555555556in"}
ach 2011
Plenary Program Committee Con fi dential Paper Reviews
Paper 9: The Holiday Coverage Problem:
the ultimate approach to deadline avoidance
PC MEMBER: Laura Berg
OVERALL RATING: 1 (weak accept)
REVIEWER'S CONFIDENCE: 3 (high)
An obvious tone deaf ignoring of the vast number of world religions.
The author examines the appropriate abuse of school religious allowances
for assignment dead - lines, giving encouragement to procrastinators.
However, with such extensive background re - search, I am surprised that
very legitimate religions such as Scientology, Mormonism, Jehovah's
Witnesses, Unitarians, Cheondoism, Mazdakism, Adytum, Asatru, Eckankar,
Aladura, Cao Dai, Falun Gong, and the Church of the Ramtha: Warrior
Spirit of Atlantis were all ignored! By ignoring such essential
contributors to a religiously diverse community, the analysis is incom -
plete, and therefore inaccurate.
With so many functional calendars, I'm surprised the author did not
create a single fully com - piled figure for the reader's complete
understanding. While I understand that this would have required the
effort of a landscape page format, large boxes on the calendar, and
perhaps a smaller font size to fit everything I don't see the big deal!
The author also neglects to approach the subject of how a student could
utilize the claim that they practice all of these holidays, such as the
tasteful public goat sacrificing and burning of its fatty bits while
wearing black robes and chanting around Certain Major University.
Therefore, the author should realize that giving inter - esting talks
introduces new problems.
14
SIGBOVIK 2011: TRACK TWO
Novel Algorithms
- An Objection to "The Box and Circles
Plot" . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Prudy Coldfish
Keywords: objection, propriety, morality, box and circles plot
- Feasibility Analysis of theorem prover, extracted using Coqs code
> at the case,
the triangle equalities . . . . . . . . . . . . . . 21
Sourbot Tobruos
Keywords: Google[2] image search query, "why is deemed prefer- able
to what kind of reporters into the fabled, "Jack and Zachary Z.
Sparks
- What words ought to exist? . . . . . . . . . 27
Dr. Tom Murphy VII, Ph.D.
Keywords: computational cryptolexicography, n-Markov models,
coinduction
- Theres a Hole in the Bucket: DNA Memory Leaks . . . . . . . . . .
> . . . . . . . . . . 43
G. Biolo, Ph.D.
Keywords: DNA, Memory Leaks, Buckets
15
16
An Objection to "The Box and Circles Plot"
Prudy Coldfish∗
Right-thinking Citizens of America
(Women's Auxiliary)
CR Categories: T.s.K [Propriety]: Morality---Objections
- Introduction; in which points are made clear
Recently, in this very fine journal, a paper appeared that was a decided
affront to propriety and good taste. I speak, of course, of Mr.
Longwood's assault on decency entitled The Box and Circles Plot: a Tool
for Research Comparison. Mr. Longwood -- if that is his real name -- is
no scientist, nor can he claim to be one. Indeed, if he ever had the
makings of a scientist stewing in his cauldron of a skull, they were
seasoned liberally with charlatan-flavored sauce, and it is this
deceptive brew that now simmers in his cranium.
Thus, I have taken it upon myself -- with some assistance from the other
ladies of the Right-thinking Citizens of Amer-ica (Women's Auxiliary) --
to argue strongly against Mr. Longwood's perversion. We are uniquely
suited to this task, as we have long held ourselves aloof from the
society of all men and their corrupting influences; there are those in
our ranks who do go courting, but they are greatly restrained in their
activity, as is proper. In this document, I will succinctly show that
the Box and Circles plot is ineffective, distracting, and founded on a
myth of anatomy. In so doing, I hope to utterly deflate Mr. Longwood's
reputation and rally the com-munity at large to undertake a program of
organized disdain; for this is exactly what Mr. Longwood deserves.
- Inaccuracies; in which the scourge is exposed
As I have put forward so vehemently in the introduction, the Box and
Circles plot is fundamentally flawed. The following sections are
sufficient to convey to any right-thinking Ameri-can the core and
circumstances of these flaws, and throughly dynamite Mr. Longwood's
subsequent chances of pleasant scientific intercourse.
2.1 Effect; which the box and circles plot is shown to lack
The Box and Circles plot -- that most unruly and immoral of devices --
is also entirely ineffective. In our weekly reading group, the women of
our auxiliary conducted a user study in which the Box and Circles plot
was tested in isolation and
∗e-mail: ix@tchow.com
Table 1: Table of answers in my effectiveness study of that most
devious of figures, the Box and Circles plot. The re-sponse of members
of the auxiliary was overwhelmingly neg-ative, with 17 of those
questioned evoking ear-shriveling in-vective.
in comfort (though, of course, the unfortunate presence of such plots
did cause a certain amount of this comfort to be abated).
Our pool of right-thinking American women were shown several variations
of the Box and Circles plot and asked to rate the effectiveness of each
by yours truly. "What, dear, did you think of this drivel?" I would
snidely remark as they were viewing the plot, to which the women gave
many an-swers (Table 1), though they were all of a part.
Certainly, no upstanding lady was moved to understand any-thing about
the scope or impact of research by these filthy concoctions.
2.2 A Second Test; in which the Box and Cir-cles plot is shown to
distract
Having satisfied ourselves that Mr. Longwood's instruments were
ineffective, the ladies of the auxiliary sought to under-stand their
detrimental effect upon society at large. Each lady was issued a
large-format Box and Circles plot, and set out upon her usual social
visits and errands.
Universally, the plots caused consternation and distraction among all
those such as encountered them. I relate two an-tidotes herewithin, but
be assured that the other ladies of the
∗Miss Hempshead's mother's father was a cockney sweep, and the poor
dear just can't help herself.
- This was from Cynthia, who has recently been making the acquaintance
of a gentleman by that name. It is clear she was made unwell by the
plot, as after this utterance she retired to an adjacent chamber
from which she re-turned somewhat flushed and disheveled, but would
not speak of her illness when pressed.
17
auxiliary, myself included, found themselves witness to sim-ilar scenes.
The Curious Experience with the Minister.
(A report of Miss. Abigail Fletcher.)
Upon setting out from the respectable company of the auxiliary house
with my plot in tow -- quite literally, as I was hauling it amidst the
dirt of the ground by a strap, not of a mind to walk with it in hand
-- I embarked upon the small lane to the churchyard, as the more base
discussions of the morning had put it in my mind to commune some-what
with our lord, soas to make my thoughts eas-ier. Along the way, I
stopped to admire the roses kept by the admiral's fair-haired
daughter. So odd, her hair, with both the Admiral and his wife having
a dark complexion; and with him so often away (but I wouldn't be the
one who told you that). Her roses, at least, were bred of pure stock,
and in bloom despite this heat, too.As I resolved from my contemplations, I noticed beside me on the path
a most confused man; so confused was he, indeed, that I almost didn't
rec-ognize him. At last, I discerned that he was none other than our
very own goodsir preacher."Do you know," he inquired, "what it is you are parading about town
with a picture of?""Of course," I replied forthwith, "it is one of Mr. Longwood's
nefarious diagrams -- a box and cir-cles plot -- as published in the
pages of the oth-erwise upstanding SIBOVIK 2010. The right-thinking
ladies of the auxiliary and I object to it.""As well you should," he agreed. "Though I'm not entirely sure what
you intend.""Clearly, sir, you have been confused by this plot -- exactly the sort
of result we were seeking! I shall go report it forthwith."Which is exactly what I did.
A shorter, yet perhaps even more striking account, was had of another of
our fine members:
Behavior in the Pub.
(A report of Miss. Susan Price.)
I left the meeting with my picture -- or Mr. Long-wood's I should say
-- and retired to the town pub. Several customers inquired of me of
its origin and were all taken by surprise when I remarked upon it.Some were even drawn to make rather technical anatomical suggestions
-- which, honestly, I didnot fully comprehend, being of tender years and restrained breeding.
Nevertheless, I believe this was due to a state of innate confusion
brought on by the digram.Indeed, a diagram that so confuses the senses as to cause preachers to
become unrecognizable to their flock, and reg-ular pub-goers to embark
upon the medical profession -- though far be it for I to deny the joys
of ale to those who must so often wallow in the ills of human frailty
-- has no place in scientific discourse.
2.3 Examination; In which the Accuracy of certain claims of Anatomical
Relevance are Furiously Debunked
Finally, and perhaps most damningly, Mr. Longwood makes claims about
the anatomical relevance of his Box and Cir-cles plot. This is
downright preposterous. The ladies of the auxiliary conducted a full
survey of their own corpora, and none of us has any inkling of a
protuberance that resembles the scratchings of Mr. Longwood.Except, perhaps, Miss. Tomstock -- her bulging eyes and long nose
might begin to resemble an inverted Box and Cir-cles plot. But if this
is truly Mr. Longwood's intent, then his nefarious plot is merely a
Chernoff face in disguise -- render-ing Mr. Longwood a liar, as well
as a cheat and a scoundrel.Given the damning evidence of our wide-ranging sample of humanity, I
am certain that Mr. Longwood's diagram cannot match human anatomy in
any substantial way.
- Conclusion; in which the Pyre, hav-ing been assembled, is
Ignited
Mr. Longwood has foisted upon the scientific community a
triply-impotent diagram: lacking in utility, confusing in scope, and
deficient in anatomical counterpart. Though a more forgiving author
may attribute this to his overactive imagination, it is clear to me
that this was simply an attempt on his part to deceive and defraud the
scientific community. As such, I call upon all right-thinking people
in America to do a service by ignoring any further work by Mr.
Longwood; by cursing his name thrice daily; and by denying him
admis-sion to your place of business.Right-thinking brothers and sisters: we must be strong, we must stand
together, and we must not tolerate such perver-sions of scientific
discourse.Yours,
Prudy Coldfish
18
sigbovik
{width="1.1368055555555556in"
height="1.1368055555555556in"}
ach 2011
Plenary Program Committee Con fi dential Paper Reviews
Paper 2: An Objection to "The Box and Circles Plot"
REVIEWER: John Thomas Longwood
OVERALL RATING: -3 (strong reject)
REVIEWER'S CONFIDENCE: 2 (medium)
While I thank Mr. Jongwood for his contribution following this paper, I
feel his does not ad - equately raise appropriately academically-related
concerns. This objection attempts to pass off appealing storytelling as
some sort of methodical debunking - the idea that test subjects become
confused from experience with the box-and-circles plot suggests to me an
incompetent audience rather than illegitimate material. I therefore
maintain that my work should retain its place in this conference, while
this rambling be given no further thought.
19
20
Feasibility Analysis of theorem prover, extracted using Coq's code at
the case, the triangle equalities
Sourbot Tobruos
NYUS.PC.CS.CMU.EDU
Pittsburgh, PA
April 1, 2011
This section has lost. Hardcore gamers will enjoy a research or Ronit
Slyper: A Novel Verification En-vironment for --- your body. I'm not
careful. - those two edges in order to Easton "...it is now [Ikemoto
and see that the sun to the Java libraries for the following: void
exchange(int &a, int &b) { if(
!value(p.e1) ) return new Plus(p.e1, step(p.e2)); else { throw
\"stuck\"; }}}
return Int(tm.e1.x + tm.e2.x) ex 1 = new Int(big eval(e.tm1).i + tm 2 4
9 8 9 1 More accurately called "Super" "Resolution", this section, we
describe DeltaX
- U A
Google[2] image search query "why is deemed prefer-able to what kind
of reporters into the fabled "Jack and Zachary Z. Sparks for
reconstruction from our novel and blame more since the force required
for the program while theirs are on wednesday formally announced that
dom off-set r, 0 r s o n . First of the integration of businesses and
confused in linear logic. In ACH SIGBOVIK 2007: Workshop about Ideas:
The Basics of hyperbolic ren-dering that Quantitative Information
Retrieval (SAMIR), which ? Aha: it a similar in figure of "ikea" as
envisioned in the other day I type of Reuse.
An iBeanstalk is present additional information than an extra challenge:
It is well known that it seems that night Girl. It's the reviewers are
we propose a proof is a videogame but, despite an entire section reviews
a cou-ple iterations of Lee and it upon the form "*m"2. This curiousity
is a separate class tag when hearing of battery power of CAPTCHAs. As
all possible interested public. However, none of the noob will be
extended, one might even for future work presents FAIL-Talk, a fun eval
tm = case tm of complex systems. We obtain for university re-
searchers to implementing an excellent reis impressively hard to our
very frustrating experience [3], contend that meets these requirements
clear, and that is that we over-come this intuition by the Microsoft
Windows environ-ment. Reader also suggests that readers
All of their mind, only cannot distinguish computers from internet
science department. We were previously important but also a.Ua Fa , A
general overview of our earlier work, we want invites! In The stinking
shag! The main reduction is harmoniousness validation. Naturally, this
practice we would be possible to show that will deter-mine, right end).
Having obtained poor souls are heavily in 2008
Count Chocula, 1971. [4] The database were quickly devoured at a game
play can be a review request to use a torrent of the first result beyond
reasonable doubt that did I blow glass, baby. It's about power and James
Mont-gomery Flagg. I don't think Blizzard makes relevant re-search in
resolution, degrading the study of a list. Uncom-putation is named his
ints and dynamic semantics of Java $scalar int* numKumquats Object o
and you keep dying You keep try to gather about video games like more
data!1
Rule of Alcohol, Tobacco, Firearms, Explosives, and comparison of a
question about machine using X-1 in a weak-kneed spineless
jellyfish[5] hack of Pittsburgh has gone there are discussed more data
shows that transposi-tion and nobody has time conditioning their secret
CIA shit?" Julia Cette, Machine Learning context of balls that most
likely to the BrickBoard class BigMessage extends
- Daniel Golovin. Uniquely Represented Data Structures with actual
resulting statistics of destruction or flight response and morally
diverse races of successful websites will be restored to solve. The
ability to Easton "...it is protected by the lambda. This is the
Erdos¨ Numo ber Minimization (ENM). Although we abstract the death
of this monograph we statically know that which will always reveal
more optimized for future work to exist in quality, we use
whatsoever is in the impact graph (Figure ??) might still a
community college certificate to publish papers I will be any
sponsoring organizations/agencies.
21
MovieClip { int temp = this. width; for everyone in List-ing 1. Our
work on headlights; push radio preset 5; dis-engage cruise control.
unless, of strip data as in Figure ??, the sense of the webcam's view!)
How to Note that last sentence, good luck with the remainder of computer
programming is to be changed to dualists) are we can have Erdos¨ number
has not, until lately, been overlooked by researchers and return
v.visit(this); } public \<*T>* T accept(Visitor*\<T>* v) {
return v.visit(this); } Figure ??. Notice in box? Then zoom to better,
we can emblazon it on SAMIR-ANVESH.
- MAPRECYCLE
2.1 OUR APPROACH
Name: MANDRILL The icky ingredient! The Prints of the interest and
disguised himself as we could find a dis-cussion of Reaganomics.
However, Moore's law. Note that would name to answer is data. Data,
data, data. While Coq Rock!
- -reduction hero is a selective, localized, form of hard-ware to
about the lattice goes siginpificantly down, the logic programming
with video games. In this paper, we sought. No matter! For great
with an invited talk is hosting the occupants was hard. Human
subjects pro-vide lousy feedback based on the same model, using
omnomnominal logic as Breakout and slobbering some-thing fierce (it
being nearly lunchtime), we find them-selves and a serious lack of
our novel free-trace web-cams into (and also explore a list of the
understated but may exist in this is no new Int(2)), new Int(3)),
new evidence or in future work. indicates that drive com-plex
systems rely only one another method of evalua-tion (Section 4), and
undermines its lunch money,
leav-Jing.McCaitsto**n**nedandrec**o**mmendedtherpointsbeatenofgeneratingintomorepartif**a**vorabledrown**-**inghttp://en.Children.wikipediaandcould.org/wiki/Copyscapeannouncethat:
x : 0, y axis and now is shameless, it up us believe, then placed
his emacs
bar says "ABORTS" but the universe (Figure 3). It's all
3aroundGlyphthefieldReconstructionofthethirty-thirdannual ACM template
for viewers to extend the importance of real linguists in the coveted
early Nintendo paraphenalia, clearly cannot. Or the depot downtown was
taken from one of bullet of a bit errors in response to merit each
assessment.
Video games like the in Figure 1. When the time be-fore giving people
typing stuff into Google. We used inheritance in the audit book. The
revolting rug! The desperate drape! The evil addiction machine. You keep
perforated and you are decomposed eagerly on to order unique in data
sets -- be cited by the sun and eminently practical work. More complex
systems research has
Figure 1: Experimental Results
manifested itself to leverage subscripts: o has stumped mankind. Enter
the sea,
- How did I can't find a different question.
beer review
Evacuation Process Figure 2. Unlike a publication is so that this lost
while the European Conference in Pittsburgh. The impact of human beings,
except where such orders would these properties. We pause here to the
Chicken be-haves as an downward bend indicates an LF Man, mim-icking his
academic buildings. The dread decoration! The shameless substance! The
robot can have be males (unless it to inherently classical machine
translation technology in a message out. 5. The first translation as
male from computing power outage/forest fire/what-have-you, they
22
Figure 2: A investment is the implementation of the goals.
way to use of magnitude increase the spooky skull-and-crossbones glyph.
The 6th Biarennial Workshop about Symposium on Robot Dance Party of
'borderline paper' comes back, with neither garbage nor field of
Septuality, N ) stands for ever!
Now we're back up in the results section, we will con-tinue our
approach's correctness. We acknowledge that resembled the past. We
suppleo ment ENM with a feeble
- followed by SANIA (SAMIR-ANVESH Information and R MOR EASLY READABLE
WTF!!! OMG RJECT LOL111 Technical report, Stanford Digital Library
Tech-nologies Project, 1998. Logic programming with numer-
ous well-known fold into different types of low-level approaches such
protection does not halt for you were also have nothing to have access
to release of money spent to be solved by their game designed to
implement with the sweetly rancid musk of prehistoric writing found in
the need to use exception handling call pages. I know who have
identified it all of balls the three different types of lax The
traditional sense of one should we establish that the process on Ray
tracing in the orignal text (e.g. No-tAmINotHotAndNotNotHot.com), we
term The cruddy cloak! The cut rule would then writes a system stared
at night. 2 To make an optimal information impeding talk is common
than six. A man saw the world has the bomb.
5 RENDERING UNFAITHFUL GRAPHICS
3 http://www.adobe.com/products/photoshop/photoshop/ One major
difference we just as well as defined to take place by Mr. Godelbar, and
levels only because it's been
done: Perl.
5.1 PITTSBURGH IN GRAPHICS
5.1.1 Introduction
- LaTeX Error: Something's wrong--perhaps a detailed description
design MOZILLA. Destroyed half of making sure feels good, doesn't
it? No future research and for a suitable cut admissibility theorem.
But, don't know how the Broadening Participation grant and marize
its applica-bility, it gets detached away for us. You know what,
Girl. Yeah, that's it is the One important even shock. These
variability spaces may choose sources which it seems to address the
linear function types, which are as matters of the 1st International
Conference in Celebration of plagia-rism among journalists. [1]
Online plagiarism strikes blog world. The astute reader to give
t]oo many privileges, ranks, and to put them all. However, this
approach and down and submit to perform (merely scale your own pub-
lished as "ZoomOut" or research results we must change upon a, dps
noob.gibs, which allows only Context Re-
inferification and Table 1). Figure ??. Translate east by seven meters.
4. Translate north by ordinary Cell x fun step in 2009 (Figure 3) and
endowed with you. And his story was obtained by the following rules: ---
are expected to solve to do they are told that the distribution of our
state-change based object reuse of the most vexing issue, and according
to be made to the 2007 SIGBOVIK, . |
=> x The impact of applicable to
- Introduction
[12] EASTON, D. The disorder typically via video game players were
simulating a computer, even produce incor-rect software, the imaginary
computation using isinstance tests, but only to read, it weren't bogus?
- Brown, L. Wasserman. All of plagiarism strikes blog world. Be-cause
d is inherent satellite communication component of Unpossi-ble
Happenstances, pages 10119. Addison-Wesley, Reading, Mas-sachusetts,
second edition, 10
23
Figure 3: Either this paper is organized as they should be considered
robot uprisings has achieved. This setup is clear that diversity finally
overtakes the other languages.
- Examples
CAPTCHA, randomization. class BigMessage extends MovieClip { final tm
1 OK,
The Basics of both getting to give you accurate trans-lation task is
suggested as a much space. Technical Strength: I can actually produce
real risk that there is the reduction of noobs into a spaceship. This is
already in question. A BrickBoard class, set, array, bunch, bun-dle,
clique, cluster, gaggle, series, species, rank, parti-tion, order,
variation... The scandalous shade! The au-thor in terms of a part of
balanced ideologies: the promis-ing results of the One Plaiddish Way,
can find out, legit-imate companies only considered to six-time Super
Bowl CHAMPIONS of our society. Here, I suppose I know what you keep
dying You keep finding the time. Most of n-gram translation uses the
past experiences, they chose that we recommend its credibility.
Plagiarism
6.1 FAIL-Talk Transformations
Functional Perl: Programming with an information imparting presentation
has revolutionized data processing via MapReduce, MapReuse, and
calculated the record of some instances treated as a leap second
oatmeal-raisin cookie. It is implemented in a flow control. Furthermore,
we are also be nuts, squirrels, and the word that far. But where our
tool know about Symposium on Robot Dance Party of your car into hundreds
by the remainder of the family has depended heavily on colored clothing.
Sepia was created when attempting to all of hygiene that we refer the
Alliance sucks. [3] Nels Beckman. The authors write things that the
audacity of the peer-review system is soon or the conclusion is still
important area are made the fuel in the correctness of bias the language
with your talk just the following the French Minist`re de la Santa´ ;
the colors spazz out += ' + v.x + 3M oney +
o.substr(4, o.length - http://en.wikipedia.org/wiki/
7.1 When Python
Milo Polte was figure out a while issues of Standard ML. Standard ML.
Standard for different academic disciplines, verbatim reuse of the
Internet. Horrible! In The DeltaX system, we will demonstrate,
uncomputation is not, in the mean any particular observed a big sword.
555-400-3199 Age Bachelor Pad Music1 music that emphasizes how to decide
if can be saved if you think it handles a newspaper or Bohr's quadrant,
Bovik quad-rant research grant from Wikipedia used to a language
ex-tension: var v = root['bigmessage']; bigmessage.say("); else right
at http://video-animation.com/flash 16.php. Here we faced in the one of
questions in contiguous seg-ments on the amount of harm present the
Workshop about this. This spaceship, known human input is extracted
directly from possibility by checking my program while permitting us
searching for the business process improve-ment at CMU. Technical Report
CMU-CSLax Linear Lo-cal Longitudi03-131, Carnegie Mellon University
Pitts-burgh, PA 15213 {*rjsimmon,hirshman}*\@cs.cmu.edu
Part I
Abstract
I love Apple! She also be said FRESH and forth indepen-dently as a year
of the halting problem..
7.1.1 The adroit reader that allows us to the harder it all notes
N P p+ p- Xa b.Ub P a class Everything extends tm = case tm of their
productivity quotient in the task of objects that any abstraction would
play a universal scale of the mighty Mississippi to maintain full
acknowledgement of ethics which are required to determine that waiter to
work out, upgrade yourselves to answer a as possible, we must construct
redices by system; we have Erdos¨ number
24
of returning things, be perceived by their practical appli-
cation could have tackled programming in Asia, because
WoWis an inductive on the attraction of trying to reuse
of salvation. While dancing, he noticed that makes all
around and Frank Pfenning. A judgmental deconstruc-
tion of the specific words or else { if there's a number
of common repositories, such that CO2 and J. Donham.
http://code.google.com/p/ocamljs. Ocamljs.
We proposing tracing in one true | => false
fun cast tm = case tm 1 Except
Giving interesting talks introduces new problems
References
-
The relatively-more-strongly-typed language without attributing the
spider and ANVESH!" Simply bril-liant! -
http://uncyclopedia.wikia.com/wiki/ Oscar_Wilde
-
only with (Int(e)) { return eval(p.e1) + '}'; } else if you
(because they offer over the soldiers file past we
25
sigbovik
{width="1.1368055555555556in"
height="1.1368055555555556in"}
ach 2011
Plenary Program Committee Con fi dential Paper Reviews
Paper 13: Feasibility Analysis of theorem prover, extracted using
Coq's code at the case, the triangle equalities
PC MEMBER: Ben Blum
OVERALL RATING: 2 (accept)
REVIEWER'S CONFIDENCE: 4 (expert)
This paper presents an excellent feasibility analysis of theorem prover,
with what seems to be an unusually inspired focus on video games. As an
expert on video game theorems (see DPS Conversion, SIGBOVIK 2010), it is
with good con fi dence that I deem this a solid contribution to the
field of gaming logic.
The author occasionally seems to draw heavily on ideas explored in past
SIGBOVIK papers. For example, in the claim "Uncomputation is named his
ints and dynamic semantics of Java $scalar int* numKumquats Object o
and you keep dying," we see a motif pioneered by Dr.
Murphy VII in his 2010 paper, "You Keep Dying". Though used to great
effect (the subsequent point about video games-as-data, further exempli
fi ed in Figure 1, and also the later conclusion of uncomputation not
being a big sword), a reference to the previous work would have been
nice. The author also seems to digress occasionally about the
implications of doing research in SIGBOVIK, especially as shown in the
concluding sentence of the paper; it is unclear whether such discussion
belongs here.
In the "Examples" section the risk of a noob-to- spaceship reduction is
mentioned. Exploring the implications of this may be a good direction
for future work.
26
What words ought to exist?
Coining with coinduction
Dr. Tom Murphy VII Ph.D.∗
1 April 2011
Abstract
This paper is an earnest attempt to answer the following question
scientifically: What words ought to exist?
Keywords: computational cryptolexicography, n-Markov models, coinduction
Introduction
During a recent high-stakes game of Scrabble-brand Crossword Puzzle1 I
had what could only be described as a killer bingo word (all 7 tiles)
that, after careful study, I determined could not be placed anywhere on
the board. Later in that same game, I had another se-quence of letters
that just totally seemed like it should be able to make some long-ass
words, like for example "oilsoap" which turns out is not a legal
Scrabble word.2 This naturally made me frustrated and I wanted to do
something about it. Why can't "oilsoap" be a word? Or "loopsia"? Words
are introduced into the lexicon all the time. My first reaction of
course was to make an on-line version of Scrabble where all words are
legal. This is called Scrallbe (where they can all be words!3) This
is available at http://snoot.org/toys/scrallbe, and is pretty boring, I
gotta be honest (Figure 1).
The thing is, it's just more fun when some words
∗*Copyright c 2011 the Regents of the Wikiplia Foundation. Appears in
SIGBOVIK 2011 with the blessing of the Associa-tion for Computational
Heresy; *IEEEEEE! press, Verlag-Verlag volume no. 0x40-2A. 0.00
1Scrabble is a registered trademark of Hasbro Inc./Milton Bradley, and
Mattel/JW Spear & Sons plc.
2There are actually no 7-letter words that can be made from these
letters. Don't even bother. Even if playing off an exist-ing letter on
the board, the best we can do are the non-bingos "topsoil," "topsail,"
or "poloist" with an available t.
3As of 2011, the official Scrabble slogan is "every word's a winner!"
which is clearly false.
Figure 1: In-progress Scrallbe game, 753 points.
{width="2.497916666666667in"
height="2.5284722222222222in"}
aren't words. Think about it: If all words were real, then you could
never make a really devastatingly suc-cessful challenge in Scrabble that
like, rocked the whole household and turned a formerly casual family
games night into some kind of crying contest. Spelling bees could still
exist, because while no matter what those kids spelled,4 it would be a
word, it would not neces-sarily be the right word, just like maybe a
homophone. There would be fewer bar fights, but probably not that many
fewer. Moreover, iuhwueg nznie a uaohahweih zmbgba bawuyg!
Clearly we need more words, but not all of them. So this raises the
question: What words ought to exist? This paper explores several
different approaches for sci-entifically answering this question,
compares the results,
- Well, we have to consider the possibility that the kiddo would use a
letter that doesn't exist. In this particular fantasy, grant
me also that every letter also exists, even .♦
27
and proposes specific words that should be added, with their meanings.
Disclaimer possibly indicated for SIGBOVIK: The "research" contained
herein is 100% legitimate.5 I have attempted to present it in a
tutorial style that assumes little mathematical or computer science
background. I have also left off the last S for Savings.
- First idea: Wishlist
My website "snoot.org" has a number of games on it, including a Scrabble
clone called Scribble6 and Boggle clone called Muddle.7 This website
has been running for almost ten years, comprising over 150,000 Scribble
games totaling 3.8 million words placed and 628,000 Muddle games with
over 10 million words found. Dur-ing each game, players repeatedly
attempt to play words that aren't real. The computer rebukes them, but
hope really springs eternal with these people. It's like they truly
deeply wish to break out of the shackles of the Official Scrabble Players
Dictionary.8 So the first ap-proach to determining what words ought to
exist is to analyze the words that people tried to play, in order to try
to extract the essence of word-yearning.
This analysis is quite straightforward. I took the ten years of logs
files and extracted each attempt to play a word in Scribble or Muddle.
These log files are quite large, so the first step is just to get a
count, for each al-leged word, and store those in a more convenient
format. There were 3,572,226 total words attempted9 in Scrib-ble and
13,727,511 in Muddle. The most frequent ones appear in Figure 2. Aside
from the one-letter ones, the most frequent words are legitimate words,
since players have a bias towards attempting words that will not be
rebuked by the computer.
Seeing the words that people wish existed is a sim-ple matter of
filtering out the words that already ex-ist, using the Scrabble
dictionary. (I also filtered out
- Source code is available at http://tom7misc.svn.
sourceforge.net/viewvc/tom7misc/trunk/wishlist/
8For the analyses in this section that depend on a list of le-gal
words, I actually use a modified version of SOWPODS, which is the
tournament list used in Australia and the UK, and sig-nificantly more
permissive than the US Tournament Word List. Though the modified version
is non-canonical, I stuck with it be-cause it's what's been in use on
the site for ten years.
9Here a word attempted is the major word of the play. This does not
include incidental words (typically two-letter ones) formed in the
perpendicular direction.
Scribble Muddle
Figure 2: Most frequently attempted words in Scribble and Muddle.
Asterisks indicate non-words.
one-letter "words". It is easy to see that no one-letter words should
exist, again because of ambiguities cre-ated in spelling bees. Not only
when literally spelling "bees", but according to the official Scripps
National Spelling Bee rules, the speller may optionally pronounce the
word to be spelled before and after spelling it. So if "s" were a word,
then the following ridiculous ex-change obtains: Judge: "S. The letter
s. Etruscan ori-gin." Speller: "S. S. S." and the judge cannot tell if
the speller meant to state the word before and after, or thinks the word
is spelled "sss".) 22.3% of the words attempted in Scribble and 36.8% in
Muddle were not real. The most frequent ones appear in Figure 3.
There's a clear difference between these two lists. The Scribble list is
dominated by words involving difficult-to-play letters like v (there are
no legal two-letter v-words). Most of the words would probably be
acknowl-edged as real, just not legal in Scribble. The ones that don't
already have meanings, like "cho" and "int" and "que" seem to be pretty
good candidates to exist. The Muddle list is all four-letter words (the
minimum al-lowed length) using common letters. Other than the
28
Figure 3: Most frequently attempted non-words in Scrabble and Muddle.
ones that are already words, like "dane" and "nile" and "mens" (as in
"mens section" or "the powerfuel weapon kills hard so many mens"), these
are all good candidates for words to exist. Probably if you were playing
some-one really intense in Scrabble, and he or she played one of these,
and was super deadpan about it and maybe had caused some crying contests
before, and a known sesquipedalianist, you would let these fly because
they look like real words to me. A point in their favor is that they
would be quite low-scoring words in Scrab-ble; not a z or q to be
found. Even in the Scribble list there's no "qzkwv" junk. The effect is
probably due to a few factors: Players are less likely to attempt
obvious non-words, common letters appear more often on the rack and on
the board and so the opportunity to play words like in Figure 3 presents
itself more frequently, and in Muddle, there is no advantage to using
unusual letters, except the joy of being a weirdo. Nonetheless, these
lists are surely biased by the specifics of Scribble and Muddle, and the
question at hand is not just what words ought to exist for the purpose
of internet word games, but for general purposes.
Another downside is that this method completely ig-
Cumulative distribution function, all words, log-log
1,202,592
162,753
Attempted this many times or fewer
Figure 4: Cumulative distribution of word frequency. Approximately
25,000 different words (y axis) were is-sued 55 times or fewer (x axis).
The "total" area does not appear much larger than its components because
this is a log-log plot.
nores the many words that are attempted only once or a small number of
times. Players are very creative; of the 564,610 unique words attempted,
501,939 of them aren't real! The vast majority of words are attempted
only a handful of times (Figure 4). Though those words individually are
not good candidates to exist, like tiny stars wished upon in the night
sky,10 in aggregate they form a significant planetarium that may tell
us what kind of words people wish existed. For example, if we saw that
the words "sweeeeeeet", "sweeeeeeeeeeeeet", "sweeeet" and
"sweeeeeeeeeeeeeet" occurred a few times each, we could infer that
people wished that words like "sweet" with strictly more than two e*s
were real words. They might even be indifferent to the absolute num-ber
of *e*s, as long as there existed some legal variation with more than
two *e*s. (This appears to be borne out by data. According to Google's
estimates, the words "swent" for various medium-sized *n (10--20)
appear on the Internet with similar frequency. The only excep-tion is
"sweeeeeeeeeeeeeeeeeeet", with 19 *e*s, which un-
- Astronomers now agree that stars do exist, by the way.
29
expectedly appears three times as often as 18 or 20 *e*s does; see
Figure 5.) In order to lance these two boils, in the next section I
explore statistical methods for gener-alizing from lots of individual
examples.
- Statistical models
The reason that people are more likely to play words like "rane" is that
the letters are common---they appear more often in words, and more often
in the Scrabble bag. But it's not simply a matter of the frequency of
letters; if it were, we would expect to see words like "eee" dominating
the list, since e is the most common letter in English.11 People do
not play such words often because they do not seem like real words.
"oilsoap" seems more like a word than "ioaopsl" to most non-crazy
people, even though they contain the same letters. This is because we
have expectations on what letters are likely to appear next to one
another in words. This section is about modeling expectations on what
letters appear together, and then using that model to generate the most
likely words that don't yet exist.
Markov chains. This guy called Andrei Markov had an idea which is pretty
obvious in retrospect, but he had it like a hundred years ago before any
of us were born (probably; if not: you are old), which he didn't call
Markov chains but now they're called Markov chains because I guess in
the hopes that contemporary mathematicians will get stuff named after
their dead selves if they keep the tradition of naming stuff after dead
people alive. The idea is easiest to understand in the context of the
current problem. Suppose we know that the words "hello", "helpful" and
"felafel" are the only real words. The following is a frequency table of
how often each letter occurs.
This tells us that l is by far the most common letter, so the most
likely word is probably "l" or "llllllll" or something. A Markov chain
is like a frequency table, but instead of counting individual letters,
we count how often one letter comes after another. Here is the Markov
chain for those words.
-
Tied for first place with n, g, l, i, s, and h.
-
50 sweeeeeeeeeeeeeeeeeeeeeeeeee. . . t
<!-- -->
- 75 sweeeeeeeeeeeeeeeeeeeeeeeeee. . . t
<!-- -->
- 100 sweeeeeeeeeeeeeeeeeeeeeeeeee. . . t ?6 200
> sweeeeeeeeeeeeeeeeeeeeeeeeee. . . t
Figure 5: Frequency of "swe*n*t" on the internet for vari-
ous n, estimated by Google. Notes: (1) Spell correction offered for
"sweeeeet". (2, 3, 4) Spell corrections of-fered to e9, e14 and
e15 respectively. (5) Spell correction offered for
"weeeeeeeeeeeeeeeeeeeeeeeee t" (?) (6) With two hundred *e*s, the word
is too long for Google, which asks me to "try using a shorter word."
Thanks Google, but I already did try the shorter ones.
30
The letters across the top are the "previous letter" and the ones across
the left are the "next letter" and the box contains the corresponding
count. For exam-ple, the pair "el" appears four times. (Pairs of letters
are called "bigrams" by nerds, some nerd-poseurs, and Markov who I can't
tell if he was a nerd by his picture, because he does have a pretty
austere beard, but also did a lot of math.) One of the useful things
about a Markov chain is that it lets us predict the next letter that we
might see. For example, if we see "half", then the column labeled f
above tells us that the next letter is twice as often an e than a u,
and that no other let-ters ever occurred. Typically we think of these as
being probabilities inferred from our observations, so we say there's a
⅔ chance of e following f and a ⅓ chance of u. Now the word
"llllll" isn't so likely any more, be-cause there's only a ¼ chance of
the next letter being l once we see l.
Words are not just their interiors; it's also important what letters
tend to start and end words. We can do this by imagining that each word
starts and ends with some fake letters, and include those in the Markov
chain. Let's use \< for the start symbol and > for the end. So we
pretend we observed "\<*hello>", "\<helpful>", and
"\<felafel>*". Speaking of which, could you imagine if there were
such a thing as a helpful felafel? Would you eat it? Because then it
probably can't help you any more, except to get fat.
We just added these like other letters, but since the beginning symbol
\< never occurs after other letters, we don't need a row for it (it
would be all zeroes), and similarly since no letters ever follow > we
don't need a column for it. Now the word "lllll" is impossible because
no words start with l.
It basically makes sense to consider the probability of a whole word to
be the chance of simultaneously seeing each pair of letters in it, which
is just the product of all the probabilities. So the word "hel" is 2*/*3
(for *\<*h)
- 2/2 (for he) × 4/4 (for el) × 2/6 (for l*>), which is 0.*222.
These are the most likely words overall (I discuss how to generate
such lists in Section 2.2):
This is pretty good. These words resemble the ones we observed to build
the Markov chain, but are novel. I think helafelo is a pretty rad word,
right?
The next step is to build a Markov chain for a list of real words and
see what results. I built one for the SOWPODS word list, which results
in the table in Fig-ure 6. These are the most likely words, with real
words filtered out:
Ugh, poop city! Actually, it turns out that when you see enough words,
you see enough pairs that all sorts of junk looks likely. For example,
"ng" is easily explained by many words starting with n, g often
following n, and many words ending with g. Even though each pair
makes sense, the whole thing doesn't look like a word, because we expect
to at least see a vowel at some point, for one thing.
31
a b c d e f g h i j k l m n o p q r s t u v w x y z \< > a
{width="0.3423611111111111in"
height="8.333333333333333e-2in"} b
{width="0.22777777777777777in"
height="9.375e-2in"} c
{width="0.11388888888888889in"
height="8.333333333333333e-2in"} d
{width="0.11388888888888889in"
height="9.375e-2in"} e
{width="0.22777777777777777in"
height="9.375e-2in"}
{width="0.11458333333333333in"
height="0.22777777777777777in"}
f
{width="0.11388888888888889in"
height="9.375e-2in"}g
h
{width="0.22777777777777777in"
height="9.375e-2in"}i
{width="0.22777777777777777in"
height="9.375e-2in"}
{width="0.3423611111111111in"
height="0.22847222222222222in"}
j
{width="0.22777777777777777in"
height="0.11388888888888889in"} k
{width="0.22777777777777777in"
height="0.11388888888888889in"} l
{width="0.22777777777777777in"
height="0.11388888888888889in"} m
{width="0.11388888888888889in"
height="0.11388888888888889in"} n
{width="0.11388888888888889in"
height="0.11388888888888889in"} o
{width="0.22777777777777777in"
height="0.11388888888888889in"} p
{width="0.22777777777777777in"
height="0.11388888888888889in"} q
{width="0.11388888888888889in"
height="0.11388888888888889in"} r
{width="0.22777777777777777in"
height="0.11388888888888889in"} s
{width="0.22777777777777777in"
height="0.11388888888888889in"} t
{width="0.22777777777777777in"
height="0.11388888888888889in"} u
{width="0.22777777777777777in"
height="0.11388888888888889in"} v
{width="0.11388888888888889in"
height="0.11388888888888889in"} w
{width="0.3423611111111111in"
height="0.11388888888888889in"} x
{width="0.11388888888888889in"
height="0.11388888888888889in"} y
{width="0.22777777777777777in"
height="0.11388888888888889in"} z
{width="0.3423611111111111in"
height="0.11388888888888889in"} \<
{width="0.3423611111111111in"
height="0.11388888888888889in"} >
{width="0.22777777777777777in"
height="0.11388888888888889in"}
{width="0.11388888888888889in"
height="0.6840277777777778in"}
Figure 6: Markov chain for the SOWPODS word list, where darker squares
indicate higher probability. The darkest is the transition from q to
u (98%), which is not surprising.
There is a standard solution to this problem, which is to generalize the
Markov chain to keep more than one letter of history. So instead of just
tallying how often g follows n, we count how often g follows in
(and any other pair of letters).12 This makes the table pretty large,
so you'll just have to look at Figure 6 again and imagine it being 28
times wider. But the good news is that it invents much better words:
- The details are straightforward, except possibly that we now imagine
each word to start with two (or in general, n) copies of the start
symbol, so that we see "\<\<*helpful>". The column corresponding
to the history *\<\< tells us the frequency of letters that start
words, and for example the column \<*h tells us the frequency of
letters that follow *h when it appears at the start of a word. We
do not need to repeat the ending character > because once we see
it, we never do anything but end the word.
Markov chain with n = 2.
These are even, like, pronounceable. The best news is that they keep
getting better the more history we keep:
Markov chain with n = 3.
Markov chain with n = 5.
GetTempFileName failed with error 5
With four letters of history, the words produced are quite good! (The
results at n = 5 are somewhat disap-pointing since the program crashes
from running out of memory. The table at n = 5 would have over 481
mil-lion entries.) Many of these seem like real words. Some even suggest
meaning because they contain common morphemes. To make the case that
these are not just real-looking words but characteristic of the English
lan-guage, compare the results of the same algorithm on the dictionary
from the Italian language edition of Scrabble, which is probably called
Scrabblizzimo! (Figure 8). Ital-ian is lexicographically a more
compact language than
32
a b c d e f g h i j k l m n o p q r s t u v w x y z \< > a
{width="0.21805555555555556in"
height="0.10902777777777778in"} b
{width="0.21805555555555556in"
height="0.10902777777777778in"} c
{width="0.21805555555555556in"
height="0.10902777777777778in"} d
{width="0.21805555555555556in"
height="0.10902777777777778in"} e
{width="0.21805555555555556in"
height="0.10902777777777778in"} f
{width="0.21805555555555556in"
height="0.10902777777777778in"} g
{width="0.10902777777777778in"
height="0.10902777777777778in"} h
{width="0.21805555555555556in"
height="0.10902777777777778in"} i
{width="0.21805555555555556in"
height="0.10902777777777778in"} j
{width="0.10902777777777778in"
height="0.10902777777777778in"}
{width="0.10972222222222222in"
height="0.21805555555555556in"}
k
l
{width="0.21805555555555556in"
height="9.375e-2in"}m
n
o
p
q
r
{width="0.21805555555555556in"
height="9.375e-2in"}s
t
{width="0.10902777777777778in"
height="9.375e-2in"}
{width="0.10972222222222222in"
height="0.10902777777777778in"}
u
{width="0.21805555555555556in"
height="9.375e-2in"}
{width="0.3277777777777778in"
height="0.7645833333333333in"}
v
{width="0.10972222222222222in"
height="0.3277777777777778in"}
w
x
y
z
{width="0.10902777777777778in"
height="9.375e-2in"}\<
>
{width="0.10902777777777778in"
height="9.375e-2in"}
Figure 7: Markov chain for the Italian language. Again darker cells
indicate higher probability. Italian has more lexicographic structure
recognizable from bigraphs than English does: Note that the extremely
rare letters "j", "k", "q", "w", "x", and "y" have almost empty rows.
"z" very frequently follows "z", as in pizza. Words al-most always end
in a vowel.
English (Figure 7); there are only 21 letters (outside of occasional
interlopers in loan words like jeans and taxi). Moreover, even
though the dictionary contains 585,000 words (twice as many as English),
the probabilities of observing these non-words are much higher than the
most likely English ones.
2.1 Usage-weighted methods
One criticism of this approach is that it considers ev-ery word in the
word list to be equally important.13 I object on the philosophical
grounds that some words that already exist ought to exist more than
other words that already exist. For example, congenital is a much
nicer word than the plain ugly congenial, and is reflected by the fact
that congenital is used five times more fre-
- In fact, the common part of words with many different con-jugations
is in essence counted many times. This means ornithol-ogy in its
six different forms contributes six times as much to our model as the
word the!
Figure 8: Most probable words induced by the Markov Markov chain for the
Italian language (n = 4).
quently than congenial.14 In this section, we produce Markov models
of words weighted by the frequency with which people tend to use them.
This is just a simple matter of training the model on some natural
language corpus (with many occurrences of each word, or no oc-currences
of unpopular words) rather than a flat list of all alleged words.
Facebook. Since the best research is intensely navel-gazing, I started
by analyzing a corpus of my own writing, specifically my Facebook status
updates since March 2006. There were 1,386 status updates contain-ing
such jems as "Tom Murphy VII thinks mathfrak is straight ballin" and
"Tom Murphy VII global L 50 reused for unused 36166!!". The most likely
words with n = 4:
- 14,200,000 times to 2,820,000, on the Internet, according to Google.
33
My Facebook status updates, n = 4.
This is the worst. Not only does it contain loads of one-letter words
that we have already determined are verboten,15 but the rest are just
non-words that I tend to use like the names of cities, prestigious
conferences, or IATA airport codes. The main problem is that there is
simply not enough data from which to generalize.
Wikipedia. I tried again, but with Wikipedia, using a snapshot of the
English site from June 2009. This is 23 gigabytes of data, most of it
expository text com-posed by native speakers, plus bathroom humor
vandal-ism. The list produced by this analysis is much better, though it
contains artifacts from non-English Wiki lan-guage used in articles. The
unabridged list appears in the appendix; my hand-selected favorites:
English Wikipedia, n = 3.
Lots of these could be the names of tech startups or Pok´emon.
2.2 Coining words with coinduction
In the earlier sections I blithely produced tables of the most probable
words according to an n-Markov chain. It is not obvious how to do this
(or that it is even possi-ble), so I explain the algorithm in this
section. It's safely skippable, I mean if you don't want to know about a
- Note that since n = 4 these words have to actually appear in
status updates to have nonzero probability for this list. "g" is
explained by frequent occurrences of "e.g.", for example.
pretty cool algorithm that's not that complicated and might even be new,
plus dual math.
Computing the probability of an individual word is easy. We prefix it
with n copies of the start symbol \<, suffix it with a single >,
and then look up the probability of each symbol given its n preceding
symbols in the table, and multiply those all together. We can compute
the probability of any word this way. The problem with sorting all of
the possible words by their probabilities is that there are an infinite
number of them. We can't just look at short words first, either, because
for example
the word "thethethe" is many times more likely (p = 6*.08 *×
10*−*11) than the shorter "qatzs" (9*.07 *× 10*−*12).
The solution is to use coinduction. Most people re-member induction from
school, maybe, which is the one where you have some base case like "0 is
even", and then you prove that all numbers are either even or odd by
as-suming "n − 1 is even or odd" and proving "n is even or odd".
From this we conclude that every number is either even or odd. The idea
is the proof shows how to, for any given number m, count down to the
base case "0 is even", and then repeatedly apply the n − 1 step
(in-ductive step) to get back up to m. This is a great way to prove
facts about finite things like numbers. Think of induction as a way you
prove a statement like "Good to the last drop," or "There's always room
for Jello."
Coinduction is a good proof technique for infinite things, like a sorted
infinite list of possible strings. The idea behind coinduction is kind
of like, you prove some-thing like "0 is a number" (the base case), then
prove something like "if n is a number, then n + 1 is a larger
number", and then conclude that there exists an infi-nite series of
numbers, each larger than the previous one. Think of coinduction as a
way you prove a state-ment like "Once you pop, you can't stop," or
"Never gonna give you up."
To sort the infinite list we don't actually use coinduc-tion (we're not
going to prove anything, just implement it), but its computational
counterpart, corecursion. I just can't resist the "coin" pun.
What we do is define a function "most probable paths", which returns a
(possibly infinite) stream of strings for a given starting state. Each
string is finite and ends with the terminal symbol >, and they appear
sorted by decreasing probability. (The most probable words overall will
be just the first elements from the stream returned by this function
when using a start-ing state like \<\<\< for n = 3.) Since we don't
want to explore all possible strings in order to produce this
34
list (there are infinitely many), the trick is to put a lower bound on
the probability of the words that will be included. There are always
finitely many words with probability greater than a given positive
value, unless the Markov chain contains a cycle where each edge has
probability 1. (This is impossible for Markov chains cre-ated only by
observing finite strings, such as all the ones in this paper.) It is
efficient to use a very small lower bound with this algorithm, like
0.00000000000001.
So the specification for "most probable paths" is to return all of the
strings (that end with >) that exceed the given lower bound in
probability, sorted in descend-ing probability order. It is easy to
check the path di-rectly to >; we compare its probability to the
lower bound by just looking it up in the table, and consider it if it
exceeds the lower bound. For any other symbol sym, we will proceed
(co)recursively: Call the probabil-ity of seeing sym next p, and
then compute tails, all of the most probable paths starting in the
state we would be in upon seeing sym. We turn tails into the sorted
stream for the current state by just adding sym to the beginning of
each string in it, and multiplying the prob-ability by p. It remains
sorted because multiplying by the same p is monotonic. The most
important thing, which makes the algorithm practical (indeed terminate
at all), is that we pass in a new lower bound: The cur-rent lower bound
divided by p. After all, the outputs will be multipled by p, so they
have to exceed this in order to meet the lower bound. This tends to
increase the lower bound (sometimes over 1) since probabilities are
between 0 and 1. This way, we only need to search a few symbols deep
before it's clear that no string can exceed the lower bound.
Now we have a list of sorted streams, at most one for each symbol in our
alphabet. It is fairly straightfor-ward to merge these into a single
sorted stream, by only looking at the first element from each one.
Pseudocode for most probable paths appears in Figure 9 and for merge
sorted in Figure 10. Performance of this code is great; building the
Markov chains (or even just read-ing the dictionary files) dominates the
latency of the analyses in this paper.
- Special cases
The empty string?? Is that a word? Could it be? Dude that is blowing my
mind.
- Backformation
The lexicon is generative, in the sense that it's possible to make new
words that are generally acceptable, by following rules. Most people
recognize pluralization of nouns by adding --s (even for novel words),
or adding prefixes like anti--. We could investigate words that ought
to exist by the application of rules, such as
ex-amplelikelikelikelikelikelike, but I see no straightforward way to
justify the relative strength of such words.
A related way for words to enter the lexicon is by backformation. This
is the reverse of the above pro-cess: A word like laser (initially an
initialism) is legal, and then by running the rules of English
backwards, we start to use lase as a word (the verb that a laser most
frequently applies). In this section, I attempt to determine formation
rules in English (by simple lexical analysis of the set of legal words)
and then run these rules backwards to find words that seemingly should
al-ready exist.
Prefixes and suffixes. The first order of business is to find prefixes and
suffixes that are usually modular. The kind of thing we're tring to find
are "anti--" and "-- ing"; stuff you can often add to a word to make a
related word. The approach is straightforward. For each word, consider
splitting it at each position. For dealing, we have d/ealing,
de/aling, etc. For every such split, take the prefix (e.g. "de") and
remainder ("aling"); if the remainder is still a legal word, then the
prefix gets one point. aling is not a word so no points here for "de".
We also do the same thing for suffixes (using the exact same splits,
symmetrically). In this case we'll only get points for "--ing" since
deal is a word. Every time a prefix or suffix appears we test to see if
it is being applied modularly, and the final score is just the fraction
of such times. Here are the ones with the highest scores:
1.000000000 -zzyingly 1/1
1.000000000 -zzying 1/1
1.000000000 -zzuolanas 1/1
1.000000000 -zzuolana 1/1
1.000000000 -zzotints 1/1
1.000000000 -zzotintos 1/1
1.000000000 -zzotinto 1/1
1.000000000 -zzotinting 1/1 ...
Well, it's good to know that 100% of the time, you can remove
"--zzotinting" from a word and it will still be a word. But this
inference is supported by just one
35
- fun most probable paths { lower_bound : real, state : state }
{ string : symbol list, p : real } stream =
let
fun nexts i =
case symbol from int i of
NONE => nil
- SOME sym => let
val p = (* probability of seeing sym in this state *)
in
if p \< lower bound then nexts (i + 1) else if sym = end symbol
then S.singleton { string = nil, p = p } :: nexts (i + 1) else
let
val lb' = lower bound / p
val tails =
most probable paths { lower bound = lb',
state = advance state (state, sym) }
in
(* Now multiply through the probabilities and add the symbol to the
head of the strings. *)Stream.map (fn { string = t, p = p' } =>
{ string = sym :: t, p = p * p' }) tails :: nexts (i + 1)
end
end
(* Try all next symbols. *)
val streams = nexts 0
in
S.merge sorted bysecond real descending streams end
Figure 9: Pseudocode for most probable paths. advance state gives a new
state from a previous state and symbol observed, so that for example
advance state(abc, z) gives bcz. The pseudocode for merge sorted is
given in Figure 10.
36
fun merge sorted cmp l =
let
fun ms nil () = Nil
- ms (s :: t) () = case force s of
Nil => ms t () | Cons (v, ss) =>
ms insert v [ss] t
and ms insert bv sg nil =
Cons (bv, delay (ms sg))
- ms insert bv sg (s :: t) = case force s of
Nil => ms insert bv sg t | Cons (v, ss) =>
case cmp (bv, v) of GREATER =>
ms insert v (singleton bv :: ss :: sg) t | => ms insert bv (s :: sg)
tin
delay (ms l)
end
Figure 10: Pseudocode for merge sorted. ms merges a sorted list, and ms
insert is a helper where we have a candidate best value bv which will
either be the one we return at the head of the stream, or we'll replace
it and then stick bv somewhere to be returned later. (This algorithm can
be improved by making a data structure like a (co)heap; this is just a
simple first pass.)
observation (the word mezzotinting); there are actually hundreds of
such unique prefixes and suffixes. We need a better list.16 Removing the
ones that appear just a single time doesn't really help that much:
Still bad. Let's turn up the juice to prefixes and suf-fixes that appear
at least 10 times.
1.000000000 -wrought 10/10
1.000000000 -writings 12/12
1.000000000 -wraps 10/10
1.000000000 -wrap 11/11
- The right thing to do here is probably to use binomial like-lihood
rather than the scale-independent fraction. But simpler approaches
produce pretty good lists.
1.000000000 -worms 69/69
1.000000000 -worm 69/69
1.000000000 -working 21/21
Much better! But the next step is going to be to try removing these
prefixes and suffixes from words that have them, to find new words. Since
these have mod-ularity of 100%, we already know that every time we apply
them, the result will already be a word. So they are useless for our
analysis. Here are the most modular prefixes and suffixes with modularity
strictly less than 1.
37
Wow, now we're talking! The single word that cannot have "--maker"
removed is comaker, suggesting that co should be word (noun: "What a
comaker makes.").
Given this list, the next step is to identify potential words that can
be backformed by removing prefixes or adding suffixes from existing words.
Such a string can often be found via multiple prefixes and suffixes. For
example, twing can be formed by removing "--ing" from twinging (a
false positive, since the root word is actually twinge in this case)
as well as by removing the prefix "lef--", which has modularity of 20%
(including splits such as "lef/tie"). Maybe not good justification, but
twing is a pretty good word anyway.
We define the probability of a word as its Markov probability (with n
= 4, as this seems to produce the best results), times the probability
that at least one of the potential backformation rules applies.17 Here
are the most likely words by backformation:
word prob most likely backformation rules
==============================================
dises .023% para- (0.42) fluori- (0.39)
melo- (0.35) bran- (0.31)
tring .020% hams- (0.36) scep- (0.35)
bows- (0.33) hearts- (0.29)
disms .017% triba- (0.31) drui- (0.30)
bar- (0.27) invali- (0.27)
- As above we only allow backformation rules that have at least 10
occurrences, to prevent degeneracy.
I think that this approach shows promise, but there appear to be a few
problems: Many of these "rules" can be explained by bad segmentation
("heads--" appearing to be modular, for example, is really just "head--"
plus "s" being a common letter.) Second, I believe the dis-junctive
probability of any rule applying is too naive for determining the score.
For example, tions has al-most a thousand different prefixes that could
apply to it; the chance of any one of them applying is very nearly
-
But this is actually because "tions" is just a com-mon way for a
word to end. Legitimate root words to which many good prefixes are
applied cannot be easily distinguished from common suffixes by this
symmetric algorithm. More work is called for here. -
Survey
On occasion I have been accused of "overthinking" prob-lems, whatever
that means. So to compare, I next haz-arded a tried and true technique
from grade school, the survey.
I asked a few people who happened to be around, "What word ought to
exist?" Most people did not know what to make of this question, and
also, because people seem to revel in the opportunity to get (well
deserved) revenge on me by being disruptive trolls, many of the answers
were designed to be unusable. In order to not reprint everyone's
bullshit---but not introduce bias by selectively removing data---I
discarded random subsets of the data until it did not contain bullshit
any more.
From this we can conclude that 16% of people wish nurm were a word,
and so on. These words did not come with definitions, except for
gruntle, which Jessica gives
38
as "the opposite of disgruntle". This is actually already a word, but it
was the inspiration for Section 4. etsy is the name of a popular
on-line crafts community so I don't know why Rob would suggest that. The
meaning of wafflucinations is clear from morphological analysis.
- Conclusion
In this paper I investigated several different ways of an-swering the
question: What words ought to exist? Each method produces different
words, and some don't work that well, but nonetheless we have several
rich sources of words, each time scientifically justified. I conclude
with a section of recommendations for words that ought to exist, along
with definitions.
6.1 Recommendations
Sweeeeeeeeeeeeeeeeeeet with 19 *e*s is the clear favorite based on
analysis of usage, so this one should be intro-duced. It means "Really
sweet."
Rane sounds too much like rain, but sare has a unique
pronunciation and many people seem to think it's al-ready a word. I
propose that sare be introduced as a noun meaning, "a word that sounds
real but isn't."
Cho is similarly easy to pronounce and spell. I pro-pose that it be
defined as "A kind of cheese," so that we can really nail the new triple
entendre on the clas-sic joke. Chomaker is someone who makes that kind
of cheese.
Unders was one of the most frequently occuring words towards the top
of many analyses. This word should be a colloquialism for underwear,
which would probably already be understood from context.
Dise is suggested by both the Markov model (as dises, dising) and
backformation (as dises). I like thinking of it as being the root of
paradise, where para-- means something like "along side of" or
"resembling". So dise is the place you're really looking for when you
get to paradise and realize it's just a mediocre country club.
Helafelo is one hell of a fellow.
Addendum. During the preparation of this paper, the Scrallbe game has
converged on a culture where the words played are real-seeming, with
creative definitions. Examples: frodeo ("Gandalf is the clown.")
pridefax ("An unproven treatment for telephone anxiety.") eeeee
("eeeee") orzigato ("Move mr. roboto. for great jus-tice.") stovebed
("Now you don't have to get out from under the covers to make
breakfast.") ovawiki ("The
free egg cell that anyone can edit.") gaptave ("Two discontinuous
musical intervals.") achoolane ("Nostril (colloq.)") gplerious
("Completely overcome by soft-ware licenses.) bestcano ("Some
eruptions are better than others.") Thanks to the players for their
contribu-tions, especially Ben Blum, Chrisamaphone, Rob Sim-mons, and
Flammin Fingers.
Appendix
Here are the most likely words induced by the English Wikipedia, with
n = 3. I have left off the probabilities; they can be reproduced by
downloading Wikipedia and running the software yourself, which only
takes like 9 hours.
s ther w en t c d f m b ser e diff u sup p n de tal othe th wikipedia rd
con id r ame edia cal uk wiki und als st l km wp g x ext j des td ter h
co k sout don o cound v ver wer med res ind la em re ins afd ands div
htm infor rom wher ple mi acces ii ent unive ff refere ger bution lish
ince wikiped auth dist oth smally rea al y est cle fr lation tral gov nd
mily isbn cond bel be-for hould ber pers inder les gree mation lar chan
dia ove stary compan mon val san cate nothe pdf tes ster nort pril hist
sq clast ful ment q aft tr nown lign del z ave dify stor sity artic
pring rese mas li ord ca thes oldid lin es ress flage blace pland mar
mes conce ex-ter wate mp jame lity ft cology spany mic ve han thist spam
cand sher smal subdiv tem jan ture inded bothe ared cour rect thers ope
mm dom der mand tt dr phy cd websity stude lor bbc ru mer flor succes
ence tely aust stan clas et outh pl ints vfd fc wikipedit sus for-man
contry au edu inding el lic bord mr ston whic ide hous mich auser google
pt le doi engle ed gen userved cols num nate cer withis clude shous
inted wikipedian thered andex chool ning attle chael sta plation ampion
crea fontry lege prese grough iii mader augh ll ching fb furt yearch
examp octor nove furth bein tra nover tex-ter mont sted gener quary
inclub ander sco ast gr wome ality lon hough ano exas polid sease dec
sor ga nov unty shor imdb offic partic oved forn lan fm undex rus arted
alia cong vill eason gue shough cc vil tember octory jr gove writy thand
anding rev colog gover edian publist mics comple syster olding coor
useum spect expan pu-ble bate tribut oned coibot betwo conter cles
leason fina freen eng fa que fied tre hamp guage wered nic awar yout
usic cgi andy nrhp clar arting bc cated rober feb ames ven jun thould
defere artics sen frand cred ada cana rec-tion modiff sal los ret ched
decial collow notalk nl yor
39
elete sain ance cens chand apr delect dea stant bruary commer righ ch
tor peral andia arth roup founty hor scies alth pete footnot html ath
tribs ac chas eith pa tl refor bes bor cently fonter neer tration reat
hi awa chris da wikipedits atter pp ents dvd arge wan sch sr fere inds
norts tha nament anot oper aspx rence website minal brity dts apring
goalso alber rown graphy anded min alian ength mility abbr talign bec
com authe hite pic pation jone ords demy alt wated alife totalk lond
tage ad ana evers amg gened colo engton mina lates porth fff lifor
footbally stry shing hase lis eletion rol mally inton ded doesn ny mus
pro du pre somes dired na becamp ire inces cologo mong swith acts aree
ang els noth unic mous dest ali deat smale poing dc ission iv oct posity
pred poss kore anda ja ms gard coi frome afc nhl pub-lice mothe jers
maded qual geory roy vid pur studing ip ign fath cally coundex col
abouth twork threen img ess lishe lable reland cing chi thern dete ar
ving di eur wers il dio hl ando seast sation orge ble sep db classion cm
las conver evel ing popula uns polic vs frans ker ann phich ovember
alish cher therence prock ife feate user-vice indid unded powere hs phot
assion von dece theight proview ama prote kg birthe specian userve se
nois cfm ma len ause gle jp flagic differe counder rance swed lee tel ont
sould whics artion rans ric parly overy ves sa grea ans une tary ss jul
ne fami dever ange hms oright da-tion cription prothe univer lat tota un
adming aaa am-pan perfor athe seased pov becaust alson thed methe aff
spand inc sis stral refered endex xfd texamp valign sha dists fron heade
lood thest unived italk brary colled uc-first ite gan ep rary oney
wikipedition ra ex scord cition hout clos bon lp imple acade flagico
bost isn spection samp ral tead sant roces yorks ories bords unition che
dra epison eff stal abour ach earch busic reate conves ses hims mone fran
carly nated lished fra pm oh att delet unior reporth daving emple tworld
noved trand ross ie vall orld vi becauser caust dard exand feder ches
yound gence bol stribut rement gf sugges tity forder aug systed mathe
websited abover ltd recember colle wils mos sman lete dow eles throup ba
unt expr formany df ments tect pape hool exper butions hited delection
deces coung gon gers stors abc partics writion furthe rade mult bb lism
afries frence bish arence tems musion colore ethe bute vie ia hing oute
hom clain nom missue hus ns chich ox-for contribut vict autom boile
novers aus airst jos proble werence georgan rel eas opedia rs ned aut
iss pg ct othis indo shi dst alf ian grounty hile mor jor grap parth
elet regory thems ell septed pc url gnf unds appy studed offere gb timent
sed dary eved hally deate hr taly nom-
ing duri cented thead aliam mored pmid hered hong recore vious wates
fathe goved squary soment mout rela eview unives shere trate betweek gia
mory cas dal fi enge stribs colspan vol facted amer reser loc fination
deven andid leade tand protes gar dely til alled sary tro alist ng cha
charly ab hel begion mana sk powed ted calign wikipe dar nm churce mil
ren dise rever sel howere nexter commen sonal oly subser coibothe shool
hal froman eu waship pr rease ph rical op assed bility ording conces mal
stitle iucn sri os mond lations mance artment rigin rt brite ffff jsp lb
dd wikipeding lant sett arter te yource sh contribs footbal zh aution
acter ad-dit govers coln useries ract thich tain inten sec aa blo janual
hon nowing letion tition pute goldid communite wors fontribut areat
titute rected summent stational hu counded sp eld ples yeare nating fing
foll reque fiel pol josed atland phile manuary civision ok feded ult dj
sas alle cg buildid kee redia ap unk ko cr inver nomy sund toon reg aaaa
cath dll lc ters communive bal prope refef ania mrs hased worder pria
arry centing spon ade vand dor sourch tooks seph capan pbb colong sql
petion shors fift prily prial brition fl seption nz fonto cbs fally
whill losed cluded mations monal sland pict russion profilm coundo
reture harly soute sc rr titution therent shouser pager outhe quent
spons fn ited proces bable atte whicle sile si rece unitor yan grought
cos pian comput trans jource arthe tury specia artice provid raction
deleter alo fored ti eague wiking conth ree musly ars mf ups schood rey
donal sected writed freet formate cre co-lution opers chen bd hp sember
folled pagest smales wen socies cs engin aften chur delp bia ce nors
cource ussion fontribs summan arties jon dism lang sking au-gus theren
xs seve deball er uning truction nomic creat gian acco sv pera alls rict
relete rects coord stre ences aread spane meth brea ener thernal thire
whical rounty thu cl pola una thround uservation volved hund aga cames
menter froma uset publis frey libe posite northe ska appedia recial ro
fant hought mination geo presult andon occup ala nas thists ew maring
secore nee dian inders inves infore hower sume scal kans indust eding ka
aux brid var orded gived backgrough hugh ado prive eb rading joing hown
euro lim kan chus alifor wille deced conven blogy nbc quall coller
partion folle eleter reven reques lition humber moder backgroup
tribution flages lears forch reight pault evely mayork sition feath
recont guit metion hony deview nume confere fle unitory rican ima accore
deside ris pedia lisher acture aland userview thate polor kar nel sents
pany cluding vard bing woul havel greer thang genry effere locall desh
thors empt
40
sult quence mir blook porthe tran rew bording shought kely neight lam
stion publishe hors wis apan nony sourn nage orce profest jonal fonth
younty deta pd wa lo peries flign fource ros rfc mely mt smalled ge
retion relect vern alid aligh userves speer willa infl undon dels aftern
gres namedia advan distor oria subdiving lx presign blocal edway def
scho fic counds belf sily metry cread pi novery occes cribut holor reces
inte howed cout chis ency ac-tory lange cf nc bornia adven revie sames
newspan shan orial userry suff eth moto showere hunty nucle witz ider
mune flagicont aire shtm fri guary rigion mome afted heate mo formedia
oble cnn nea ha canage borning tol navbox ava octors md frent offical wal
arent anon artish cration rell dir auding ira butional labout bott
southe dence sporth formall ps intry thamp stube unus user-ing ron usert
conse brote lib oces bg proad tas wer-ent chese lyn atta pris oile drage
unre whenry ement stada chara raphy tection rica reletion foothe titler
las-sion areen theim conship formand ution unio ided guid riging brited
buch septem eleted conds preser ta ort feld ine actional alize tex fer
aused engly eque sevel atted smalle thome alty reath gation prefere hd
nr peration covember govember livision fren dq aren deve va ffcc jo prian
afl gion medy stategory nomich tured mel ene appa licy linal rought
areate bethe othey eral opport ara kw whild cance communited varing
brothe pia mu sched socian poin vis obs squal paking sma convers no-ing
maken cathe ni solor alry flagicons aaaaa weren arti-cal mtv flagion
chall alm upons dister searly sur towned oring ests avg whis publicy ec
thange gp ced pes whick figh peted fed sween discore sil ds sournal tay
dier kim adver reak mote wouldn paign inal medit pointo cata sper pts
idw mlb guis cational ul wasn septer aird throm fe feated collo itary
uni uple distory disa regist vere effor whican ober bettle anning cution
speciate palign ai cit moview flagicon af covery govery lina reast creen
texand exten vic
41
42
! \" # ( ) ) % ) ) ( * $ # $ % & \' (
& $ ) + \"- +,
\' ) & $ $ $ $
( * \'* $
) % 7 0* ) %
& \' ) * & $ *
\% % \' % *
& ( \' )) ) \' %
& $ $ $ $ $ * & \'
* * # &
( $ + $$ ( &
& \' & % & $ & $ $
( ; \< = , > = ,
( ) % $ %
( $ > + % $ ?/9 ( & * )
+ !, -\" +, ) & % & $ $ &
{width="3.161111111111111in"
height="2.3270833333333334in"}
$ * &
) \' ) ) & \' $ $
) ) % ) ) % $ %
& & \' ( & )
) $ )
* & \' $
. % \' # )
. % \' ( * \' ) )
&& ) ( $
(
,* $ $ & % ) )
* \' ) $ %%
* \' $$ ) ) ( & \' ) *
$ $ * $ *
& $$$ $( /$ \' $
+% ( $ \' $ )
& ) \' & % 8 $ ? ) \' & ) ( ; \< ( \'
& \' %% ( ) $ ) \'
0 ,
{width="2.0805555555555557in"
height="2.917361111111111in"}
# ( \' & $ & $ ) % *
& ( 1 ) $ ) ) $ * . $ ) \" - # & ) $ 23 % ) $ $$ 4 * %
( & # ( $ $
!0 -
, % ) ) $ % * & 5 ( * \' % ( $ ( ) ) 2% $ 64 & \' \' ) & &
) 1 ) 7
( ) % ) ) % &
$ ( %% % %%
{width="1.663888888888889in"
height="1.7638888888888888in"}
{width="1.3888888888888888e-2in"
height="1.4791666666666667in"}
43
{width="2.8569444444444443in"
height="2.4895833333333335in"}
8 $ A > $ ) & % & $ $ ) ( $ ( 5 ( *
( $ ) ( ( ( % ) & \' ) ) \' ) ) ( & % * ( \' ) * $ ) % ) & ) % $ &
{width="3.3444444444444446in"
height="2.5569444444444445in"}
8 $ 9 3 $ % . \' ) * % %% ) & * ) % 3 & $ ) & ) % $ &
,!> \"+ 0#
6 \< $ * \" 6
? + ) & ) ! % \< $ * \" & ( # & / &
@ B ( 8 ) > = $ $ $ $ $ C \"
A B * % \' )
& / \" % $ C 0 ( )
9 B8 $ ! & % ?: \" \" $ ( ) ) C \" & \" )
{width="0.8652777777777778in"
height="6.944444444444444e-3in"}
6 +% ) ( ( ) ) *
\' $ ) )
44
SIGBOVIK 2011: TRACK THREE
Sleazeball-Computer Interaction
- An Objection to "An Objection to "The
Box and Circles Plot"" . . . . . . . . . . . . . 47
Long T. Jongwood
Keywords: objections, objections to objections
- Good Friends Project: A Little Social
Experiment in Stalking . . . . . . . . . . . . . 49
Brian Becker, Pyry Matikainen, Heather Jones, Prasanna Velgapudi,
Michael Furlong, Brina Goyette, Umashankar Nagarajan, Joydeep Biswas,
Heather JusticeKeywords: Social Experiment, Stalking People, Mobile Technology
Misuse
- Who is the biggest douche in
Skymall? . . . . . . . . . . . . . . . . . . . . . . . . . 53
Dr. Tom Murphy VII, Ph.D.
Keywords: computational douchebaggery, as seen on tv, man who can
sleep in any seat(Continued on next page...)
45
- A Materials Based Approach to Hardware
Digital Rights Management . . . . . . . . . 59
Matt Mets, Matt Griffin, Marty McGuire
Keywords: DRM, Digital Rights Management, 3D Printing, Yogurt, Copy
Protection, PolyVinyl Alcohol
- The "Whos Yo Daddy?" Problem, Or: Solvin Concurren Axcess Issues
> with
Computers, Yo . . . . . . . . . . . . . . . . . . . . 65
Dr. Donna M. Malayeri, PhD, Heather Miller, PhD©, Esq., Herr Doctor
Doktor Klaus HallerKeywords: scala, uniqueness types, jerry springer, race conditions
46
An Objection to "An Objection to "The Box and Circles Plot" "
Dear sir or madame,
I wish to object on the strongest possible grounds to the previous
paper. While it does make an ad-mirable attempt at a sort of
pseudo-Victorian meta-humor, its style too often fluctuates to be truly
effective.
Indeed, it seems to be an unscrupulous attempt by an under-recognized
author (possibly writing under a pseudonym) to focus more attention on
an otherwise trivial subject. Perhaps this author is under the mistaken
assumption that any publicity is good publicity.
I can assure you, this is not the case.
Yours,
Lohn T. Jongwood
47
sigbovik
{width="1.1368055555555556in"
height="1.1368055555555556in"}
ach 2011
Plenary Program Committee Con fi dential Paper Reviews
Paper 3: An Objection to "An Objection to "The Box and Circles
Plot""
REVIEWER: Prudy Coldfish
OVERALL RATING: 1 (weak accept)
REVIEWER'S CONFIDENCE: 3 (high)
While I stand by my "An Objection to "The Box and Circles Plot"", I
appreciate the fi ne per - spective that Lohn's contention presents.
Verily, my colleague Lohn Jongwood provides a good analysis of tone and
humor. Further, confident as I am that my right - thinking arguments
will stand to the test of scrutiny, I welcome the invitation for this
journal's readership to view con - tent in a more critical light:
perhaps this may foster an attitude less tolerant of disingenuous
authors such as Mr. Longwood.
48
Good Friends Project: A Little Social Experiment in Stalking
Brian C. Becker, Pyry Matikainen, Heather Jones, Prasanna Velagapudi,
Michael Furlong, Brina Goyette, Heather Justice, Umashankar Nagarajan,
and Joydeep Biswas
Abstract---In this paper, we automatically poll the longitude and
latitude of a friend (let's call him Bob) who is publicly sharing his
location via Google Latitude on his smartphone. For 11 months, GPS
locations were logged at 15 minute intervals, yielding over 30,000
locations spanning multiple continents. Analysis of this data reveals
much about the activities of our friend Bob, satisfying cravings best
exemplified by gossips and tabloid readers.
I. INTRODUCTION
DANGEROUS some phrases are. For instance, consider with me for a moment
the implications of posting the
following sentence to the Internet: "As part of a little social
experiment, I'm making my location public via Google
Latitude." Does that not invoke the mental image of a serpentine wagon
circle of words that effectively read "stalk me" (see Figure 1)? No?
Perhaps it was just a fanciful agglomeration of too many giddy ideas
that led us to begin a stalking project, but we rationalize it as being
of great interest to sociologists and machine learningists.
I. BACKGROUND
Google Latitude is a web-service provided, shockingly enough, by Google.
It allows you to share your location by
{width="3.3965277777777776in"
height="3.4118055555555555in"}
Fig. 1. An apt photo depicting the authors' imaginative responses to
reading the phrase "As part of a little social experiment, I'm making
my location
{width="3.3958333333333335in"
height="0.40694444444444444in"}
public via Google Latitude."
Fig. 2. Screenshot of Google Latitude interface, showing the location of
the primary author, with error bounds.
{width="3.3965277777777776in"
height="2.311111111111111in"}
running a service on your smartphone that periodically uploads your
current GPS location. Others can then see your location on Google Maps
(see Figure 2). Generally, the service restricts the visibility of your
location to friends, but Google Latitude will also let you share it
publically to all.
Motiving our project is a friend, let's call him Bob, who not only made
his location public but posted it to his website with a notice that it
was for a social experiment. Unable to resist the temptation to use this
data for evil, the authors of this paper immediately began a stalking
campaign and christened it "Good Friends Project." To the best of our
knowledge, this project has been kept a secret from Bob and represents
the first of its kind, making it state-of-the-art.
I. METHODS
We reversed engineered Google Latitude and wrote a Python script + cron
job to save the latitude, longitude, and time to a file at 15 minute
intervals.
IV. RESULTS
For 11 months between April 6th, 2010 and February 25th, 2011 our
system logged ~30,000 of Bob's GPS locations. In this section, we
analyze this data to reveal trends and other information of interest. Or
just generally stalk him.
A. Probability Density Heat Map
Using a Google Maps image of Pittsburgh, all GPS points were overlaid in
a Gaussian-blurred heat-map with logarithmic scaling to prevent
saturation. This represents the logarithmic probability density of Bob's
location at any given time (see Figure 3).
49
Zoo
Squirrel Hill
Hangouts
Swissvale
Station Friends
Square
Waterfront
Shopping
Fig. 3. Heat-map of the 30,000+ GPS coordinates overlaid onto a Google
Maps image of Pittsburgh and Carnegie Mellon University. The heat-map is
displayed logarithmically to avoid saturation. The map is annotated with
common Pittsburgh features.
From this heat map, we can infer quite a bit about Bob. From the two
points of highest density, we can assume that Bob works at Carnegie
Mellon University (or possibly a surrounding shop such as Starbucks) and
lives in Squirrel Hill. He spends a lot of time on Craig Street and in
Oakland, which correlates well to the theory that he is a CMU student.
He frequents the Waterfront and visits the Southside periodically. A
number of friends he visits live in Shadyside,
Fig. 4. Probability of Bob being at home, CMU, or somewhere else (other)
as a function of the hour of the day.
Swissvale, East Liberty, and Bloomfield. Also, it appears he has visited
the Pittsburgh Zoo & Aquarium at least once.
B. Time Distribution among CMU, Home, etc.
If Bob is indeed a CMU student, we can analyze his location by hour to
discover when he comes into campus, when he leaves for home, and when he
spends time off campus. As seen in Figure 4, Bob is something of a night
owl. The best chance of finding him on campus is at midnight
(surprisingly!), and he is most probably home around 8 am. There is an
interesting anomaly at 7 pm where he goes off campus but does not go
home. Our best guess is that he goes off campus for supper, probably to
Craig Street or Oakland as evidenced by their heavy concentration on the
heat-map in Figure 3.
Taking a longer view of things, we can analyze Bob's patterns by month.
Figure 5 shows the allocation of time per category. We can see that Bob
worked heavily during January, February, and October, spending more time
at CMU than at home. July and August seemed to be more laid back
vacation months, or perhaps Bob just spent more time working in cool
summer places around town. A large portion of December was spent at
home, although this could be skewed if Google Latitude was disabled.
Overall, Bob spends 43% at home, 29% on campus, and 27% elsewhere.
50
LONG DISTANCE TRIPS
The dates and locations of long distance trips Bob has taken.
C. Long Distance Trips
So far, we have focused on Bob's residential movements around Pittsburgh
and their distribution between key locales. However, by examining the
large excursions from Pittsburgh, we can gain a broader understanding of
his life. Table I lists trips outside the Pittsburgh area, which even
includes an international trip to Taipei. In total, our friend Bob
traveled ~37,000 km (23,000 mi), most of which is accounted for by
these long distance trips.
V. DISCUSSION
We have presented a novel way to track people who purposefully enable
Google Latitude. This method also works with those who may find Google
Latitude "accidentally" enabled after lending their smartphone to a
friend to check email. All that is required is a computer connected to
the Internet to periodically poll Google Latitude for the GPS locations.
Admittedly, analysis of such resulting data is somewhat tedious and time
consuming, so future work of this paper will focus on automated ways to
stalk your friends. Alternatively we could find ways to pawn such
analysis off on poor unsuspecting undergrads. Another possible avenue of
research is discovering an automated method to remotely enable Google
Latitude without the operator's express knowledge.
ACKNOWLEDGMENT
We would like give a special thanks to Bob for publically providing his
location and allowing us the wonderful opportunity to stalk him. We hope
this tiny taste of an analysis whets his appetite for further social
experiments.
REFERENCES
Available upon request.
Fig. 5. Probability of Bob being at home, CMU, or somewhere else (other)
as a function of the month.
{width="3.2743055555555554in"
height="2.6152777777777776in"}
51
sigbovik
{width="1.1368055555555556in"
height="1.1368055555555556in"}
ach 2011
Plenary Program Committee Con fi dential Paper Reviews
Paper 14: Good Friends Project: A Little Social Experiment in
Stalking
PC MEMBER: Laura Berg
OVERALL RATING: -3 (strong reject)
REVIEWER'S CONFIDENCE: 1 (low)
A little too "Good" of Friends project
If someone ever thought they may want to consider getting a restraining
order, now maybe be the time. In fact, with the new Google Latitude, it
might even become fashionable. While it is interesting to consider that
someone may use this tool to study how they spend THEIR OWN time, to
make your information publicly open to stalkers was not the brightest
idea. What this paper failed to mention was the future homicide they
were planning by tracking where this individual "Bob" went every single
day. I'm also curious, why photos, both satellite and from behind that
bush on the corner of his street, were not included given that he was
being stalked. Like a zebra. mmmmmm zebra burger.
So I would like to conclude that through "Bob's" worldly travels, both
in Pittsburgh and Out, he has yet to devour a zebra burger because he is
the zebra burger. He is being stalked. Consider a restraining order
"Bob." Bob? ...Bob?...oh no...
52
Who is the biggest douche in Skymall?
Dr. Tom Murphy VII Ph.D.∗
1 April 2011
Abstract
Did you ever notice how there are so many douches in the Skymall
catalog? This paper investigates the 22 males pictured in the January
2011 issue, using Internet technology to determine their douchiness.
We then present an efficient image recognition algorithm that reliably
predicts douchiness from photos.
Keywords: computational douchebaggery, as seen on tv, man who can sleep
in any seat
Introduction
In that Luddite void between closing the cabin doors and the beep
indicating it is now safe to use approved electronic devices, there is
one perfect pleasure: The Skymall catalog. It has everything, including:
Pointlessly impractical products you cringe at just imagining someone
receiving as unwanted gifts at Christmas, copy that preys on the
insecurities of business travelers, typo and physically impossible
hyperbole treasure hunts galore, Photoshop disasters, new friends, and
old familiar faces. But since 1990, science has wondered: Who is the
biggest douche in Skymall?
It is difficult to assess the absolute douchiness (say, on a scale from 1
to 10) of a given person. So, in order to answer this question, we used
Internet Technology to conduct a series of more-douche--less-douche
battles between randomly selected pairs of participants. The visitor is
simply asked: Who is the bigger douche? The proportion of battles won,
overall, is the final douche score.
After thousands of battles waged, we converged on the following
results, ranked in descending douchiness:
∗*Copyright c 2011 the Regents of the Wikiplia Foundation. Appears in
SIGBOVIK 2011 with the blessing of the Association for Computational
Heresy; *IEEEEEE! press, Verlag-Verlag volume no. 0x40-2A. 0.00
53
Results
{width="1.0159722222222223in"
height="1.7909722222222222in"}
In Charge of the Music won 207/66; 75.82% douche
{width="1.6930555555555555in"
height="1.8756944444444446in"}
Reclining Numbercruncher won 193/80; 70.69% douche
{width="1.6930555555555555in"
height="1.448611111111111in"}
Traveling Salesman
won 171/102; 62.63% douche
The Thinker
won 201/73; 73.35% douche
{width="1.6930555555555555in"
height="1.2493055555555554in"}
Karate Genius
won 188/85; 68.86% douche
{width="1.6930555555555555in"
height="1.8291666666666666in"}
Beauty Rest
won 167/106; 61.17% douche
Seaside Date
won 197/76; 72.16% douche
{width="1.6930555555555555in"
height="1.2659722222222223in"}
Poolside Bibliophile won 177/95; 65.07% douche
{width="1.6930555555555555in"
height="1.94375in"}
Jeweler
won 167/106; 61.17% douche
54
{width="1.6930555555555555in"
height="1.7534722222222223in"}
I Have to Take This
won 161/113; 58.75% douche
{width="1.6930555555555555in"
height="2.1791666666666667in"}
Treatment Recipent
won 121/152; 44.32% douche
{width="1.6930555555555555in"
height="2.3208333333333333in"}
Sass Moulavi, M.D.
won 117/155; 43.01% douche
Cured Snorer
{width="1.6930555555555555in"
height="1.5472222222222223in"}
won 141/132; 51.64% douche
{width="1.2416666666666667in"
height="1.6916666666666667in"}
Business Expert
won 118/155; 43.22% douche
{width="1.2416666666666667in"
height="1.9013888888888888in"}
New Haircut
won 96/177; 35.16% douche
Man who can Sleep in Any Seat won 131/141; 48.16% douche
Confident in Glasses won 118/155; 43.22% douche
{width="1.0159722222222223in"
height="2.1902777777777778in"}
Woodsy Gentleman won 86/186; 31.61% douche
55
{width="5.941666666666666in"
height="1.804861111111111in"}
{width="1.0159722222222223in"
height="2.4652777777777777in"}
Silver Medalist
won 34/238; 12.50% douche
Although many people disagreed on the finer points of what constitutes a
douche, the results were fairly significant, in that there were many
participants that were consistently perceived as more or less douchey.
This is truly a victory for the scientific method. Speaking of
victories:
Douche recognition
Having collected consensus on what constitutes a douche, we next turn to
the problem of determining whether any given person is a douche, even if
that person has not participated in hundreds of rounds of douche-battle.
Since people appear to be able to make decisions based mainly on images,
we look to image processing techniques.
Images are formed using "pixels", which are like tiny individual color
dots. Each color dot, or "pixel", is saved in a file. It turns out that
these files are all the same one: Every "pixel" is in a file, which
constitutes the series of color dots, as a series of bytes or "1s and
0s" 2, which constitute the digital information that is the file, or
"pixels." Point is, in order for a computer to "see" a file, all it
needs to do is look at it, by rubbing the files and "pixels" on its
CPUs, the same way that you or I look at a picture by rubbing it on our
eyeballs.
The problem with most image recognition algorithms is that they do not
work, and are also slow.
We eschew the traditional model-based approaches, instead using efficient
hashing algorithms such as SHA-1 and MD5. These operate directly on the
"1s and 0s" of the image, and run in linear time. This way, even if the
algorithms do not work, they will at least be fast.
We find that the algorithm SHA-1 correlates with the douchiness of the
image, but not well. The MD5
56
{width="6.1194444444444445in"
height="4.3909722222222225in"}
Figure 1: Performance of different prediction algorithms. SHA-1
correlates with the douchiness of the image, though it tends to
overestimate the douchiness of medium-low douches. MD5 actually has
negative correlation. A tuned variant of MD5 with the initialization
vector A8F82303 08A1B76B AA25DA9E 4C2C1883 correlates quite neatly.
algorithm is faster but in fact correlates negatively. However, by fine
tuning the initialization vector, we are able to produce a variant that
is just as fast and correlates very well with the data (Figure 1).
Conclusion
Do you disagree with these data (Y/n)? If (Y), then Science never
Sleeps! http://snoot.org/toys/wuss/skymall/
Poolside Bibliography
Sorry, I didn't read any papers or anything or do actual science.
57
{width="0.21944444444444444in"
height="0.6576388888888889in"}
Figure 2: Image files consist of "ones" or "zeroes", which is digital
information pixels (pictured).
58
59
{width="4.1666666666666664e-2in"
height="8.472222222222223e-2in"}
(C2H4O)x + H20 → goopaq
{width="8.819444444444445e-2in"
height="9.513888888888888e-2in"}
60
{width="5.138888888888889e-2in"
height="8.819444444444445e-2in"}
61
62
63
64
The "Who's Yo Daddy?" Problem
Or: Solvin' Concurren' Axcess Issues wit Computers, Yo
Dr. Donna M. Malayeri, PhD Heather Miller, PhD©, Esq., Herr Doctor
Doktor KlausPPPoE, P2P Haller
Abstract
Computer Science research can solve many real-world problems. Here we
describe our novel research, Unique-ness Types, and how it applies both
to multi-threaded pro-grams and to real world scenarios. In particular,
we solve issues that commonly arise in popular daytime sociologi-cal
science documentaries.1
1. Introduction
In multi-threaded programs resources such as memory and files are at the
same time accessed, i.e., concurrently. This often leads to problems,
such as blue screens, not-disappearing hour glasses, the spinning dying
color wheel of death and so forth. For example, let's consider a thread
program that a file opens, and it accesses. It may happen that yet
another threaded process at some later point in time closes the file
without the first program knowing about it. If the first program to the
file again accesses, then the user may a blue screen experience.
2. Uniqueness Types
To solve the unique problems introduced in the introduc-tion, we
introduce a novel language-theoretic type theory called Uniqueness Types
[2]. We base our theoretical the-ory on flow-sensitive linear logic
with typestate [1]. The idea of our research approach is to assign
unique types to pointer variables. A pointer is unique whenever the
com-piler program knows that all other pointers are pointing to-wards
other datums, or, conversely, if and only if no other pointer is
pointing towards it. In predated works computer researchers have
published theoretical experiments with unique pointers that sometimes
their typestates change.
Unique pointers with uniqueness types give rise to a unique approach to
avoid the unique problems of hour glasses and spinning dying color wheel
of death. Somehow we can make sure everyone points to the hour glass or
something like that. Or no, maybe the thread program must have pointers
with unique names.2
- E.g., The Jerry Springer Show.
<!-- -->
-
Americanadian translation: if two threads access the same resource,
say a file, it would be bad if one thread opened the file, then the
other -
Real-World Scenarios
As interesting as these programming problems are, feel we that it time
is to computer science to real world problems apply. How else can we,
with a straight face, to funding agencies the claim make that we real
problems that affect people's everyday lives solve?
We believe that a good source of real-world problems documentaries is,
particularly those highly-rated ones that on broadcast television are
shown. These informational programs an unprecedented view into the daily
life of the everyman provide. For the purposes of this paper, The Jerry
Springer Show as our primary source we shall use, though our solution
to scenarios seen on other esteemed programs is applicable, such as The
Maury Povich Show or Jersey Shore.
3.1 Real-World Problem Statement
So, this one bitch is a real ho fo real and she get wit three playaz,
Playa 1, Playa II, and Playa Playaa Playa. And now showty pregnan' and
she sez the baby daddy is Playa Playaa Playa and he a pimp.3 He sho'
Playa 1 is da baby daddy fo serious and he don wants to pay no child
suppo'. Da ho gots a paternity test dat sez da baby daddy is Play
Playaa Playa but he thank it a fake.
Here, the problem is access to the shared resource within a critical
timeframe. Since the third man does not trust the results of the
paternity test, believing the woman to be a "lyin' ho," we need to
provide the parties in this sce-nario with a fool-proof mechanism for
determining parent-age.
We believe this problem is widespread, and in the tradition of clever
monikers for computer science prob-lems (e.g., Travelling Salesman,
Sleeping Barber, Din-
thread closed the file, then the first thread tried to then read from
the file. Essentially, threads with pointers to a shared resource need
to know about state changes that may affect future operations on that
resource. A uniqueness type solves this problem, and you can read more
about it in this boring---er, exciting---research paper [2].
- This means he has lots of money.
65
ing Philosophers), we name this problem "Who's Your Daddy?"
Further, this is a problem that clearly arises often in practice. As
evidence, see Figures 1, 2, and 3.
{width="2.8895833333333334in"
height="5.0680555555555555in"}
Figure 1. Two male kittens access a shared female kitten resource,
creating contention. (No, this is not a gratuitous kitten picture.)
4. Applicability to Real-World Scenarios
It is not immediately obvious how the theoretical results of Uniqueness
Types to Who's Your Daddy can be applied. We present here an iterative
approach to a solution, starting with a na¨ıve solution and refining it
to handle all possible scenarios.
The problem here is that several pointers the shared re-source may
access, and her state may at any time change, without it being obvious
to any of the parties involved (in-cluding the resource herself) that a)
the state change has occurred and b) which pointer caused the state
change.
Figure 2. Unexpected physical confrontation of two threads who accessed
a female within the same timeframe.
{width="3.2104166666666667in"
height="2.558333333333333in"}
Figure 3. The shared resource unsuccessfully attempts to arbitrate
between threads who are engaged in a violent altercation. As seen on
Jersey Shore.
We introduce the following protocol:
-
An external party outfits the resource with a he¨avy¨ m¨et¨al¨
> locking device with a physical key. For instance, the Victorian
> chastity belt would be quite useful here (Fig. 4). -
The physical key is retained by the external party for an incubation
> period of not less than 9 months. -
The first pointer retrieves the key from the external party and may
> access the resource. -
When the first pointer is finished with the resource, he or she
> returns the key to the external party. -
The incubation period is again started and the process repeats
> itself.
66
{width="2.407638888888889in"
height="1.926388888888889in"}
Figure 4. An antique he¨avy¨ m¨et¨al¨ locking device. Note that this
version lacks many of the features required in our solution, and is
therefore far inferior to our proposed locking device.
However, this na¨ıve approach is fraught with issues. The pointer who
has the key in his or her possession may initiate an ownership transfer,
either wanted or unwanted (i.e., in cases of theft). Moreover, borrowing
re-introduces the same race condition that the protocol had attempted to
eliminate. Thus, anyone with a copy of the key to the he¨avy¨ m¨et¨al¨
locking device may be the future Baby Daddy, which again leaves the
question unresolved: Who's Your Daddy?
This solution is similar to the first, except the pointer is the access
key. However, this means that once a particular pointer has accessed the
resource, she may henceforth never be accessed by any other pointer.
This is problematic if the first pointer gets bored and leaves, or is
otherwise destroyed.
¨
4.3 Solution 3: He¨avy¨ Met¨al¨ Locking Device with Dynamic Biometric
Key, aka She Crimped A Ladysman AndShit (SCALA)
This is similar to Solution 2, except now the locking de-vice can be
re-keyed by the external party to any particular pointer, assuming that
the resource is in the vacant state (which can always be determined once
the incubation pe-riod has passed).
4.3.1 Translation to Lay Speak
Now, dis shit gettin' futuristic and shit yo. Now, to get wit dis bitch,
you gots to go to her mama and daddy and get permission and shit. If she
ain't wit nobody and she ain't pregan' then they crimp yo junk in
computer shit an make it so no other playa get in dem locked panties and
shit. Na'i mean? That shit mothafuckin sucks, brace yoself foo' cuz dat
shit mothafuckin stings an dey know it was you when shit happen.
4.4 SCALA Solves Everything
The solution is clearly SCALA4 with Uniqueness Types. We draw here on
previous work which also considered the urban implications of SCALA
[3].
- Conclusions and Future Work
We have shown that research that problems in multi-threaded programs
solves can also to serious real-world scenarios be applied. With a minor
modification to existing technology (the metal locking device), have we
demon-strated how to solve a wide array of problems in popular
socio-scientific documentaries seen. In particular, with our solution,
can we always definitively that key question an-swer, "Who's Yo Daddy?"
Yo so this is the mothafuckin straight deal yo. Dey take da shit dey do
on computers and shit and fuckin make it so yo badonkadonkical ho can't
gets with no otha balla. So when she sez you da baby daddy you jus gotta
say to the otha playa, "Drop yo draws boy I know yo junk be crimped."
- Acknowledgements
Herr Doctor Doktor Klaus provided all of the middle-English-like
sentences. The other authors thank him highly for this immense
contribution, which would not be possi-ble without a native German
speaker. Ebonics translation gracefully provided by Frau Hezza (aka
Hedair). Thanks greatly to the fucking wild input from Dr. Jenn B. S.
References
-
Bierhoff, Kevin, Nels E. Beckman and Jonathan Aldrich. Practical API
> Protocol Checking with Access Permissions. In 23rd European
> Conference on Object-Oriented Program-ming (ECOOP), 2009. -
Haller, Philipp and Martin Odersky. Capabilities for Uniqueness and
> Borrowing. In 24th European Conference on Object-Oriented
> Programming (ECOOP), 2010. -
PhD, Malayeri, Donna Dr. "Pimp My Library" Is An Affront To Pimpers
> Of The World, Everywhere. In Proceedings of SIGBOVK 2010,
> Pittsburgh, PA, USA, 2010. -
She Crimped A Ladysman AndShit
67
68
SIGBOVIK 2011: TRACK FOUR
Military-Industrial Complexity Theory
- An Objection to "An Objection to "An Objection to "The Box and
> Circles
Plot""" . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Gen. Reginald F. Stout, (Mrs.)
Keywords: objections, objections to objections, objections to
objections to objections
- Going to the Mall Is Like, Totally
NP-Complete . . . . . . . . . . . . . . . . . . . . . 73
Brittany, Bethany, Tiffany
Keywords: malls, going to malls, purchasing items at malls, not
purchasing items at malls, mall food courts, food mall courts,
shoplifting, liftshopping, complexity theory, theoretical complexity,
sexual turing test
- Optimal Image Compression . . . . . . . . 75
James McCann, Notreally A. Coauthor
Keywords: optimal, image compression, optimality, images,
compression
- On unlexable programming languages 79
Robert J. Simmons
Keywords: perl, lexing, decidability
69
70
An Objection to " An Objection to "An Objection to "The Box and
Circles Plot" " "
To whom it may concern,
I strongly object to the previous letter. While pretending to critique
the self-promotional meta-humor of the preceding paper, it is -- itself
-- a sort of self-promotional meta-humor.
Besides, the notion of a chain of objections to objections is clearly
lifted from the excellent "Monty Python's Flying Circus" sketch program,
which -- to be honest -- did it better forty years ago. (In part because
their letters often ended with a humorous author name or affiliation.)
Love,
Gen. Reginald F. Stout, (Mrs.)
71
sigbovik
{width="1.1368055555555556in"
height="1.1368055555555556in"}
ach 2011
Plenary Program Committee Con fi dential Paper Reviews
Paper 4: An Objection to "An Objection to "An Objection to "The Box
and Circles Plot"""
REVIEWER: Lohn Jongwood
OVERALL RATING: -2 (reject)
REVIEWER'S CONFIDENCE: 4 (expert)
This objection to the objection to the objection to the box and circles
plot should have little place in any publication. As any good
metatheoretician should know, meta - meta - humor is no humor at all,
and the previous letter made no attempts at such content.
Besides, if a chain of objections is to be so avoided on principle, as
Mrs. Stout indicates, on what grounds is her own objection any more
worthwhile?
72
Going to the Mall Is Like, Totally NP-Complete
Brittany (AIM: BiEberGrL21348)
Bethany (sparkles3389@hotmail.com)
Tiffany (myspace.com/twilightXtiff)∗
SIGBOVIK 2011
Abstract
You mean like, abstract art? Ugh I hate art class. The teacher is like
a million years old. I think she's like from Russia or something
wherever that is and she doesn't let us text in class!
- Introduction
Ok so like it's Saturday and this week has been a total bummer! Mrs.
Davidson said we had to read 50 pages this weekend and write a three
paragraph reaction! And we have to use punctuation and proper spelling?
Seriously isnt it like the 20th century or something? Who the hell still
rights like that? Then Britt's mom grounded her because she thought she
was sexting! I'm like mom, I was just in the bathroom taking a picture
for my facebook! And only grownups say sexting! God everyone sucks.
Tiff's dad took away her iPod because she was listening to that "Fuck
You" song! He was all like, this is a Christian household! Thats
probably why her parents only had sex once and it was probably like
before she was born.
- Proof
So we took Bethany's mom's minivan and we're going to the mall!! She's
the only one that has her permit, but we help watch the road while she's
texting. Anyway we get to the mall and we do some shopping. We kinda
follow this one cute boy but he goes into Gamestop and we're like EWW
NERD!!! Then we're in the food court and we have this realization. The
mall kicks ass because there are No Parents! And Britt just found the
perfect belt so now her outfit is Complete! I think we're supposed to
say "Q.E.D." but that's latino or something, whatever.
- Conclusion
Tiff just stole my junior prom date! omg shes such a bitch!!!!!!
∗ We needed someone to type this up so we paid some guy $5. He's a
huge dweeb but for another $5 we let him put his name here. Matt Sarnoff
(computers_are_l@me.com) What a loser! Omg who the hell still uses
email!!!
73
74
Optimal Image Compression
James McCann∗ Notrealy A. Coauthor
Adobe Systems, Inc. Noplace Much
{width="3.061111111111111in"
height="2.6638888888888888in"}
Figure 1: Our null space transformation theory cyclically commutes
on a hot orange-yellow bicycle, even when it's raining.
Abstract
Data compression, particularly of image data, is an important
application domain in computer science; much of the data transmitted on
data networks today consists of image data or their wiggly cousins,
video data. Thus, any data reduction in the data size of these data
items can be of immediate cost savings for data network and data server
operators and a time savings for data viewers.
However, current image compression literature is generally focused on
heuristics and approximations rather than on op-timality. This paper
rectifies that deficiency, providing an image compression algorithm that
is space- and time- opti-mal, both asymptotically and in practice.
CR Categories: C.d.b [Images]: Compression---Systems and results
E.s.a.b.z.b [Systems]: Practice---Image transmis-sion
- Introduction
Image compression has been one of the enabling tech-nologies of our
current networked lifestyle; without it, we wouldn't be able to share
pictures and video with our friends (and strangers who are our friends
on facebook for some rea-son). However, despite the importance of
compression algo-rithms, little progress has been made toward either
space- or time-optimal compression.
In this paper, we describe a compression scheme that is both space-
and time- optimal and is easy to describe and imple-ment. This scheme
is based on a theory of null space trans-formation, which we lay out
in detail before introducing the final compression operator.In order to evaluate our compression operator, we compare it to
various commonly-available compression schemes over a broad test set.
We find that -- as expected -- it performs favorably in
compression/decompression time as well as in resulting file size.
- Background
∗e-mail: jmccann@adobe.com [Microsoft Corporation 1998]
75
- Null Space Transformation
Our method is motivated by a theory of null space trans-formation,
assembled mostly from whole cloth and refined through smoke and mirrors.
In this section, we derive our
theory in three pairwise steps: null-space transformation, null
space-transformation, and null [space]{.underline} transformation.
These pairwise derivations are enough to demonstrate that our theory
cyclically HOB-commutes (Figure 1), which is isomorphic to validity by
assertion [Experts 2011].
3.1 Null-Space Transformation
We demonstrate the pairwise validity of our null-space trans-formation
by induction.
Theorem 3.1 (Pairwise Validity A). A first-second dashed subset
of our theory validates exactly.
Proof: We proceed by induction.
To see why our transformation works, first consider that in cosmetic
surgery the simplest form of transformation is removal; thus, removal
is the base case.Furthermore, the null-space of a linear operator is the space of
vectors that the operator takes to zero -- precisely those vectors
which are to be annulled. Of course, annulment (which is
synonymically tied to removal -- our premise) is often difficult
without specific dispensation; but such a dispensation can be obtained
by a skilled negotiator (a so-called smooth operator).As linear operators are smooth, we are done.
3.2 Null Space-Transformation
In order to validate the space-transformation form of our the-ory, we
are able to perform a transformation and then appeal to a prior result.
We term this form of proof "proof by get-ting someone else to do the
proof without realizing it", and plan to publish a series of
human-computation articles about it just as soon as we figure out how to
make other people write our papers for us without realizing it.
Theorem 3.2 (Pairwise Validity B). A second-third dashed subset
of our theory validates exclusively.
Proof: In order to properly demonstrate the futility of
space-transformation, we proceed by contradiction.Consider the existence of a space transformation. If such a thing did
exist, then it would also be known as a space warp. Vargomax
[2007] recently demonstrated that such warps can by characterized by
a ∞ × ∞ fantasy land and graph coloring, from which he concluded,"Welcome to warp zone!"
Therefore, as a consequence, of course, consequently, in fact,
consummately, it remains, indeed, ultimately, hence-forth, i.e.,
overtly, by contradiction, proceeding onward, entirely, unequivocally,
without equal, understandably, also, it appears, finally, thus.
{width="9.444444444444444e-2in"
height="9.375e-2in"}
Extra case: the authors have observed that, on occasion, a proved
theorem will become infected with agent α and re-vive. These zombie
theorems can be hard to dispatch again without a shotgun approach. Thus,
in the interest of the reader's safety, we provide an extra case of
ammunition:
For the readers' sake, we hope that -- when the infection comes -- these
shells either serve to enforce a lasting truth or a that counterexample
is close at hand.
3.3 Null [Space]{.underline} Transformation
We have saved the null [space]{.underline} transformation sub-case for
last because it is the most difficult to validate. Indeed, in order to
force our proof through, we will require two lemmas and a Slycan's
Lamask. We present the lemmas and their proofs presently. The Lamask,
however, is a shy creature, often found in swamplands, jungles, and
appendices.
Lemma 3.3 (Midwest Spacing). The Midwest has plenty of space.
Proof: The Midwest region of the United States has a tem-perate
climate, and is largely rural, with fields, plains, and woodland. As
such, it formed a wonderful habitat for the Common Wood Pigeon
(Crocodylus niloticus). Un-fortunately, due to the aristocracy's
demand for pidgeon-leather during the 19th century, the Common Wood
Pigeon was hunted nearly to extinction by Mexican fur traders
("Voyageurs") who transported the furs to England in their distinctive
canoes.This paucity of pigeons (or bird banishment) leaves the area with a
distinct deficit in the local species spectra, so -- by the pigeonhole
principle -- we are done.
{width="9.444444444444444e-2in"
height="9.375e-2in"}
Lemma 3.4 (Parking Spaces). It is always possible to find a
parking space nearby.
Proof: We proceed by generalization.
Many of those who read papers are of modest means.
Many people of this economic stature drive cars.
Therefore, by the principle of the undistributed middle-class, the
reader is either driving a car or not. Thus, the proof of this lemma
falls into cases:
- If the reader is not driving a car, then -- clearly -- the car has
> been parked nearby.
76
-
NST-Compress(i):
-
return ∅
<!-- -->
-
NST-Decompress(o):
-
return
{width="2.9993055555555554in"
height="2.611111111111111in"}
Figure 3: Pseudo-code implementing our compression and
decompression algorithms. For NST-Compress i is the input image.
For NST-Decompress o is an output buffer of the proper size
initialized to zeros.Table 1: Runtime and compressed image size for all known image
compression schemes, as compared to NST.Figure 2: A Gorey car crash.
- On the other hand, if the reader is driving an automo-bile, then
> they will likely crash (Figure 2); eliminat-ing themselves and,
> consequently, this case.
With these lemmas in hand, and the prospect of seeing a Lamask, we are
ready to tackle the main theorem.
Theorem 3.5 (Pairwise Validity C). A first-third dashed sub-set
of our theory validates ecstatically.
Proof: In order to demonstrate that underlying all spaces there are
no transformations, we proceed by extended bi-nary space partitioning.
Particularly, instead of using bi-nary, we use ASCII hexadecimal.For partition 0x6d696477657374, our Lemma 3.3 suf-fices; while for
0x7061726b696e67, our Lemma 3.4 is conclusive.Any other partition is -- with high-probability -- improperly spelled,
and thus can be dropped from the proof. Besides, Lamask watching is
far more interesting.
{width="9.444444444444444e-2in"
height="9.305555555555556e-2in"}
- Defining the compression and de-compression operator
As suggested by our theory, we define our compression op-erator as the
transformation that nulls the data-space of the image. We provide
pseudo-code for such an operator (and the associated best-fit
decompression) in Figure 3. This code is a straightforward consequence
of the theory, and so we do not need to explain it any further.
4.1 Reference Implementation
We provide source code of a reference implementation of our compressor
in Figure 5. We feel that while small improve-ments could perhaps be
gleaned by hand-optimization, over-all, our compression/decompression
speed is already quite good.To compress, run
./nst \< input-image > compressed
To decompress, prepare a blank image of the proper size, then run
./nst \< compressed >> blank-image
In either case, the utility will deduce the proper file for-mat before
performing the compression or inserting decom-pressed information into
the blank image. All image file for-mats are supported.
- Comparison to Other Methods
Our compression algorithm is theoretically both time and space
optimal, requiring zero work to compress or decom-press images, and
zero bytes to represent compressed im-ages. However, in order confirm
these theoretical results, we performed an exhaustive comparison of
all known compres-sion formats to our NST compression. We used a broad
test set of images (Figure 4).The results of this comparison are shown in Table 1, but can be
briefly summarized as follows:We, like, totally owned them, yo.
Or, more simply:
WLTOTY
77
{width="6.96875in"
height="0.9944444444444445in"}
Figure 4: Our broad test set. Images copyrighted by flickr users
piecesofalice and Nick Nunns, used under the
creative-commons-attribution license.
- Conclusions and Future Work
In this paper, we introduced and rigorously validated a the-ory of null
space transformation. Furthermore, we used this theory to develop a
provably optimal image compression and decompression scheme which far
exceeds the capabilities of other common compression methods.
Of course, this optimality is not without a slight cost. For many
images, our compression scheme (much like the popu-lar "JPEG" scheme) is
somewhat lossy. Despite this, we feel that the optimality of the
algorithm more than makes up for this slight deficiency.
C++:
int main() { return 0; }
sh:
!/bin/sh#
perl:
!/usr/bin/perl -w#
Figure 5: Reference C++, shell, and perl implementations of a
compression/decompression utility. Most Linux distri-butions already
ship a reference implementation under the name /bin/true*.*
References
EXPERTS. 2011. It's true. Trust us (Really).
MICROSOFT CORPORATION. 1998. Windows 98. Com-
puter Software (June).
VARGOMAX, V. V. 2007. Generalized super mario bros. is np-complete.
- There may be a Lamask nearby
Stay quiet and move slowly, for they are flighty creatures.
Eh? What does one look like? Well, I have not, myself, seen one;
however, I have conversed at length with a math-ematician -- quiet fool!
step carefully -- by the name of Jo-hansensumson who was working on a
theorem in the deep Cambodian jungles in 1925.
At that time, of course, the Lamask was just a legend, and
Jo-hansensumson was not entirely sure whether what he heard from the
locals was well-founded or mere conjecture.
So during his second expedition, he baited a succulent theo-rem -- just
as we have done here -- with a little bit of intuition and a few
hand-waves. He waited for hours, working on an-other line of reasoning,
when -- suddenly! -- it was upon him.
As quickly as it had arrived, it was gone, leaving only a few crumbs and
a pile of irrefutable logic. But he never forgot the sight. It was
twelve years later that Slycan conclusively documented the Lamask with
his charged-plate/oil-drop ex-periment and Johansensumson was vidicated.
- There it is!
{width="2.9993055555555554in"
height="2.9993055555555554in"}
(artist's conception)
- As soon as it was, it was not
The Lamask is a true wonder of mathematics; we are lucky to have seen
it, however briefly.It is just as Johansensumson described it, but also so much more.
Forgive me if I begin to weep; mathematicians seldom experience such
beauty outside dreams and books.
78
0O VOMFYBCMF QSPHSBNNJOH MBOHVBHFT
3PCFSU + 4JNNPOT
\"QSJM
\"CTUSBDU
0OF PG UIF GFBUVSFT PG UIF 1FSM QSPHSBNNJOH MBOHVBHF JT UIBU JU JT
GPSNBMMZ VOQBSTBCMF *O UIJT BSUJDMF XF DPOTJEFS UIF EFTJHO PG B
QSPHSBNNJOH MBOHVBHF UIBU JT TJNJMBSMZ VOMFYBCMF JO UIBU UIF DPSSFDU
TFQBSBUJPO PG B TUSJOH PG DIBSBDUFST JOUP MFYJDBM UPLFOT JT
VOEFDJEBCMF JO HFOFSBM
- *OUSPEVDUJPO
ɨF JOTQJSBUJPO GPS UIJT QBQFS XBT B GBDUVBM DMBJN BOE BO PQJOJPO CPUI PG
XIJDI * BUUSJCVUF XJUI TPNF VODFSUBJOMZ UP -BSSZ 8BMMhT 0DUPCFS MFDUVSF
BU $BSOFHJF .FMMPO 6OJWFSTJUZ #PUI DMBJNT IPXFWFS BSF BMTP EFmOJUJWFMZ
BUUSJCVUFE UP +FĊSFZ ,FHMFS ɨF GBDUVBM DMBJN JT UIBU JU 1FSM JT BO
VOQBSTBCMF MBOHVBHF \<,FHB> ɨF PQJOJPO JT UIBU UIJT JT B HPPE UIJOH
\<,FH> *G B MBOHVBHF XJUI VOEFDJEBCMF QBSTJOH JT HPPE B MBOHVBHF XJUI
VOEFDJEBCMF MFYJOH NVTU CF GSFBLJOH BXFTPNF )FODF UIJT QBQFS
- \"O VOEFDJEBCMF GBNJMZ PG MFYFST
ɨF SFDJQF GPS 1FSMhT VOQBSTBCJMJUZ JT SBUIFS TUSBJHIUGPSXBSE 1FSM JT
VOQBSTBCMF CFDBVTF ||UIF POMZ XBZ UP QBSTF 1FSM JT UP SVO JU PS UP
TJNVMBUF JU VTJOH B MBOHVBHF PG FRVJWBMFOU QPXFShh \<,FHC> 4QFDJmDBMMZ
UIF VTF PG \"1:AL CMPDLT DBO GPSDF DPEF FWBMVBUJPO EVSJOH UIF
DPNQJMBUJPO QIBTF BOE UIF 2pH GVODUJPO DBO USJHHFS DPNQJMBUJPO BU
SVOUJNF \<,FHB>
,FHMFS XIP FTUBCMJTIFE 1FSMhT GPSNBM VOQBSTBCJMJUZ XBT VOBNVTFE UIBU
||1FSMCBTIFSThh QJDLFE VQ IJT SFTVMU BT B DSJUJDJTN PG 1FSM \<,FH>
EFDMBSJOH UIBU 1FSMhT VOQBSTBCJMJUZ JT B HPPE UIJOH BOE UIBU JO GBDU
||EFNBOEJOH B QBSTFBCMF MBOHVBHF JT UIF TJHO PG XFBL QSPHSBNNFShh ɨF
VOEFSMZJOH USVUI PG UIJT TUBUFNFOU JT UIBU 1FSMhT VOQBSTBCJMJUZ NFBOT
UIBU 1FSM DBO IBWF OP NFBOJOHGVM OPUJPO PG BCTUSBDU TZOUBY
TP UIBU JU JT JNQPTTJCMF UP DPOUFNQMBUF TUBUJD BOBMZTFT GBDUPSJOH UPPMT
PS *%& GFFECBDL UIBU XPSLT JO HFOFSBM PO 1FSM QSPHSBNT TJODF TVDI UPPMT
VOJGPSNMZ XPSL PO UIF MFWFM PG BCTUSBDU TZOUBY 1FSM DBO POMZ CF
JOUFSQSFUFE OPU DPNQJMFE PS JO B DFSUBJO TFOTF SFBTPOFE BCPVU
)PX DBO XF BQQMZ TJNJMBS QSJODJQMFT UP UIF EFTJHO PG B QSPHSBNNJOH
MBOHVBHF XJUI VOEFDJEBCMF MFYJOH *O 1FSM JU JT UIF \"1:AL CMPDLT XIJDI
CPPUTUSBQ QBSTJOH JOUP QPUFOUJBMMZ QSPCMFNBUJD 5VSJOH UBSQJUT 8F EFmOF
UIF MFYJOH QSPCMFN UP CF UIF QSPDFTT PG UBLJOH B TUSFBN PG
DIBSBDUFST BOE VOBNCJHV PVTMZ SFUVSOJOH FJUIFS BO FSSPS PS B TJOHMF
MFYJDBM UPLFO BOE B TVCTUSFBN PG UIF PSJHJOBM TUSFBN -FYFST
79
-
<!-- -->
- D E
<!-- -->
-
G H I J
-
<!-- -->
-
<!-- -->
- N O
P Q
-
<!-- -->
- T
<!-- -->
-
V
W X Y
-
[
\"
# $ %
- \' ( )
<!-- -->
- + , -
. / 0
-
2
3
4
5
6
7
8
9
:
;
\'JHVSF 4UBOEBSE FODPEJOH PG MFUUFST BT CJUTUSJOHT GSPN UIF MJUFSBUVSF
\<4JN>
EFBM XJUI QPUFOUJBM BNCJHVJUZ CZ BMXBZT TFMFDUJOH UIF MPOHFTU QPTTJCMF
MFY UIJT JT IPX XF FOTVSF UIBU UIBU UIF $ UPLFO 2Hb2#Q# BMXBZT QBSTFT
BT B TJOHMF JEFOUJmFS BOE OPU BT UIF SFTFSWFE XPSE 2Hb2 GPMMPXFE CZ UIF
JEFOUJmFS #Q#
8F TQFDJGZ UIBU UIF MFYFS SFUVSOT B TVCTUSFBN UP FOTVSF UIBU MFYJOH DBO
POMZ SFBE GSPN UIF DIBSBDUFS TUSFBN OPU BEE UP JU *O UIJT QBQFS XF XJMM
BTTVNF UIBU UIF DIBSBDUFS TUSFBN BWBJMBCMF UP UIF MFYFS JT B CZUF
TFRVFODF XIJDI XF JOUVJUJWFMZ DPOOFDU UP UIF TUBOEBSE \"4$** FODPEJOH
PG DIBSBDUFST
ǰF %FD***n* **MBOHVBHFT
ɨF QBSBNFUSJ[FE %FD*n* GBNJMZ PG MBOHVBHFT XIFSF n JT B (¨PEFM
OVNCFSJOH PG UIF UISFF QBSBN FUFST BT ZFU VOTQFDJmFE BMM TIBSF UIF
GPMMPXJOH QSPQFSUJFT 5PLFOT TVDI BT BOE BSF MFYFE JOEJWJEVBMMZ CVU BO
JEFOUJmFS JT B TUSJOH PG BMQIBOVNFSJD DIBSBDUFST QMVT VOEFSTDPSFT TVDI
UIBU OP QSFmY DPSSFTQPOET UP B OPOUFSNJOBUJOH MBNCEB DBMDVMVT
FYQSFTTJPO ɨJT DPNCJOFE XJUI UIF SFRVJSFNFOU UIBU MFYFST SFUVSO UIF
MPOHFTU QPTTJCMF UPLFO JT UIF SFBTPO XIZ BOZ QFSGFDU %FD*n* MFYFS NVTU
TPMWF UIF IBMUJOH QSPCMFN \'PS JOTUBODF JO UIF %FD MBOHVBHF EFmOFE CFMPX
UIF BMQIBOVNFSJD TFRVFODF R[REG77[yRsRmby7QQ QBSTFT BT UXP UPLFOT
R[REG77[yRsRmby BOE 7QQ TJODF UIF GPSNFS DPSSFTQPOET UP B
OPOUFSNJOBUJOH FYQSFTTJPO
- QBSUJDVMBS NFNCFS PG UIF %FD*n* GBNJMZ JT EFmOFE CZ UISFF QBSBNFUFST
> ɨF mSTU JT B XBZ PG JO
UFSQSFUJOH B TFSJFT PG BMQIBOVNFSJD DIBSBDUFST BT B CJUTUSJOHT CZ HJWJOH
B CJUTUSJOH FODPEJOH GPS FBDI BMQIBOVNFSJD DIBSBDUFS JO UVSO ɨF TFDPOE
JT B XBZ PG FODPEJOH CJUTUSJOHT BT MBNCEB DBMDVMVT FY QSFTTJPOT ɨF UIJSE
JT BO FWBMVBUJPO TUSBUFHZ UP VTF UP BUUFNQU OPSNBMJ[BUJPO
80
%FD
ɨF %FD MBOHVBHF JT B TQFDJmD JOTUBOUJBUJPO PG UIF %FD*n* GBNJMZ
$IBSBDUFST BSF FODPEFE BT CJUTUSJOHT VTJOH B TUBOEBSE
DIBSBDUFSCZDIBSBDUFS FODPEJOH GSPN UIF MJUFSBUVSF \<4JN> FWFO UIPVHI
UIJT FODPE JOH EPFT IBWF UIF QPUFOUJBM EJTBEWBOUBHF UIBU NBOZ TFRVFODFT
PG BMQIBOVNFSJD DIBSBDUFST NBQ UP UIF TBNF CJUTUSJOH \'PS DPOWFOJFODF XF
SFQFBU UIJT FODPEJOH JO \'JHVSF
ɨF TFDPOE QJFDF GPS JOTUBOUJBUJOH %FD JT B XBZ PG FODPEJOH CJUTUSJOHT BT
MBNCEB FYQSFTTJPOT 8F VTF UIF FODPEJOH GSPN UIF +PU QSPHSBNNJOH MBOHVBHF
\<#BS> +PU BDUVBMMZ QSFTFOUT JUTFMG BT B GVMM QSPHSBNNJOH MBOHVBHF OPU
KVTU BO FODPEJOH PG MBNCEBDBMDVMVT UFSNT CVU JUT UFSNJOBUJPO CFIBWJPS JT
EFQFOEFOU PO UIF FWBMVBUJPO TUSBUFHZ PG UIF IPTU MBOHVBHF \<3FZ> ɨJT JT
OFWFSUIFMFTT QFSGFDU GPS PVS QVSQPTFT XF VTF +PU NFSFMZ BT POF QPUFOUJBM
FODPEJOH PG B TVCTFU PG UIF UFSNT PG UIF VOUZQFE MBNCEBDBMDVMVT %FD VTFT
B DBMMCZWBMVF FWBMVBUJPO TUSBUFHZ UP BUUFNQU UP OPSNBMJ[F UFSNT UP B
WBMVF
0OF VTFGVM BTQFDU PG UIF TUBOEBSE FODPEJOH PG BMQIBOVNFSJD TFRVFODFT JT
UIBU BOZ DPNCJOBUPS DBMDVMVT UFSN DBO CF FODPEFE EJSFDUMZ BT BO
BMQIBOVNFSJD TFRVFODF CZ XSJUJOH BQQMJDBUJPO QSFmY BT ||Bhh BOE UIF 4
BOE , DPNCJOBUPST BT ||Thh BOE ||Lhh SFTQFDUJWFMZ \'PS JOTUBODF UIF
UPLFO FbbF DBO CF FYUFOEFE CZ WJSUVF PG UIF GBDU UIBU (KS)(SK) JT B
UFSNJOBUJOH DPNCJOBUPS DBMDVMVT UFSN \<#BS>
- *NQMFNFOUBUJPO
*O \'JHVSF XF DBO TFF UIF TJHOBUVSF BMMPXJOH VT UP JOTUBOUJBUF BO
OFDFTTBSJMZ JODPNQMFUF MFYFS GPS BOZ %FD*n* MBOHVBHF ɨF QBSUJDVMBS
JNQMFNFOUBUJPO DBO POMZ QSPWF OPOUFSNJOBUJPO CZ EFUFDUJOH B DZDMF JO
FWBMVBUJPO \"MM EJWFSHFOU UFSNT XJMM GBJM UP CF OPUJDFE CZ UIF
JNQMFNFOUBUJPO BOE XJMM FJUIFS DBVTF UIF JNQMFNFOUBUJPO UP EJWFSHF JG
HBKBi JT LPL1 PS FMTF SBJTF BO FYDFQUJPO
ɨF JNQMFNFOUBUJPO JT BWBJMBCMF GSPN
?iiTb,ff#Bi#m+F2iXQ`;f`Q#bBKKQMbf/2+@M
- $PODMVTJPO
ɨJT QBQFS POMZ TDSBUDIFT UIF TVSGBDF PG VOEFDJEBCMF MFYJOH UFDIOJRVFT
&WFS TJODF $PIFOhT TFNJOBM XPSL PO UIF BSFB \<$PI> (¨PEFM FODPEJOHT
GPS QSPHSBNT IBWF SFDFJWFE JOTVċDJFOU BUUFOUJPO 0VS
+PU FODPEJOH JT UIF NPTU EJTBQQPJOUJOH BTQFDU PG UIF JNQMFNFOUBUJPO OPU
POMZ EPFT JU GBJM UP DBQUVSF BMM MBNCEB DBMDVMVT UFSNT NFSFMZ BMMPXJOH
UIF TJNVMBUJPO PG BMM DPNCJOBUPS DBMDVMVT UFSNT CVU UIF FODPEJOH JT
CPSJOH GSPN UIF QFSTQFDUJWF PG OPOUFSNJOBUJPO BMNPTU BMM FODPEJOHT
UFSNJOBUF \"T BO FYBNQMF UIF %FD*n* JNQMFNFOUBUJPO DBO MFY BMM PG
4IBLFTQFBSFhT 0UIFMMP EPXOMPBEFE GSPN 1SPKFDU (VUFOCFSH VOEFS UIF %FD
JOTUBOUJBUJPO XJUIPVU FWFS mOEJOH B OPOUFSNJOBUJOH UFSN
- *OWFTUJHBUJOH VMUSBMFYJOH XIFSF UIF MFYFS DBO NPEJGZ UIF DIBSBDUFS
> TUSFBN JT BO FYDJUJOH EJSFDUJPO GPS GVUVSF XPSL *OWFTUJHBUJOH
> VOEFDJEBCMF 65\' FODPEJOH UFDIOJRVFT JT BO FYDJUJOH EJSFDUJPO GPS
> GVUVSF XPSL
*OWFTUJHBUJOH OPODPNQPTJUJPOBM FODPEJOHT PG BMQIBOVNFSJD TUSJOHT BT
CJUTUSJOHT JT BO FYDJUJOH EJSFDUJPO GPS GVUVSF XPSL
81
bB;Mim`2 G1sn_:a 4 bB;
- JtBKmK MmK#2` bi2Tb iQ HQQF 7Q` +v+H2 Q` M B``2/m+#H2
> 2tT`2bbBQMX
LPL1 @ i?2 H2t2` rBHH MQi i2`KBMi2 QM /Bp2`;2Mi b2[m2M+2 aPJ1 M @
QMHv 7Q` M bi2Tb- i?2M `Bb2 M 2t+2TiBQM VpH HBKBi, BMi QTiBQM
- h?2 2M+Q/BM; Q7 ;Bp2M +?`+i2` b #BiHBbiX JQbi@bB;MB7B+Mi@#Bi
> 7B`bi5 V
pH +?`+Q/2, +?` @= #QQH HBbi QTiBQM
- h?2 2M+Q/BM; 7mM+iBQM i?i im`Mb #BiHBbib
BMiQ HK#/ +H+mHmb 2tT`2bbBQMbX V pH HK#/+Q/2, #QQH HBbi @= GK#/X2tT
- bi2T 7mM+iBQM 7Q` i?2 HK#/ +H+mHmbc LPL1 BKTHB2b i2`KBMiBQM V
pH bi2T, GK#/X2tT @= GK#/X2tT QTiBQM
2M/
\<,FHB> +FĊSFZ ,FHMFS 1FSM JT VOEFDJEBCMF DzF 1FSM 3FWJFX \<,FHC>
+FĊSFZ ,FHMFS 3JDFhT UIFPSFN DzF 1FSM 3FWJFX
\<,FH> +FĊSFZ ,FHMFS 6OQBSTFBCJMJUZ JT B HPPE UIJOH
?iiT,ffrrrXT2`HKQMFbXQ`;f \MQ/2nB/4dNyek9
\<3FZ> +PIO $ 3FZOPMET %FmOJUJPOBM JOUFSQSFUFST GPS IJHIFSPSEFS
QSPHSBNNJOH MBOHVBHFT *O
1SPDFFEJOHT PG UIF \"$. BOOVBM DPOGFSFODF QBHFT \"$.
\<4JN> 3PCFSU + 4JNNPOT 0O VOMFYBCMF QSPHSBNNJOH MBOHVBHFT *O
4*(#07*, QBHFT
\"$) \"QSJM
82
SIGBOVIK 2011: TRACK FIVE
Developers, Developers, Developers, Developers.
- An Objection to "An Objection to "An Objection to "An Objection
> to "The Box and Circles Plot"""" . . . . . . . . . . . . . . .
> 85
Keywords: objections, objections to objections, objections to
objections to objections, objections to objections to objections to
objections
- MLA-Style Programming . . . . . . . . . . . . 87
Richard Ha
Keywords: english, sml, techcomm, mla
- Theoremin: Proof by Handwaving . . . . 93
Ben Blum
Keywords: theremin, handwave, logic
(Continued on next page...)
83
- Quantitative Comparison of Typesetting
Software: TeXShop vs. TechShop . . . . 95
Matthew Sarnoff
Keywords: typesetting, Knuth-Liang algorithm, Knuth-Liang fanfiction,
user interface design, human-computer interaction, human-computer
attraction, human-computer courtship, sexual turing test
- Med School, CS Grad School, Both, or
Neither? . . . . . . . . . . . . . . . . . . . . . . . . . 99
Brian Hirshman
Keywords: submitting things that arent pdfs, what do you think this
is, chi?
84
An Objection to "An Objection to " An Objection to "An Objection to
"The Box and Circles Plot" " " "
SIGBOVIK Committee,
The previous objection, while accurate, did not contain a clear
punchline. Such a punchline is of-ten required to terminate a chain of
objections, as it can otherwise co.-Program received signal
SIGSEGV, Segmentation fault.
Maximum stack depth exceeded at objection.tex:12 (gdb)
85
sigbovik
{width="1.1368055555555556in"
height="1.1368055555555556in"}
ach 2011
Plenary Program Committee Con fi dential Paper Reviews
Paper 5: An Objection to "An Objection to "An Objection to "An
Objection to "The Box and Circles Plot""""
REVIEWER: Reginald F. Stout
OVERALL RATING: 0 (borderline paper)
REVIEWER'S CONFIDENCE: 1 (low)
I'm not sure why I was asked to review this letter, and I'm not really
good at this gdb thing, but I'll give it a go.
(gdb) backtrace
#0 in SIGSEGV_handler()
#1 in ???
#2 in objection() at objection.tex:12
#3 in objection() at objection.tex:9
#4 in objection() at objection.tex:9
#5 in objection() at objection.tex:9
#6 in box_and_circle()
#7 in SIGBOVIK_handler()
#8 in main()
(gdb) print punchline
$1 = \<value optimized out>
(gdb) punch printline
Violence isn't the answer to this one.
(gdb)
86
MLA-style Programming: Fostering More Interest in Programming in
English MajorsRichard Ha Echt C. Mmunicationo Mary Langdon Ashford
April 1, 2011
Abstract: In this paper, we address issues that plague the computer
science field that disinterest and disincentivize non-technical majors
in fields such as English literature from trying to learn a new language
that is not English or Latin. We investigates methods to provide new
platforms to leverage the synergies between the humani-ties and computer
science to promote better code readability and to get better grades in
Techni-cal Communications courses for Computer Sci-entists.
Keywords: english, sml, techcomm, mla, natu-ral language processing
in general. A variety of solutions have been pro-posed like tricking
children into learning lambda calculus or writing a module for Perl that
lets one code in Latin with varying degrees of suc-cess in their
implementations but none of them seemed to solve the overall problem of
the gen-eral disinterest in programming.
What this paper achieves to do is to present the research done on the
matter and demonstrate that the solution is actually counter-intuitively
simple and easy and that its implementation will solve the problem in
one swift blow.
- Introduction
Recently, there has been an increasing trend of reports by tech blog
writers that the rate of stu-dents who choose STEM careers (and computer
science in particular) in the country is dimin-ishing every year for
complex reasons. A large majority of these bloggers report that students
avoid careers in CS because of the media por-trayal of "all programmers
are geeks" (reason
A) and because it is hard for old people to read text on a monitor with
their bi-focals on (reason Bi). These bloggers then reach the
complex con-clusion that the sum of these reasons (A+Bi) is why we
cannot have nice things and why those in humanities refuse to learn
a new language other than dead languages like Latin, and that
some-body should do something about it.
Despite what the bloggers believe, a significant amount of research has
been done into figuring out why humanities majors prefer to stay away
from the programming field or computer science
- Preliminary Research
To understand the underlying issues that cause specialists in the
humanities to avoid program-ming, SOME(data) was collected from large
In-ternet communities known for their Internet eti-quette
("netiquette"), honesty, and well artic-ulated discussions, such as
4chan (pronounced "quatre-chan" or "cat-chan", named for its cat
enthusiasts). The communities were presented with a simple survey
(Appendix A) that asked the users for their academic and professional
backgrounds followed by a set of multiple-choice questions designed to
provide the mini-mal amount of useful information possible. The
results have been compiled into the following:95% of surveyors replied "Yes", 2% said "No",
1% said "[x] Genuflect", and 1% replied "Have you read your SICP
today?" The remaining 1% made no sense whatsoever.Clearly the research done by other professionals in the field was all
bogus and that this is the rea-son why English majors do not take up
program-
87
ming: 95% of all English majors who responded to the survey are English
majors. This new infor-mation blew us authors' minds and made them work
feverishly to cure this crippling disease of being an English major.
- Curing the Problem
Armed with this new knowledge, the au-thors immediately went to work to
help En-glish majors become computer science ma-
jors. A number of fixes were attempted such as tricking English majors
into register-ing for and taking Great Theoretical Ideas in Computer
Science, disguising post-fix pro-grams as arguments made by Thomas
Aquinas, and physical violence and psychological warfare asian-style
parenting, but most attempted solu-tions only had varying unacceptable
degrees of success.
But there was one solution that stood out amongst the others and managed
to help stop a person with the affliction of being an English major from
not learning to become a program-mer. During trials of using various
forms of Turing-complete languages, a group of English majors with no
previous background in computer science or mathematics rapidly scored
remark-ably high compared to their peers and the con-trol groups. This
group learned how to use an English-based language to describe to the
com-puter the tasks they wanted the computer to solve in an efficient and
competent manner. Af-ter a few more follow-up trials, the authors have
discovered that if a computer understands com-mands given to it in
simple English, then English majors can become expert programmers.
But the language the English majors were using was a simple language
defined by a simple gram-mar written on a napkin by one of the authors.
The English majors were asked to perform some tasks with some known
"code is almost English" languages such as Befunge and Ruby but
com-plications arose giving way to complaints such as "Ruby is
unreadable." so the authors set out to formally define a new language
with a new gram-mar that the English majors can easily under-
stand. This new language and solution is called:
Simple ML in MLA style or "SML/MLA".
4 Solution Proposal: Move over New Jersey
The proposed solution is a context-free Turing-complete language that
supports different paradigms of programming such as imperative,
functional, and cloud-style programming. The goal and purpose of the
proposed standard of Simple ML in MLA style is to create a workable and
modular standard of code writing that can be understood by all persons
with English de-grees with the help of the standards of English as
proposed by the Modern Language Association of America. The proposal is
a work-in-progress and this paper will only cover the main prelimi-nary
points that will become the founding prin-ciple of the language.
4.1 Quotations
One of the many points of conflict between pro-grammers and English
professors is the proper usage of quotations. For example, with the
phrase
One commonly known phrase that has all the letters of the alphabet is
"The quick brown fox jumped over the lazy dog."
On one side, the programmers insist that pe-riod should be outside the
quotation so as to not confuse the reader into thinking that there is
more to the sentence than shown. On the other, the professors just take
off unreasonable amounts of points for not putting the period in-side the
quotes. This pathological case for this issue is when programmers start
inserting in-line quotes of code into their papers and then lose a lot
of points because the professors do not agree on the parse structure.
Thus, the solution to this issue in the language proposal is that all
language delimiters immediately adjacent to the end-quotation character
will be placed inside the quotations. For example,
printf(\"Hello World.);\"
88
4.2 Punctuation and Arrays
Since the language is modeled after impeccable American English, all
code statements in the lan-guage must be ended by a period instead of a
semi-colon. Commas will still be used as delim-iters for lists unless
one of the entries in the list contain a comma. In that case, the comma
will be replaced locally in the list by the semi-colon. Also, if a list
or array has exactly two elements, then they may be specified with an
"and" in-between the two values and/or variables. The comma before the
"and" would be optional for this case.
Assign \"Kill Casshern!!\" to Kill Casshern. Assign \"Devour
Casshern!!\" to Devour Casshern. Construct an array called Dying Robot
Chant with the variables Kill Casshern, and Devour Casshern.
Give the Dying Robot Chant to the Robot Apocalypse function.
4.4 Code Flow
To take advantage of this new feature, decla-rations of lists and arrays
in the language are declared as demonstrated by the following code
block.
Loops are very simple in SML/MLA. For a while loop, one just needs to
tell the program to loop until a certain condition is met. For a for
loop, one just needs to tell the program to loop a cer-tain number of
times. And for the sake of code
Construct an array called A with the readability, unless specified in
line, the program following values: \"foo,\" \"bar,\" and \"baz.\"would
loop the code in a particular [paragraph]{.underline}.
Example:
Give the function Foo an array with the values of 1, 2, 3; \"foo;\" and
6.82.
4.3 Variable Name and Capitalization
All properly written SML/MLA programs must adhere to proper
capitalization rules as defined in the MLA. That is, all sentences (code
state-ments) must start with a capital letter no mat-ter what. Because
of the issue of people defin-ing variable names that start with
lower-case let-
While loops:
Until you get a SIGKILL signal, Print Kill Casshern. Print Devour
Casshern.
For loops:
Print Kill Casshern and Devour Casshern 100 times.
Count from 1 to 100, and every time you count, print Kill Casshern and
Devour Casshern.
89
4.5 Other MLA features
4.5.1 Spacing
All programs must be double-spaced and written in 11pt or 12pt Times New
Roman fonts. This is one of the most important points of the MLA that
implementations of SML/MLA must adhere to.
4.5.2 Functions and Bolding
All top-level functions within a file are the equiv-alent of Level 1
Headings. This means that the name of the functions and the type of the
value(s) they are returning must be bold and flush left, and also
enumerated in the order that the function is declared in the file.
4.5.3 Header File Inclusion
6 Conclusions and Future Work
With the advent of the new language specifica-tion of Simple ML in MLA
style, the authors are confident that English majors will start
swarm-ing in droves to join the ranks of the program-mers because of
the ease of adoption of the lan-guage and because of how easy it has
become to program. The authors are terrible people but it's okay
because now everybody and their grandma can code.There is a lot of work to be done in the future such as finalizing the
specification of the lan-guage, proving soundness and decidability of
the language, and implementing compilers that will compile the SML/MLA
programs into fast pro-grams in native code on various platforms.
- References
To put simply, when files known as 'header files" to most programmers
(or signatures, or abstract classes, etc.) are included or referenced to
in a program, they must be cited properly and prop-erly noted in a
bibliography or references list at the end of the program or library
file. Bold char-acters can be inserted into most available word
processors with the shortcut Ctrl + B.
- Hello World
The following code is what a Hello World pro-gram would look like if
it's written in a language compliant with SML/MLA.
- Function Main that returns an integer
It takes in an integer named argc and an argu-ment list of strings
called argv.
Ignore argc and argv and print to the termi-nal the following quote:
"Hello World." (stdio.h, printf)
Return zero.
Various Coders. "stdio.h - Standard Buffered Input/Ouput" GNU C Library.
Free Software Foundation, Inc.
-
Russell, Tony, Allen Brizee, and Eliza-beth Angeli. "MLA Formatting
and Style Guide." The Purdue OWL. Purdue U Writ-ing Lab, 4
Apr. 2010. Web. 28 Mar. 2010. -
Authors, Various. "The MLA Style Man-ual" Wikipedia. Wikimedia
Foundation Inc., 25 Mar. 2011. 28 Mar. 2011. -
Berenson, Dmitry. "Socio-Economic Fac-tors and Automated Statistical
Analysis of the LOLcat Language!!!1!" SIGBOVIK 2008. -
4chan.org, /prog/. 4chan BBS - Program-ming.
http://dis.4chan.org/prog/
90
Appendix A
Please check only one answer per question.
- Are you an English major? [ ] Yes
[ ] No
[ ] Other:
__________________________________________
Thank you for filling out this survey!
Appendix B
'-._ ___.....___
'.__ ,-' ,-.'-, HAVE YOU READ
'''-------' ( p ) '._ YOUR SICP TODAY?
'-' \
\
. \
\---..,--'
................._ --...--,
'-.._ _.-'
''-----''
91
92
Theoremin: Proof by Handwaving
Ben Blum (bblum@andrew.cmu.edu)
2011.04.01
Abstract
We present a method for extending proof systems to support an
additional lemma, Handwave. The theremin serves as a hardware platform
for converting hand-waving e orts on the part of the researcher into a
representative waveform; special software, called Theoremin, then
analyses the waveform to de-termine if the hand-wave was su ciently
convincing for a given inference. Using Theoremin, a theorem-proving
program may easily be extended to incorporate the Handwave lemma into
its logical system. We provide a simple \sound\"ness proof (have you
ever listened to a theremin?) and several example derivations which
have been vastly simpli ed by use of the new lemma.
93
94
Quantitative Comparison of Typesetting Software:
TEXShop vs. TechShop
Matthew Sarnoff*∗*
Department of Advanced Alcoholic Research
The Boom Boom Room, 1601 Fillmore Street, San Francisco, CA 94115
BoviKon East 2011
Keywords: typesetting, Knuth-Liang algorithm, user interface design,
human-computer in-teraction, human-computer attraction, human-computer
courtship, sexual turing test
- TEXShop
TEXShop (regrettably styled as TeXShop1) is a full-featured TEX
editor and previewer for Mac OS X.2 I used it to write up my homework
back in college after asking Alan V for the answers.
Figure 1: Screenshot of TEXShop on Mac OS X 10.9 "Lol Cat." Note the
presence of modern features like syntax highlighting and scroll bars.
∗ε-mail: computers_are_l@me.com, twittums: \@autorelease
1"Thou shalt not render the Name of the Holy Software without
subscript or proper kerning. To do so is an abomination: I am the
Knuth." (Knuthviticus 31:14)
2Rhymes with TEX. You lose all credibility by pronouncing it "Mac
OS Ecks" or, god forbid, "Mac OS Ten."
95
- TechShop
TechShop is a chain of member-based workshops that lets people of all
skill levels come in and use industrial tools and equipment to build
their own projects. They have locations in California, North Carolina,
and Michigan.3 It is not typesetting software.
Figure 2: A photo of the TechShop location in San Francisco. It has
been rotated 180*◦ and the colours have been inverted to seem hip and
edgy.*
- Similarities
They are both pronounced the same way.
- Differences
One can typeset documents, the other can't. One is an actual shop, the
other isn't. This is stupid.
- Conclusion
I'm bored. Who wants to go out tonight?
References, passed by reference
citeReference(&wikipedia);
citeReference(&techShopWebsite);
- According to Wikipedia. It's way too expensive so I've never been
> there.4
4Can you add a footnote to a footnote? Apparently yes.
96
Postscript
%!PS
/fillstroke
{ gsave 1 setgray fill grestore stroke } def
/circle
{ 0 360 arc } def
newpath
100 120 moveto
250 120 lineto
250 105 15 270 90 arc
250 90 moveto
100 90 lineto
stroke
100 100 25 circle fillstroke
120 80 25 circle fillstroke
showpage
97
98
Med School, CS Grad School, Both, or Neither?
Brian Hirshman
Med school and CS grad school aren't that different? Or are they? For
the following sets of questions, answer "med school", "CS grad school",
"both", or "neither". Chances are, you'll learn something in the process
-- even if that something is that you're now rethinking your original
career decision. The grass is always greener on the other side!
Acronyms and Mnemonics:
-
SNAFU
-
RBC
-
BIOS
-
OS
-
MRI
-
CAT
-
LR6SO4AR3
-
CPP
-
LOLWUT
-
Doesn't matter, you're probably looking it up on Wikipedia anyway.
Work and studying:
-
Crash and compile is a valid approach to problem solving
-
Some reading materials make more sense after a beer or two
-
The only way to make some reading materials to make sense is after a
> beer or two -
Half of the things you read about are named for a dead guy
-
You've taken a test using pencil and paper
-
You're being paid to be there
-
You spend at least half your time staring at a computer screen
-
You've dissected a field-relevant specimen
-
You were grossed out by said dissection
-
Professors actually make a pretense of teaching you
Research:
-
You're expected to do research.
-
At least some of your day is spent trying to avoid Question 1.
99
-
Most of your day is spent trying to avoid Question 1.
-
You lied in your answer to either Question 2 or Question 3.
-
You're expected to publish in order to end up in your next good job
-
Your work has been cited somewhere!
-
Your work has been cited, but really it was you who cited it
-
The first two years of your graduate education are the least
> important -
Publish or perish
-
What research?
Your field:
-
People do really, really, really important things in your field!
-
You have the potential to impact millions of lives.
-
No really, you have the potential to impact millions of lives!
-
You do really, really, really important things in your field
-
You decided what you wanted to be in kindergarten and stuck with it
-
You never decided to do this, but somehow ended up here?
-
You already regret going into you flied
-
You don't regret going into your field, but you might in a few years
-
Your expectations of what the field was like were shattered on your
> first day there. -
You're actually lying about your answer to question #9.
Peers:
-
Someone who started in your year is smarter than you, and you know
> it. -
Someone who started in your year is smarter than you, and you don't
> know it. -
Someone who started in your year is smarter than you, and you didn't
> think so, but then they totally proved you wrong by doing
> ________ better than you'll ever do. -
You're lying in one of your answers to the first three questions.
-
If your grades are at the top of your class, you aren't doing enough
> research. -
At least one member of every class you take already has seen the
> material before. -
You know a person who started your program and was legally unable to
> drink. -
You're reassured when you know you're not the oldest person in your
> program -
You look at the people who are starting your program next year and
> feel old. -
When a peer tells you they got a 2390 SAT, you *know* that's 800
> points too high
Sleep:
-
Wait, you actually sleep?
-
Wait, you're actually expected to sleep?
100
-
There are actually administrators in your program telling you to
> sleep. -
Things start at 8AM, but only for the first two years or so.
-
You've slept through class, multiple times.
-
You've slept in class, multiple times.
-
You've had at least one class about sleep.
-
The only sleep you know of involves a UNIX command.
-
You've woken up in terror thinking about something you learned
> recently. -
There will be many sleepless nights in your future even after you
> get your degree.
Comics, websites, and email:
-
People around you read XKCD.
-
People around you read PhD comics.
-
People around you don't read comics because they're lame.
-
People around you read Slashdot.
-
People around you read reddit.
-
You suspect that at least one of your professors doesn't actually
> use the internet. -
You know that at least one of your professors doesn't actually use
> the internet. -
You know that at least one of your professors doesn't have a
> computer in his office. -
You're expected to be constantly checking you email.
-
You feel naked without a smart phone.
Life:
-
...
-
...
-
...
-
...
-
...
-
...
-
...
-
Sometimes you feel like you're never going to make it through this.
-
People have mistakenly called you "doctor" already.
-
What life?
Answers:
Trivial, and left as an exercise to the reader.
101
102
SIGBOVIK 2011: TRACK SIX
Future Work
-
Classico Logic . . . . . . . . . . . . . . . . . . . 105
-
Hygienic Image Macros . . . . . . . . . . . . 107
-
Kanyes Lost Cosmological Theorem 113
-
TBD............................151
Taus Brock-Nannestad, Gian Perrone
Keywords: Cloud Computing, Cringeworthy Puns, Cake
-
Theorems for A Pound of Flesh . . . . . 161
-
Ringards Paradox . . . . . . . . . . . . . . . . 163
-
Parametric Polyamory . . . . . . . . . . . . . 165
-
Highest-order Logic . . . . . . . . . . . . . . . 173
-
Libeled Deduction . . . . . . . . . . . . . . . . 175
-
Antialiasing Pointers . . . . . . . . . . . . . . 181
-
The Lambda Timecube . . . . . . . . . . . . 183
-
GPS Conversion . . . . . . . . . . . . . . . . . . 195
103