A Visit to Al Quds Universitycs.brown.edu/about/conduit/conduit_v19n1.pdfA Visit to Al Quds University ALSO INSIDE: The Kanellakis Legacy Lives On Storytelling About Lighthouses 10E18342

A RESEARCH ANDALUMNI NEWS MAGAZINEDEPARTMENT OF COMPUTER SCIENCEBROWN UNIVERSITY

A Visit toAl Quds University

ALSO INSIDE:

n The Kanellakis Legacy Lives Onn Storytelling About Lighthouses

10E18342 fc Page 1 29-APR-10

Cyan Magenta Yellow Black The Signature Group

Notes from the Chair:the Latest News from 115 WatermanGreetings to all CS alums, supporters and friends.We are well into the second semester of the year and are as busyas ever. Fantastic things continue to happen in the departmentand I am thrilled to be able to share some highlights with you.

The �rst few months of the year brought several honors to ourfaculty. Congratulations are in order to Ben Raphael, who wasrecently awarded the prestigious Sloan Research Fellowship forhis work on developing novel computational and mathematicalapproaches to problems in biology, especially cancer genomicsand comparative genomics. Besides his Sloan Fellowship, Benhas previously received other recognitions for his work, includ-ing a Career Award at the Scienti�c Interface from the Bur-roughs Wellcome Fund.

I’m also delighted that the Association for Computing Machinery(ACM) recently elevated Tom Dean to Fellow for his developmentof dynamic Bayes networks and anytime algorithms. Tom, who wasa faculty member in the department from 1986 until 2007 andserved as department chair from 1997 to 2002, is currently a staffresearch scientist at Google, Mountain View, and maintains anadjunct professorship here at Brown. His research accomplish-ments include work on the role of prediction in planning, controland decision-making where uncertainty and the limited timeavailable for deliberation complicate the problem. His temporalBayesian networks, later called dynamic Bayes networks, made itpossible to factor very large state spaces and their correspondingtransition probabilities into compact representations, using thetools and theory of graphical models. Tom joins the department’sseven other ACM Fellows: Maurice Herlihy, Franco Preparata, JohnSavage, Eli Upfal, Andy van Dam, Peter Wegner and Stan Zdonik.

I’m pleased to announce that Brown University and the NationalUniversity of Singapore (NUS) have established a concurrentcomputational biology degree program. Students who completethe program will receive concurrent degrees from NUS andBrown: a bachelor’s degree in computational biology from NUSand a master’s degree in computer science with a specialdesignation in computational biology from Brown. FrancoPreparata, who has been a visiting faculty at NUS for severalyears, is the primary architect of this program at Brown. TomDoeppner has also been involved in setting up the program.Thanks to both Franco and Tom for their leadership in develop-ing this exciting new program.

In other news, the second year of the department’s Distin-guished Lecture Series was quite a success! This year’s lectureswere given by Michael Goodrich, University of California, Irvine

(October 20), Michael Littman,Rutgers University (October 29),Renee J. Miller, University ofToronto (March 4), Joel Emer,Intel (March 11), JenniferChayes, Microsoft Research(April 8), Jeannette Wing,Carnegie Mellon University andNational Science Foundation(April 19), Rob Schapire,Princeton University (April 22),and our own John Savage (April29), who gave a Special 70thBirthday Lecture.

The Distinguished Lecture Series also included the NinthAnnual Paris C. Kanellakis Memorial Lecture, which was deliv-ered on December 3, 2009 by John C. Mitchell of StanfordUniversity. This year’s Kanellakis Lecture was particularly specialsince Dr. Mitchell was a close friend of Paris’s as well as aresearch collaborator.

It is with great sadness that I share the news of the passing ofGeneral Kanellakis in late January, at the age of 97. I was luckyenough to spend a day with General and Mrs. Kanellakis at theirhome in Athens, Greece in September 2008, when they sharedwith me and Ph.D. student Babis Papamanthou memories ofParis and his family, as well as an extensive archive of documentsabout Paris’s professional life. I was struck by the warmth,generosity and intellect of General Kanellakis and know that hewill be truly missed by the department. Our thoughts are withMrs. Kanellakis during this dif�cult time.

Turning to the future, I look forward to seeing you at the nextComputer Science Reunion on Saturday, May 29, 2010. Weencourage all alums, friends and supporters to stop by. Pleaseregister for the reunion at www.cs.brown.edu/events/reunion/home.html

Finally, we urge you to contribute your professional and personalstories for inclusion in upcoming issues of Conduit. Yoursupport of and participation in department activities is alwaysappreciated and we are grateful to have such a tight-knitcommunity— thank you!

Roberto TamassiaPlastech Professor of Computer ScienceChair, Department of Computer Science

Spring | Summer 2010

10E18342 ifc Page 1 29-APR-10


Published twice yearly by the Departmentof Computer Science at Brown University,Conduit is distributed free to all computerscience alumni, faculty, staff, students andindustrial partners.

Queries about this publication can bedirected to: or115 Waterman StreetCampus Box 1910Providence, RI 02912

PUBL ISHER

Industrial Partners Program,Department of Computer Science

ED I TOR- IN - CH IEF

Amy Tarbox

GR APH IC DE S IGN

Brown Graphic Services

F E AT URED COLUMNIS T S

Sorin IstrailShriram KrishnamurthiClaire MathieuNikos Triandopoulas

CONDU¡T ON T HE WEB

John BazikKathy Kirman

PR IN T ING A ND DIS T R IBU T ION

Brett Turner, Information ServicesBrown Graphic Services

FE ATURES

A Visit to Al Quds University .................................PAGE 4

The Kanellakis Legacy Lives On ..........................................................PAGE 8

RESE ARCH

Storytelling About Lighthouses: When Professor DijkstraSlapped Me in the Quest for the Beautiful Code...............................PAGE 10

FACULT Y

The Faculty Speak Out .........................................................................PAGE 18

Faculty Notes........................................................................................PAGE 19

DEPARTMENT NE WS AND HAPPENINGS

Department Awards and Honors.........................................................PAGE 21

Recent Ph.D.s .......................................................................................PAGE 20

Parenthetically Speaking.....................................................................PAGE 24

Around the Department.......................................................................PAGE 26

P ING! ....................................................................................................BACK COVER

Conduit is printed on Burgo’s ChorusArt, an acid free and elemental chlorine free paper containing

50% recycled content and 15% post-consumer waste. You are the difference – reduce, reuse, recycle.

3 Condu¡t Spring | Summer 2010

Volume 19, Number 1 Spring | Summer 2010

25%

Cert no. SCS-COC-001154

10E18342 text_001 Page 1 29-APR-10



A Visit toAl Quds Universityin thePalestinian territories

The Holy Land is a special place for me.

It is the country of many of my friends and

colleagues in the academic world, and of my

religious roots. But virtually everyone

I know there is Jewish Israeli. What about

Arab Israelis, and what about Palestinians?

I was curious to meet them, so as to give

a concrete reality to their existence, and

to be able to associate faces of real people

to the abstract “Palestinian population”

sometimes mentioned in the news. On

the occasion of a trip to the Hebrew

university of Jerusalem in December 2009,

I arranged to give a seminar at Al Quds

university in the Palestinian territories.

By Claire Mathieu

10E18342 text_002 Page 2 29-APR-10



One Saturday morning, my driver, a short man witha mustache, picks me up in the hotel lobby. I climbinto his light, unmarked truck, and off we go, himdriving silently around the Old City and away intothe mountains, me wondering where we are headedand feeling a little bit nervous. The campus of AlQuds is in Abu Dis, a few miles from Jerusalem.Since the wall has not yet been built across the roadwe take, the only barrier we pass on the way is asmall roadblock on the other side of the highway. Iwatch the dry mountainous landscape and theMiddle-Eastern towns, noticing a good number ofconstruction projects and of brand new buildings. Inone valley, we pass a “USAid” sign mentioning somewater project. We may be in a developing nation, butinternational help does seem to be arriving!

Once at Al Quds university, my host greets me inthe very modern information technology centerand takes me for a tour of campus, followed by awoman introduced as the person in charge of

“public relations.” She hardly says anything all day,but, because of her presence, I never have aone-on-one conversation with any one; is it

chaperoning or a discrete surveillance? I am notsure. The campus is sunny and very pleasant, andeverything is neat and well-kept. We hear the gentlesound of water trickling from the fountains. Scat-tered groups of students are chatting, studying orrelaxing on the stone benches under the olive trees.Most of the women wear colorful silk scarves, andthe peaceful scene would look idyllic, if my Westernmind didn’t remind me that scarves are supposed tobe a symbol of oppression of women. Throughoutmy visit, to mark their deference for their guest, mycompanions are constantly stepping aside to let mego �rst through doors and lead the way, hence arather complex ballet of foot movements. I am notused to such manners and it is a bit awkward for me,especially since I usually do not know whichdirection we are supposed to go next! I can onlyhope I haven’t committed any serious breach ofetiquette; if I have, my hosts are too polite for me tobe able to tell.

After meeting the dean of science, we go to visit the“museum of political prisoners” where I see a map ofprisons (interestingly, I note that the 1967 borders

10E18342 text_003 Page 3 29-APR-10



A Visit to Al Quds University

are drawn on the map) and documents aboutPalestinians in Israeli prisons: photos, art created byprisoners, letters to family members. Although I tryto keep in mind that this is one particular perspec-tive of the con�ict, it is rather striking. Previously, Ihad suggested to an Israeli paci�st-leaning friendthat he, too, might wish to visit there, but now I amhaving second thoughts about that idea. It is onething to want to be a peacemaker and wish to talkwith people of goodwill on the Palestinian side, butquite another to watch an exhibit detailing exac-tions committed by one’s fellow countrymen! Wethen meet the director of the museum, who offersme a cup of tea. Since he does not speak English,my host translates. He gives me various statistics. Iask: “How do you see the future?” What a question!A deluge of words follows, summarized by the �rstsentence: “The future is dark.” I ask my host: “Doyou have any students who have been to prison?”He smiles at my naïveté: “Yes, of course,” and thenproceeds to give me an extraordinary quote:

“Prison is a �ower that every young Palestinian hasto smell!” I am stunned. But he adds immediatelythat, as for himself, by good luck he has never goneto prison so far.

The wall is there, tall, dark, forbidding. It standsright next to the campus entrance, winds down intothe valley and goes back up to block the horizon onthe hill facing us. Its presence is a constant

reminder of the political situation. No morerational reasoning on statistics, no more carefulweighing of objective pros and cons: I am pained byits ugly, intrusive sight.

I give a talk to an audience of a dozen students andfaculty, in a well-equipped room. There are a fewcomments, including one by a woman with a scarf,in the back of the room: bravo! People seeminterested to learn how search engines (such asGoogle) auction their ad slots, and I get some goodquestions. We then have lunch in a restaurant oncampus. I am surprised to learn that in ComputerScience the majority of the students are women.Apparently, this phenomenon can be traced back tothe �rst intifada, when many young men went intohiding. In addition, the women also tend to getbetter grades, but after graduation their salaries aremuch lower than men’s - hardly a surprise. Mostgraduating students go on to teach, although lowsalaries often require teachers to take on a secondjob. We discuss the problems of education at theprimary and secondary levels, entirely based on rotelearning and from which creative thinking iscompletely absent. Al Quds was the �rst universitywith a computer science department. I ask: does it

10E18342 text_004 Page 4 29-APR-10



A Visit to Al Quds University

have the best computer science students? Notnecessarily. Because of the dif�culty of traveling andgoing through checkpoints that risk causing delaysof several hours, students tend to choose theiruniversity according to accessibility. Hence Ramal-lah students tend to go to Bir Zeit, whereas Al Qudsattracts students from the neighboring region. Moststudents are self-supporting and work full-time, sothey only study part-time, one or two days per week,and naturally that delays graduation by years. Thereare undergraduate and masters students (no Ph.D.program). Among the faculty, some hold a mastersand some have a Ph.D. I am told that one is workingtowards a Ph.D. in Math at the Hebrew university,and another is doing a Ph.D. in education at theWeizmann institute.

What do they need? They have long-term plans ofreforming the high-school curriculum, with theintroduction of problem-solving and of mathemati-cal competitions, once teachers become receptiveto the idea. They are very interested in e-learningand have had contacts with CNAM in France on thesubject. They are looking for scienti�c collabora-tions with opportunities for visits, student exchang-es, etc. They wish to introduce their mastersstudents to research and would welcome long-dis-tance co-advising on research projects, suggestionsfor good research topics, and other ways to helpthem develop their research. In short, these are thenatural priorities of faculty who wish to bettereducate the youth of their country and fosteradvanced learning. In a way, in spite of the adversar-ial context they sound more optimistic about thefuture than the Israelis I have talked to!

We wait for my return taxi, standing near the wallthat once again takes center stage. My host tells methat his sister, being married to a Jerusalemite, nowhas an Israeli ID that makes it dif�cult for her tocross the checkpoints into the Palestinian territories.

“She was born here and wants to come and visit hermother who still lives here. How can it be that theborder patrols don’t want to let her through, for herto see her own mother?” he sighs. But he carefullyrefrains from following up with political statements.

As for me, when the taxi crosses the checkpoint Iam simply waved through and, just like on my tripto Bethlehem, no one even looks at my passport.For a tourist like me, the borders are wide open!

In academia, we like to believe that in the long termeducation cures all problems. The Israeli-Palestin-ian con�ict is much too complex for outsiders likeme to even begin to gauge but, where there arestudents trying to learn and professors trying toteach and to develop a knowledge of research, howcan it not be a good idea to try to participatetowards their general goal of academic learning?

10E18342 text_005 Page 5 29-APR-10



It is with great sadness that

we announce the passing of

General Eleutherios Kanellakis,

Paris’s father, on January 27, 2010

at the age of 97.

Paris Kanellakis was a distinguished computerscientist and an esteemed and beloved memberof the Brown Computer Science Department.He joined the Department in 1981 and was pro-moted to the rank of full professor in 1990. Hisresearch area was theoretical computer science,with emphasis on the principles of database sys-tems, logic in computer science, the principlesof distributed computing and combinatorialoptimization. He died in an airplane crash onDecember 20, 1995, along with his wife, MariaTeresa Otoya, and their two young children,Alexandra and Stephanos.

Eleutherios Kanellakis was born in the area ofCorinthos in Peloponnesus, studied Law at theUniversity of Athens and worked as magistratein the island of Ikaria. During World War II hejoined the Greek army as a reserve of�cer whoparticipated actively in the efforts liberatingGreece, where he received many distinctions. In1942, he left occupied Greece to join the GreekForces in Minor Asia. From 1947 he served as asupreme of�cer in the Military Department ofJustice of the Greek Air Forces, where he built a

long career with many distinctions. He complet-ed a two-year masters degree in “Penology andCrime” at the University of Michigan, where hepublished a book on his thesis. For many yearshe taught Constitutional and Governance Law inthe Military School of the Greek Air Forces, andpublished many books. He spoke �ve languages.

Overcoming the tragic loss of their only son, Gen-eral and Mrs. Kanellakis established two graduate-student Fellowships in his memory in 1998, oneat Brown and one at MIT, where Paris earned hisdoctoral degree. Since then, the Kanellakis Fel-lowship has been received by 26 Ph.D. studentsat both institutions. Many members of the BrownComputer Science Department and all the Fel-lows developed a strong relationship with Gen-eral and Mrs. Kanellakis and have visited themduring trips to Greece. General Kanellakis wasparticularly happy to see the Fellows makingprogress in their careers, having Paris’s academiclegacy as an excellent example. Up until very re-cently, General Kanellakis had been consistentlymaintaining an archive with Paris’s accomplish-ments, and last September he offered this archiveto our department. This book is available for any-one to see in the Department Chair’s of�ce.

General Kanellakis’s funeral was held on January28th at the Papagou Cemetery in Athens. Funer-al wreaths were sent on behalf of the CS Depart-ment and the Kanellakis Fellows at Brown andMIT. Fellows Aris Anagnostopoulos (Ph.D. 2006,Brown), Christos Kapoutsis (Ph.D. 2006, MIT)and Ioannis Tsochantaridis (Ph.D. 2005, Brown)attended the funeral. In addition, the CS depart-ment provided �nancial support, in honor ofthe memory of General Eleutherios Kanellakis,to the Hellenic Student Association of Brown fortheir Greek Orthodox Easter celebration.

Paris C. Kanellakis FellowsGeneral Kanellakis’s vision to help youngscientists through a graduate-student fellowshipin honor of Paris has certainly come true. Thelist of PCK Fellows is already long and increas-ing. Current students are making good prog-ress in their graduate programs and alreadygraduated students are pursuing successfulcareers as research scientists and faculty inresearch labs and academia.

The Kanellakis Legacy Lives OnEdited by PCK fellows Yola Katsargyri (MIT) and Nikos Triandopoulos (Brown). Special thanks to Manolis Kellis and Amy Tarbox for their help.

Brown1997–1998 Emmanuel Manos Renieris

1998–1999 Ioannis Tsochantaridis

1999–2000 Costas Busch, EmmanuelManos Renieris & IoannisTsochantaridis

2000–2001 Aris Anagnostopoulos,Nikos Triandopoulos

2001–2002 Aris Anagnostopoulos,Nikos Triandopoulos

2002–2003 Olga Papaemmanouil

2003–2004 Christos Amanatidis,Alexandru Balan

2004–2005 Aris Anagnostopoulos,Tomer Moscovich,Nikos Triandopoulos& Yannis Vergados

2005–2006 Aris Anagnostopoulos

2006–2007 Daniel Acevedo Feliz,Glencora L. Borradaile& Eric Ely Rachlin

2007–2008 Charalampos Papamanthou

2008–2009 Aggeliki Tsoli

The Department has nominated FoteiniBaldimtsi for the 2009–2010 KanellakisFellowship.

MIT1999–2000 Manolis Kamvysselis (Kellis)

2000–2001 Christos Kapoutsis

2001–2002 Aristidis Karalis

2002–2003 Christos Christoudias

2003–2004 Anastasios Sidiropoulos

2004–2005 Nikolaos Andrikogiannopoulos

2005–2006 Apostolos Fertis

2006–2007 Georgia–Evangelia Katsargyri

2007–2008 Georgios Papachristoudis

2008–2009 Vasilios–Marios Gkortsas

2009–2010 Georgios Angelopoulos

Revisited: The Kanellakis Scholars

10E18342 text_006 Page 1 29-APR-10



Revisited: The Kanellakis Scholars

Paris Kanellakis Fellows Meet

By Yola Katsargyri, MIT PCK Fellow

Two years ago the Kanellakis Fellows from bothBrown University and MIT took the initiative toorganize a dinner and meet each other. The dinnerwas hosted by Manolis Kellis, who, being one of the�rst fellows, brought us all together and tried toinculcate in our minds the idea of the KanellakisFellows family. Although some of us were meetingfor the �rst time, it only took a few hours to realizethat we were not just a few people who happened tohave received the same fellowship. Paris’s short lifeand General and Mrs. Kanellakis’s generosity andcourage bonded us strongly.

In 2009, the second PCK Fellows dinner wasorganized by Nikos Triandopoulos and OlgaPapaemmanouil. They warmly offered their homeas a place to reconnect and also welcome the newmembers to the Kanellakis Fellows family. Thisdinner provided the opportunity for the Fellows toshare creative ideas that aim at further preservingthe PCK Fellowship and the memory of Paris, suchas creating a Kanellakis Fellows webpage andorganizing a yearly scienti�c symposium for Ph.D.students in memory of Paris, among others.

The parental love that we all received from Eleuthe-rios and Argyroula Kanellakis was a strong incentivefor the Fellows to support each other and to form astrong community that keeps Paris’s memory andlegacy alive for the years to come.

Ninth Annual Paris C. KanellakisMemorial LectureOn December 3, 2009, John C. Mitchell of StanfordUniversity delivered the Annual Paris C. KanellakisMemorial Lecture. This lecture series has been heldannually by our department in honor of Paris on oraround his birthday. In a standing-room onlylecture, Prof. Mitchell kindly devoted a longintroduction to Paris, describing their close friend-ship, as well as two articles that they coauthored.The talk, titled “Safety on the Wild and WoolyWorld-Wide Web: Sandboxing Untrusted JavaScript”outlined some of the practical security problemsthat have arisen in recent years as a result ofcombining trusted and untrusted content anddescribed several methods for solving these prob-lems. A reception followed the lecture, where Prof.Mitchell met PCK Fellows from Brown and MIT andspent time talking with them, sharing memories ofParis and providing advice for successful comple-tion of Ph.D. studies.

MIT: Manolis Kellis, Georgios Angelopoulos, Georgia-Evangelia Katsargyri, GeorgiosPapachristoudisBrown: Olga Papaemmanouil, Aggeliki Tsoli, Nikos Triandopoulos, Foteini Baldimtsi,Charalampos Papamanthou, Yannis Vergados, Aris Anagnostopoulos(Also in the picture: 5 friends of the Fellows.)

10E18342 text_007 Page 1 29-APR-10


Research


THE SLAPI �nally met E.W.Dijkstra, the brilliant computer scientistwhom I had admired for years, in Newport, Rhode Island. Itwas 1986, three years after my immigration to the UnitedStates from Romania, and Dijkstra, by then the SchlumbergerCentennial Chair in Computer Sciences at the University ofTexas at Austin, had invited me to join him for a day at theSummer School on Program Construction being held at SalveRegina University.

The day alternated between Professor Dijkstra lecturing andwriting on a blackboard with his exquisitely precise handwrit-ing. (Even now, I can recall how he returned to the black-board to redraw a capital letter M that he felt lacked parallelvertical lines.)

At lunch, he sat near me at a long table with about ten others.Everybody there was eager to hear what this master would say,and because he spoke softly, few people if any were talking.We had just about �nished eating when Dijkstra raised hisvoice to address me.

“Sorin - I have a problem for you.”

“Sure,” I said, thinking that this was the moment I had beenwaiting for since 1983, when I �rst solicited Dijkstra’s commentand guidance on a technical report I had sent him previously.

The silence at the table became more pronounced.

“Suppose we play a two-player game played withidentical coins on a table,”Dijkstra began. “We havea bag of coins with asmany as we need. Thetwo players alternativelytake a coin from the bagand place it on the table.The rules of the game forbid the coin to sit on top of anothercoin on the table, but it could hang off the table as long as itdoes not fall off. The player who puts the last coin on the tablewins the game.”

There was something in Dijkstra’s expression – a combinationof smile and an intense Don Quixotesque gaze – that seemedto raise the stakes at the table. I awaited the question.

“Is there an algorithm for one of the players to win always?” heasked with the quiet, calm assurance of a man who saw it all,who knew what would happen next.

I said nothing for a minute or two, anticipating the excite-ment that was to come. Then, as if jarred awake from a pleas-ant dream, I did what one does in such a moment: With myright hand, I reached down the neck of my sweater to removea pen from my shirt pocket.

My hand was not even halfway there when – slap! Dijkstrasmacked my hand away from its goal. Those few seconds thatfollowed were long but I was slow to �gure out, and I couldnot believe the strength of the slap. Instinctively, I drew awayfrom him and the others at the table recoiled.

Dijkstra broke the silence. “Don’t mess up your thoughts,” heproclaimed. “Keep it simple, in your head.”

It goes without saying that for the next few minutes I wasuseless at solving the puzzle and I bubbled nonsense. Hewaited a bit, then began offering clues. “Did I tell you any-thing about the table?”

Storytelling About Lighthouses:When Professor Dijkstra Slapped Mein the Quest for Beautiful CodeBy Sorin Istrail

* Arrogance in computer science * Empirical science? * Philosophy * Goddess Reason* A Critic’s Dilemma * Pajamas algorithms * Pastiche pie prize

“If 10 years from now, when you are doingsomething quick and dirty, you suddenlyvisualize that I am looking over your shoul-ders and say to yourself: ‘Dijkstra would nothave liked this.’ well that would be enoughimmortality for me.” [1]

Edsger Dijkstra

10E18342 text_008 Page 8 29-APR-10



Research

“No,” I responded. I began to think out loud. “As it must betrue for all tables ...”

He interrupted with a question. “What is the smallest table?”

“One point?” I said trying to recover.

“Who wins then?” he asked.

By now rational, I responded. “The �rst player. So the �rstplayer always wins.”

“Which table to consider next?”

“As big as a coin; again �rst player wins.” I thought I was on a roll.

“No.”

Me: “How about as big as two coins?”

Dijkstra: “Yes. Who wins then?”

Me: “First player puts his coin in the middle of the table. AndI am not sure if the second can put his coin, but if the secondcan, so can the �rst again symmetrically; if not, neither can ...so �rst wins again!”

Dijkstra: “Do you have the algorithm now?”

Me (excitedly, knowing what he wanted to hear): “The invari-ant is Every point on the table, except the center, has a sym-metric point with respect to the center.”

Dijkstra: “Bravo.”

(To translate the answer into Dijkstranese: The �rst player alwaysputs his �rst coin on the center of the table. The second playerputs his coin somewhere on the table that is free, and then the�rst player makes his move in exactly the symmetric point tothe center, which is always free by the invariant. And so on,mutatis mutandis. The invariant assures the correctness of thealgorithm no matter how many moves/how big the table.)

One of the Joys of Life and the Cruelty ofReally Teaching Computing ScienceDijkstra was our Professor-in-Chief. How to teach computingscience was one of the most fascinating life-long themes ofre�ection for Dijkstra. He wrote many articles about teachingmethodology, he was the patriarch of posing beautiful prob-lems and silly games, which gave him the opportunity to teachthe art of problem solving. His algorithms, solutions of hisgames, de�ne beautiful code, raising the derivation of thecode from the program speci�cation to an art performance.Those are not just beautiful games; they go to the heart of thedif�culty of programming methodology, and illuminate hard-to-grasp complexities. It is an art form to invent such sillygames; this art form should be encouraged and rewarded.They make the perfect opening of a lecture on the subject,especially because they satisfy the right axioms: unique solu-tion, simplicity of algorithmic solution and rational step-by-step derivation of algorithm, generalizations become extreme-ly dif�cult problems (try the above silly game with coins on aconvex table), and most of all, your can �nd their solutionwithout pen and pencil, in your head. I had the pleasure ofattending some of Professor Dijkstra’s lectures, to be trainedone-on-one in the art, to have read many of his articles onteaching computing science, and to be inspired to uncoverand present silly games occuring in everyday life, although

their lessons are not as beautiful and important as thoseconcocted by the Master of Silly Games. For the youngergeneration, here is a list of some his gems: the dining philoso-phers, the elephant made of mosquitoes humming in harmo-ny, the toilet and trains, the plateau problem, the self-stabiliza-tion problem, average page fault frequency problem, theDutch national �ag, banker’s algorithm, [the cryptographicgame from[EWD 666]“a problem solved in my head.”

Dijkstra examined radical innovation in teaching, and his thesisis always that universities should have guts in teaching.

“Teaching to unsuspecting youngsters the effective use of formalmethods is one of the joys of life because it is so extremely rewarding.Within a few months, they �nd their way in a new world with ajusti�ed degree of con�dence that is radically novel for them; withina few months, their concept of intellectual culture has acquired aradically novel dimension. To my taste and style that is whateducation is about. Universities should not be afraid of teachingradical novelties; on the contrary, it is their calling to welcome theopportunity to do so. Their willingness to do so is our main safe-guard against dictatorships, be they of the proletariat, of thescienti�c establishment, or of the corporate elite.” E. W. Dijkstra [2]

In his “How do we tell truths that might hurt” [3] Dijkstrawrote that “it is practically impossible to teach good programming tostudents that have had a prior exposure to BASIC: as potentialprogrammers they are mentally mutilated beyond hope of regeneration”and that “the use of COBOL cripples the mind; its teaching should,therefore, be regarded as a criminal offense.” Computer sciencefaculty might consider such statements dramatic, insulting,even ridiculous, but I learned from a colleague that Dijkstrarefused to accept students in his class if they had been ex-posed to BASIC or COBOL.This, as well as his incorporation of “radical novelties” such assolving a problem without pencil or paper may help youunderstand the “cruelty” Dijkstra referred to in EWD1036.

Dijkstra was nothing if not consistent, holding himself to thesame standards he held others.

He arrived at his most famous algorithm, known as The Short-est Path, in his head “while I had a cup of coffee with my wifeon a sunny café terrace in Amsterdam,” he has said.

“The algorithm for The Shortest Path was designed for the purposeof demonstrating the power of ARMAC at its of�cial inaugura-tion in 1956, the one for The Shortest Spanning Tree was de-signed to minimize the amount of copper in the backpanel wiringof the X1. In retrospect, it is revealing that I did not rush topublish these two algorithms: at that time, discrete algorithms hadnot yet acquired mathematical respectability, and there were nosuitable journals. Eventually they were offered in 1959 to Nu-merische Mathematik in an effort of helping that new journal toestablish itself. For many years, and in wide circles, The ShortestPath has been the main pillar for my name and fame, and then itis a strange thought that it was designed without pencil and paper,while I had a cup of coffee with my wife on a sunny café terrace inAmsterdam, only designed for a demo ...” E.W. Dijkstra [4]

He also believed that solutions could – should – be elegant,and that elegance could prove elusive if a programmer’s �rststep is to reach for a pen or pencil. (A 10 years earlier expla-nation of a future “slap”:)

10E18342 text_009 Page 9 29-APR-10


Research


“I observed a few years ago that the moment at which mathemati-cians introduce avoidable complications very often coincideswith the moment that they resort to such mechanical aids aspencil and paper,” Dijkstra wrote in “A problem solved in myhead.” [5]“It was then that, for the sake of clarity of my ownthinking, I decided to be less liberal with the use of pencil andpaper and not to use them when I could avoid using them.”

“ARROGANCE IN COMPUTER SCIENCE ISMEASURED IN NANO-DIJKSTRAS”

Alan Kay 1997

“Arrogance in computer science is measured in nano-Dijkstras,”computer scientist and Turing Award winner Alan Kay saidduring his 1997 OOPSLA keynote. The quip, which produceda roar of laughter from the audience, and his ensuing criti-cism of Dijkstra is preserved on YouTube.

I listened carefully to the YouTube video, I found Kay to be farfrom eloquent. He began his keynote with an anecdote:

“He [Dijkstra] once wrote a paper of the kind that he liked towrite a lot of which had the title ‘On the fact that the Atlantichas two sides’ [EWD611] and it was basically all about how dif-ferent the approaches to computing science were in Europe,especially in Holland, and in the United States. In the U.S. here,we were not mathematical enough and, gee, in Holland, if youare a full professor you were actually appointed by the queen,and there were many other important distinctions made betweenthe two cultures.

“So I wrote a rebuttal paper and it was called ‘On the fact thatmost of the software was written on one side of the Atlantic,’and it was basically about – ’cause I have a math degree, too– that computers formed a new kind of math … they don’t really�t well into classical math… It was about a kind of practicalmath. The balance was between making structures that weresupposed to be consistent of a much larger kind than classicalmath had ever come close to dreaming of attempting, andhaving to deal with the exact same problems that classical mathof any size has to deal with, which is being able to be convincingabout covering all the cases.” [6]

Defending “Arrogance”Your Honor, ladies and gentlemen of the jury, ProfessorDijkstra is accused of “arrogance.”

Well, it is well know that at times, Professor Dijkstra expressedhis strong opinions with critical irreverence, infuriatinglyinsensitive, but always with eloquence. To cite some extremes,he called the great logician Bertrand Russell a “dilettante”regarding his mathematical notation, accused John von Neu-mann of bringing (contra-productive) anthropomorphicterminology to computer science inspired by his work on thebrain as “medieval speculations”, and one of his favoritetirades against Software Engineering, coined “the DoomedDiscipline,” and even harshly, “How to program if you cannot.”Even more extreme, he recommended that students in

introductory programming courses should be prevented fromthe temptation to execute their programs, as they should betaught through mathematical logic to infer the correct pro-gram hand in hand with their proof of correctness.

Despite these points of view, I believe that some of theseextremes were part of his dramatics, theatrical, and sometimeshumorous avenues to deliver a forceful message for change, apoke in the eye for those asleep at the wheel, about the needto breakthrough deadlock, and the need to be bluntly honest.A model of “how to say truths that might hurt.” A number ofhis critics address these extremes of his writings as if they arehis entire position

Professor Dijkstra was indeed arrogant but about honesty,about the programming elegance, and about radical noveltiesin education. Even if extreme, I still prefer his Don Quixo-tesque exceedingly demanding goal of “logic proof” as sci-ence-base driving force, to the faith-based “every program hasbugs” convenience. John von Neumann used to say that it iseasier to explain science with god than without god.

I would argue that Professor Kay’s and Professor Dijkstra’spoints of view are at the two at extremes; “not really fond ofmathematics in programming” vis-à-vis of the de-empiriciza-tion of programming “craftsmanship” through mathematicstowards the science of programming.

Although the two have pursued magni�cent bodies of work thatinspired many, they disagree in ways which have nothing to dowith “truth.” They each have been forcefully articulating theirown philosophy and there is nothing wrong with that. On thecontrary, philosophical discourse is a must when dealing withthings as complex as Computing Science. Who wants to talkabout the obviously neglected empiricism in specifying require-ments of large codes, a really embarrassing subject? What is thena programmer to do?

In what follows, I will argue the Dijkstraian quest for program-ming de-empirization though mathematics. And to bringhome the point I want to make about bringing philosophy outof the closet in computing science, I will go though an irrever-ent, and infuriating to some, tour of the principles of philoso-phy using the writing of the great mathematician Gian-CarloRota. The “axioms” formulation are my attempt to presentRota’s argument and to show how relevant philosophy is toaddressing the bottlenecks and failures in the software designof large systems, e.g., “Inevitability of failure” or the “Myth ofprecision” or the “Dictatorship of de�nitiveness.”

I believe that if you choose to critique someone as eloquent asDijkstra, you must at least strive to do so in a similar vein. It isprobably unfair to ask someone to match Dijkstra’s eloquence.Few can. But as Dijkstra offers this view about non-principlebased criticism: “I love to be corrected. (Besides being a mostinstructive experience, being corrected shows that the other onecares about you.) If, however, they only get infuriated because Idon’t play my game according to their rules, I cannot resist thetemptation to ignore their fury and to shrug my shoulders in themost polite manner. ... I have come to the conclusion that thereare such things as ‘disabling prejudices.’” [7]

10E18342 text_010 Page 10 29-APR-10



Research

In fairness, it is hard for me to believe that Professor Kay’spiece recorded on Youtube, with its ramblings and profanitywas part of a prepared text for his Keynote. Probably whathappened was his delivery got emotional, and then inarticu-late; “disabling prejudices” indeed! I see this as a sign of adeeper problem in computer science, called by some TheSoftware Crisis.

Is Computer Programmingan Empirical Science?Re�ecting on the nature of computing science, in general,and programming in particular, one needs to focus on theempirical aspects, mostly belonging to the software engineer-ing focus of the discipline, as well as on the work towards thede-empirization of programming, via mathematical sciences.

John von Neumann talked about the de-empirization of thenatural sciences with the exquisite clarity of his writing. Hepresented it by talking about the double face of mathematics.

“The most vitally characteristic fact about mathematics is, inmy opinion, its quite peculiar relationship to natural sciences,or, more generally, to any science which interprets experi-ence on a higher than purely descriptive level.” [8] He gavetwo such glorious examples: one being Geometry and theother Calculus which both started as natural, empiricalsciences. Then its de-empirization happened by the mathe-matical method. “Some of the best inspirations of modernmathematics (I believe, the best ones) clearly originated inthe natural sciences. The methods of mathematics pervadeand dominate the “theoretical” divisions of natural sciences.In modern empirical sciences it has become more and morea major criterion of success whether they have become acces-sible to the mathematical method or to the near-mathemati-cal methods of physics. Indeed, throughout the naturalsciences an unbroken chain of successive pseudomorphoses,all of them pressing towards mathematics, and almost identi-�ed with the idea of scienti�c progress, has become moreand more evident. Biology becomes increasingly pervaded bychemistry and physics, chemistry by the experimental andtheoretical physics, and physics by the very mathematicalforms of theoretical physics… One has to realize this duplici-ty, to accept it, and to assimilate it into one’s thinking of thesubject. This double face is the face of mathematics, and I donot believe that any simpli�ed, Unitarian view of the thing ispossible without sacri�cing the essence.”

Von Neumann’s deep questions about the nature of intellectu-al work in mathematics can serve as a guide into our analysisof computing science: Is computing science an empiricalscience? Or, more precisely: Is computing science actuallypracticed in the way in which an empirical science is prac-ticed? Or, more generally: What is the computing scientist’snormal relationship to his subject? What are the criteria ofsuccess, or desirability? What in�uences, what considerations,control and direct his effort?

Dijkstra addressed these questions in his writings. In “Crafts-man or Scientist?” [9] Dijkstra discusses the two extreme tech-niques in teaching programming. At one extreme is

the “craftsmanship style,” similar to the work of the futurecraftsman joining a master and “learning by osmosis, so tospeak, the skills of the craft … a well-guarded secret.” At theother extreme is the “scientist style.” The future scientist learnsfrom a teacher who formulates knowledge and skill as explicitlyas possible through free interchange of knowledge and insights

– “being non-secretive is one of their rules of professionalconduct.” A physician and a physicist, respectively, are exam-ples of people who, more often than not, practice the two styles.

However, “mathematicians are somewhere in between: math-ematical results are published and taught quite openly, butthere is very little explicit teaching on how to do mathemat-ics, and publishing besides the results also the heuristics thatled to them is regarded by many as ‘unscienti�c’ and there-fore, bad style: quite often the editor’s censorship will try toprohibit their publication.”

Dijkstra asks: "Where along this scale should we place the teachingof programming?"

Twenty-two years before Kay’s 1997 comment, Dijkstra charm-ingly alluded to coming trouble: “To make implicit knowledgeexplicit … we should realize that changing a craft into a science, andmaking public property of the secret knowledge of the guild, will alwayscause the guild members to feel threatened. For many a ‘puzzle-minded’virtuoso coder of the early sixties, the [recent] scienti�c development …has been most unwelcome. … He feels like the medieval painter thatcould create a masterpiece whenever his experience enabled him torender proportion well, who suddenly found himself overtaken by allsorts of youngsters, pupils of Albrecht Dürer and the like, who had beentaught the mathematical constructions that were guaranteed to surpasshis most successful, but intuitive renderings. And with nostalgia helooks back to the good old days when his experience and feeling madehim an outstanding craftsman. And we should realize that, as far asprogramming is concerned, the battle is still going on. -- "craftsmen"[proposals] … had a pronounced anti-intellectualistic �avour: itstressed that students should be taught how to solve the problems of ‘thereal world’ and that, therefore, the curriculum should pay as littleattention as possible to ‘abstract subjects.’”

Dijkstra advocated a blending of the two teaching styles.The “disastrous blending, viz. that of the technology of thecraftsman with the pretence of the scientist” is not the solu-tion because “the craftsman has no conscious, formal gripon his subject matter, he just ‘knows’ how to use his tools.If this is combined with the scientist's approach of makingone's knowledge explicit, he will describe what he knowsexplicitly, i.e. his tools, instead of describing how to usethem! … It deserves a special warning because, besidesbeing disastrous, it is so respectable!”

His preferred blending: “As teachers of programming we should tryto blend the technology of the scientist with the pretence of the crafts-man.” Sticking to the technology of the scientist means beingas explicit as we possibly can about as many aspects of our tradeas we can. “Now the teaching of programming comprises the teachingof facts -- facts about systems, machines, programming languages etc.-- and it is very easy to be explicit about them, but the trouble is thatthese facts represent about 10 percent, of what has to be taught: theremaining 90 percent is problem solving and how to avoid unmasteredcomplexity, in short: it is the teaching of thinking, no more and no less.”

10E18342 text_011 Page 11 29-APR-10


Research


Computing Science and Philosophy“Experimental psychology, neurophysiology, and computer science mayturn out to be the best friends of traditional philosophy.” Gian-CarloRota [10]

Perhaps the best way to explore the Dijkstra-Kay argument is todetour brie�y into philosophy. You may think it is impracticaland plays no role in computer science, but I would argue that thephilosophy of computer science is at the heart of their debate.

We begin with the great Gian-Carlo Rota, whose writingsmatched the eloquence of Dijkstra. In his essay “The Perni-cious In�uence of Mathematics upon Philosophy,” Rotashowed us the parallels between the double life of mathemat-ics and the double life of philosophy. (We will talk about thedouble life of computer science as well.)

Rota characterized the double life of mathematics as truthand proof. “In the �rst of its lives mathematics deals with facts,like any other science,” he wrote. “The facts of today’s mathe-matics are the springboard for the science of tomorrow.” In itssecond life, Rota wrote, mathematics deals with proofs. “Everyfact of mathematics must be ensconced in an axiomatic theo-ry and formally proved if it is to be accepted as true.”

In contrast, “In its �rst of its lives, philosophy sets itself the taskof telling us how to look at the world. … Philosophical de-scription make us aware of phenomena that lie at the otherend of the spectrum of rationality that science will not andcannot deal with.” Then “In its the second life, philosophy,like mathematics, relies of method of argumentation thatseems to follow the rules of some logic.”

But philosophy “has not been quite as comfortable with itsdouble life,” Rota wrote. In its �rst, philosophy “sets itself thetask of telling us how to look at the world … making us aware ofphenomena that lie at the other end of the spectrum of ratio-nality that science will not and cannot deal with. The assertionsof philosophy are less reliable than assertions of mathematicsbut they run deeper into the roots of our existence. Philosophi-cal assertions of today will be the common sense of tomorrow.”

Axiom 0: Goddess ReasonIn its second life, “philosophy, like mathematics, relies on amethod of argumentation that seems to follow the rules ofsome logic,” but – unlike mathematics – the rules have “neverbeen clearly agreed upon by philosophers, and much of the

philosophical discussion since its Greek beginnings has beenspent on method,” Rota wrote. “Philosophy’s relationship withGoddess Reason is closer to a forced cohabitation than to theromantic liaison which has always existed between GoddessReason and mathematics.”

Are we to believe that Professor Kay would issue a restrainingorder to keep Goddess Reason from darkening computerscience’s door?

Axiom 1: Philosophical disclosures are metwith anger that we reserve for the betrayalof our family secretsRota: “Philosophical arguments are emotion-laden to a great-er degree than mathematical arguments and written in a stylemore reminiscent of a shameful admission than of a dispas-sionate description. Behind every question of philosophythere lurks a gnarl of unacknowledged emotional cravingswhich act as a powerful motivation for conclusions in whichreason plays at best a supporting role. To bring such hiddenemotional cravings out into the open, as philosophers havefelt their duty to do, is to ask for trouble. Philosophical disclo-sures are frequently met with anger that we reserve for thebetrayal of our family secrets.”

When Jonathan Edwards, a research fellow with the SoftwareDesign Group at MIT, was asked to contribute a chapter toBeautiful Code, published in 2007 by O’Reilly, he declined.

“Beauty is an idealistic fantasy,” he later explained on the blog Alarm-ing Development [11] “I hope that someday we will discover suchprinciples. But in the meantime software design is still a matter ofwisdom, not knowledge, and is therefore largely unteachable.”

He con�ded: “I am having trouble with this assignment. Telling aninspiring story about a beautiful design feels disingenuous. Yes, we allstrive for beautiful code. But that is not what a talented young pro-grammer needs to hear.”

Then, as if betraying a family secret, he wrote: “I wish someone hadinstead warned me that programming is a desperate losing battleagainst the unconquerable complexity of code, and the treachery ofrequirements. I can’t teach you how to design beautiful code, because Idon’t know how myself. I may have managed to get a few thingsalmost right. … A lesson I have learned the hard way is that wearen’t smart enough. … and above all to prize simplicity. Anotherlesson I have learned is to distrust beauty. It seems that infatuation

10E18342 text_012 Page 12 29-APR-10



Research

with a design inevitably leads to heartbreak, as overlooked uglyrealities intrude. Love is blind, but computers aren’t. A long-termrelationship – maintaining a system for years – teaches one to appreci-ate more domestic virtues, such as straightforwardness and conven-tionality. Beauty is an idealistic fantasy: what really matters is thequality of the never-ending conversation between programmer andcode, as each learns from and adapts to the other. Beauty is not asuf�cient basis for a happy marriage.”

Axiom 2: Dictatorship of De�nitivenessRota: “Philosophers in this century have suffered more thanever from the dictatorship of de�nitiveness. The illusion ofthe �nal answer, what two thousand years of Western philoso-phy failed to accomplish.”

Axiom 3: Inevitability of FailureRota: “A dispassionate look at the history of philosophy disclos-es two contradictory features: �rst, these problems [of philoso-phy] have in no way been solved, nor are they likely to besolved as long as philosophy survives; and second every philos-opher who has ever worked on any of these problems hasproposed his own ‘de�nite solution,’ which has invariantlybeen rejected by his successors. … Philosophers of the pasthave repeatedly stressed the essential role of failure in philoso-phy. The failure of philosophers to reach any kind of agree-ment does not make their writings any less relevant to theproblems of our day. We reread with interest the mutuallycontradictory theories of mind that Plato, Aristotle, Kant andComte have bequeathed to us, and �nd their opinions timelyand enlightening, even in problems of arti�cial intelligence.”

Axiom 4: The Myth of PrecisionRota: “The prejudice that a concept must be precisely de�nedin order to be meaningful, or that an argument must be pre-cisely stated in order to make sense, is one of the most insidi-ous of the twentieth century. … Looked from the vantage pointof ordinary experience, the ideal of precision seems preposter-ous. Our everyday reasoning is not precise, yet it is effective.Nature itself, from the cosmos to the gene, is approximate andinaccurate. … The ideal of precision in philosophy has its rootsin a misunderstanding of the notion of rigor.”

We misunderstand the concepts of philosophy if we force themto be precise. One insightful metaphor due to Wittgenstein isthat philosophical concepts are like the winding streets of an oldcity, which we must accept as they are, and which we must famil-iarize ourselves with by strolling through them while admiringtheir historical heritage. [12]

Axiom 5: Appeal to PsychologyRota: “ … to justify their neglect of most of the old and substan-tial question of philosophy [they argue] that many questionsformerly thought to be philosophical are instead ‘purely psycho-logical.” … Experimental psychology, neurophysiology, andcomputer science may turn out to be the best friends of tradi-tional philosophy. The awesome complexities of the phenome-na that are being studied in these sciences have convincedscientists (well in advance of philosophical establishment) thatprogress in science will depend on philosophical research in themost classical vein.”

“And if I have to describe the in�uence PL/1 can have on its users, theclosest metaphor that comes to my mind is that of a drug. I rememberfrom a symposium on higher-level programming language a lecturegiven in defense of PL/1 by a man who described himself as one of itsdevoted users. But within a one-hour lecture in praise of PL/1 hemanaged to ask for the addition of about �fty new "features," littlesupposing that the main source of his problems could very well be thatit contained already far too many "features." The speaker displayedall the depressing symptoms of addiction, reduced as he was to thestate of mental stagnation in which he could only ask for more, more,more. ... When FORTRAN has been called an infantile disorder, fullPL/1, with its growth characteristics of a dangerous tumor, could turnout to be a fatal disease.” – Dijkstra [13]

Axiom 6: The Illusion of De�nitivenessRota: “The results of mathematics are de�nitive. No one willever improve on a sorting algorithm which has been provedbest possible. … Mathematics is forever. … The problems ofphilosophy are the least likely to have ‘solutions.’ ... The realitywe live in is constituted by a myriad contradictions, whichtraditional philosophy has taken pains to describe with coura-geous realism. But contradiction cannot be confronted byminds who have put all their eggs in the basket of precision andde�nitiveness. The real world is �lled with absences, absurdities,abnormalities, aberrances, abominations, abuses, with Abgrund.”

A Critic’s DilemmaIn my previous Conduit article about Dijkstra (part 1) [14], Iderived (with a slightly different notation) a “criticism equa-tion”: E=mc2 In the equation, the impressionistic quantitiesare: “E” is the “energy” of criticism, “m” is the “substance” ofthe critical message, and “c” the “authority” of the critic. Well,Kay’s Turing award surely quali�es him for authority, but his mis so small that he got no E at all. There are two obvious desid-erata for a conscious critic that creates a dilemma for her asthey are somewhat in con�ict. The �rst axiom, “non-dema-gogy,” says that you should be critical only in areas where youhave signi�cant and recognized achievements. The secondaxiom, “non-personal,” says that the critique is more effectivewhen it stays away from the personal biases of the critic. Thehigher your achievement in an area, the tougher your criticismcould be of lesser achievers, so you clearly are biased in yourcritique towards your kind, failing to satisfy the second axiom.The Spartan criticism of Dijkstra, hard to take by many, is atthe root of the “disabling prejudices.” He lived a Spartan life,holding himself �rst at the same high standards that he usedto critique others’ shortcomings.

“As a scientist Dijkstra was a model of honesty and integrity.Most of his publications were written by him alone. The fewpublications that he wrote jointly with his colleagues bear theunmistakable trait of his writing style. He never had a secretaryand took care of all his correspondence alone. He neversought funds in the form of grants or consulting and neverlent his name to the initiatives to which he would not contrib-ute in a substantial way. When colleagues prepared a Festschriftfor his sixtieth birthday, published by Springer-Verlag, he tookthe trouble to thank each of the 61 contributors separately, ina hand-written letter. His supreme self-con�dence went

10E18342 text_013 Page 13 29-APR-10


Research


FEARIn 1984, soon after my family �ed to the United States, Ibegan telling my friends about life in communist Romania– hard-to-believe stories from a weird world of dictatorshipand limited freedom. I told them about the Carpathianpresident who forced people to attend rallies and chant hisname. I described a government composed entirely of theCarpathian’s family members; long lines at supermarketswhose shelves were barely stocked; mile-long lines of carswaiting a turn at the gas pump; spending the night in line sothat the next morning we could claim our two-bottle ration ofmilk – available only to parents with young children.

This crazy dictator had a wife who, though only a labtechnician in a pharmaceutical research institute, somehowmanaged to earn her Ph.D. in chemistry in just six months.What an achievement, the newspapers declared. She wasimmediately promoted to director of the pharmaceuticalresearch institute, where she was listed as co-author -- and�rst author – of the hundreds of papers published annually bythe institute. But just in case this did not do justice to herleadership, her name was listed in bold-faced type that wasdouble the size of the other authors.

How can one not promote such a talent to a position alignedwell with her stature? It may come as no surprise that thedictator’s wife was appointed Minister of Science. Newspa-pers celebrated the achievement with patriotic pride, andnoted that no one had anything critical to say about theappointment, at least on the record.

Our new friends had dif�culty believing our stories. Why wasno one brave enough to stand up to the regime, theywondered? Such a thing could never happen here in theUnited States, they said. Could it?

It is hard to talk about fear. I wished for a metaphor, a storyto offer my friends as a way to defend my seemingly fearfulcompatriots. We dealt with the daily problems by tellingpolitical jokes – oh, the safe haven of artistic ambiguity! –

a folkloric form of proli�c and creative protest that occasion-ally got some of us called up to a certain of�ce where wewere told that “on so and so day you told a joke about … weare concerned about you …”

One thing about dictators: Their time for justice comes. ThatCarpathian dictator and his wife were executed in therevolution. It is said that the members of the executionplatoon could not restrain themselves when they marched toexecute the couple; some starting shooting before reachingthe wall.

They say that patriotism is the last refugeTo which a scoundrel clingsSteal a little an' they throw you in jailSteal a lot an' they make you kingThere's only one step down from here, babeIt's called the land of permanent blissWhat's a sweetheart like you doin'in a dump like this.

Bob Dylan [16]

“Funding in genomics is measured in nano-Landers!” was acolleague’s attempt at survival humor at a recent computa-tional biology conference. The similarity with the nano-Dijk-stras quote is only that. These two cannot be more opposite! Iam afraid that just about now, the nano could become pico.

Afraid? I guess there are many types of fear. Fear of losingfreedom scars you. It wasn’t until 2002 that I recognizedsomething similar in the United States. I was working atCelera Genomics, and from time to time we would receive insecret a message from a genomic scientist of stature – anapology for the actions of some of his colleagues. Clearlygenomics people were afraid to say positive things in publicabout Celera for fear of losing their research freedom. Irecognized this fear from the experiences in Romaniathat I was trying to forget. This time, though, I was in thecommunist republic of genomics.

together with a remarkably modest lifestyle, to the point ofbeing spartan. His and his wife’s house in Nuenen is simple,small and unassuming. He did not own a TV, a VCR or amobile telephone, and did not go to the movies. In contrast,he played the piano remarkably well” [15].

As I put the �nal touches on this article, the recent announce-ment of Grigory Perelman’s solution of the Poincare Conjecturemarked another illustration that we must work on what we love.His is a victory of doers over talkers, a victory of the deep theoryscientists over craftsmen. We should all celebrate Perelman’sachievement and his receipt of the $1 million Clay Mathemati-cal Millennium Prize.

“The question ‘What is Mathematics?’ is as unavoidable and asunanswerable as the question ‘What is Life?’ In actual fact Ithink it’s almost the same question.” Dijkstra [15]

Dijkstra’s Axioms:Axiom 0: Life = MathematicsAxiom 1: Computer Science = Mathematics + Murphy’sLawAxiom 2: Beauty is Our BusinessAxiom 3: Simplicity is a prerequisite for reliability

10E18342 text_014 Page 14 29-APR-10



Research

Father-in-law vs. PajamasI sent a silly game of my own to Professor Dijkstrabut it has never been published before. I amincluding it here asking for algorithmic solutions, andoffer a prize for their optimal algorithmic solutions,but they have to be “in your head” solutions. ThePrize, like the ones I use for my students solvingthe most dif�cult parts of the extra credit homework,a slice of Providence’s famous, Pastiche Fruit Pie.Write to me at and I will publishwinners in the next issue of the Lighthouses.

A Silly GameA story. A young man lives withhis father-in-law, a very activeretired man. Among the dutiesof the father-in-law at home wasto wash dirty laundry. It followsthat he washes the son-in-law'spajamas too. The son-in-law’s Nclean pajamas, all different, are stored on a certainshelf S, in a white closet. They are arranged, as usual,in one stack as shown in Fig.1 stack S of pajamasLet B be the basket where the used pajamas aredeposited in order to be cleaned. The son-in-law is avery absent-minded young man, and when he changeshis pajamas, he acts as follows: throws the usedpajamas in B, and puts on the pajamas from the topof S. As already mentioned, the father-in-law is veryactive, so he puts the laundry in the washing machineas soon as they occur in B. That is, till the washing

moment no more than one pajamas has time toappear in B. After washing, the pajamas are returnedimmediately to the top of S.

After some time (say, years), at a moment whenthe son-in-law comes to change his pajamas, hediscovers a very strange thing: the pajamas he iswearing and the one from the top of S are extremelyworn out, while the rest of pajamas are almost new!Then he understood that all the time he has beenwearing these two sets of pajamas. The explanationcan be obtained as follows: while the son-in-lawwas wearing pajama 1 the father-in-law quicklywashed pajama 2 and placed it on top of S. Whenhe changed the pajamas, he took pajama 2 and putpajama 1 in B which ended up on top of S, and soon. The son-in-law has been wearing the sequence ofpajamas 1,2,1,2,1,2 … Our �rst problem is how toavoid this unfair wearing of pajamas. Let us de�ne afair wearing of the N pajamas to be a sequence of Npajamas that is any permutation of them.

The Problem. Give an algorithm (if possible thesimplest; solvable in your head) for a fair wearing ofthe pajamas. There is a caveat (inspired by the real-lifesituation): no communication between son-in-law andfather-in-law should be required.

The story, continued. The son-in-law discovered theoptimal algorithm for Problem 1. He started using it,so everything seemed to be okay. However, a happyevent brought some changes. A son was born. Amongthe reorganizations involved in the house was the oneconcerning the clothes. Now, the son-in-law’s pajamaswere assigned to a smallshelf. The pajamas werenow arranged in severalstacks, say M stacks ofmaximum height K.

Concerning washing, the father-in-law acts now… non-deterministically. He returns the washed pajamas tothe top of any stack he wishes. The nondeterministicreturn proves troublesome. A new algorithm is neededfor the problem. Find the two optimal algorithms forthe two versions of the problem.

References1. E.W. Dijkstra, EWD 1213: “Introducing a Course on Calculi” (1995)

2. E.W. Dijkstra, EWD 1036, “On the Cruelty of Really TeachingComputing Science” (1988)

3. E.W. Dijkstra, EWD 498, “How Do We Tell Truths that Might Hurt”(1975)

4. E.W. Dijkstra, EWD 1166, “From My Life (Written Because PeopleAsk Me for These Data)” (1993)

5. E. W. Dijkstra, EWD 666, “A Problem Solved in My Head” (undated)

6. Transcript of YouTube video athttp://www.youtube.com/watch?v=s7ROTJKkhuI

7. E.W. Dijkstra, EWD 1227, “A Somewhat Open Letter to David Dries”(1995)

8. John von Neumann, “The Mathematician,” published in Works of theMind Vol. 1, No. 1 (University of Chicago Press, Chicago, 1947)

9. E. W. Dijkstra, EWD 480, “Craftsman or Scientist?” (1975)

10. G.C. Rota, “The Pernicious In�uence of Mathematics uponPhilosophy” (1991)

11. From J. Edwards’ blog, Alarming Development,http://alarmingdevelopment.org

12. E. W. Dijkstra, EWD 340, “The Humble Programmer” (1972)

13. S. Istrail, “Storytelling About Lighthouses: Criticizing ProfessorDijkstra Considered Harmless,” Conduit, v.17 no. 2. (2009)

14. K.R. Apt, “Edsger Wybe Dijkstra (1930–2002): A Portrait of aGenius,” Formal Aspects of Computing (2002)

15. E.W. Dijkstra, EWD 720, “Why Correctness Must Be a MathematicalConcern” (1979)

16. B. Dylan, “Sweetheart Like You”

10E18342 text_015 Page 15 29-APR-10



Faculty

Q&A with James HaysHow did you �rst become interested incomputer science?

While I was growing up, personal computerswere rare. However, video game consoles werenot. The endless possibilities of these interactivevirtual worlds intrigued me from a young age.One of my �rst signi�cant software developmentattempts was a random map generator for aCivilization-style game (I still think this is aninteresting problem – speculative geology).I’ve also always been inspired by visions of thefuture, such as Masamune Shirow’s Ghost in theShell series, in which the lines between humansand computers, and virtual and real worldsbecome blurred. Computing and its manyapplications will shape our society dramaticallythis century. I knew I wanted to be a part of that.

How do you pick your research problems?

I think the massive amount of data on the Inter-net offers great potential for advancing com-puter vision and related applications in com-puter graphics. Some of my research ideas are

driven by the data itself – e.g.�nding a source or type ofdata that nobody else hasutilized such as photo geotags.I’m also interested in attackingthe fundamental problems inimage understanding byleveraging Internet datasources. I’m fairly �exible inmy research pursuits, though.

What do you consider the most interesting andexciting challenges of your research?

Many of the challenges I face are similar to otherimaging researchers– �nding better features todescribe visual content, �nding ef�cient andeffective machine learning algorithms. But myresearch tends to emphasize data more thanalgorithms. It is challenging to �nd ways toleverage in-the-wild Internet data that is plentifulbut often unstructured and dif�cult to use.

Do you have a favorite project that you’veworked on?

My favorite project is probably “Scene Comple-tion” in which image holes are patched up withdata from matching Internet photos. It’s a veryclear illustration of the power of large data sets

– the algorithm is simple, and seems ineffectivewith small amounts of training data, but it seemsvery intelligent with millions of training examples.

How do you see your �eld evolving overthe course of your career?

Computer vision is going through the same tran-sition that computer graphics went through inthe past two decades – the transition fromlimited research successes to widespreadadoption. Computer graphics is such a successthat people can scarcely distinguish real andvirtual imagery, and most computers sold todayinclude graphics-speci�c hardware. On theother hand, many vision researchers think thatimage understanding is an AI-hard problem thatwe are decades away from solving. I tend todisagree. There are AI-hard questions you canask about images, but I think many interestingcomputer vision problems are solvable in thenear future to a degree that will allow otherimportant �elds, such as robotics and graphics,to advance signi�cantly.

What’s the “next big thing” in computer vision?

That’s a dif�cult question. I’m not sure there isa single “next big thing” in computer vision. Itend to push on research problems related todata and scale, but other researchers are makingsigni�cant gains by exploring better features,better learning, or better imaging.

Scene understanding will remain a core prob-lem in computer vision, just as photorealisticrendering is a core problem of computergraphics, but as the �eld matures we are likely tosee entirely new research problems emerge.Interactivity and privacy could be more signi�-cant concerns.

If you had enough extra time to study oneadditional area, what would it be?

I have very broad interests in history and scienceand I tend to follow up on them. Wikipedia is anamazing resource, but it only gets you so far. It’shard for me to pick just one area that I wouldwant to study more! That’s one of the fantasticthings about computing – it’s relevant toeverybody, so it gives you an avenue to connectwith any research area.

The Faculty Speak Out

By James Hays

“Computing and its manyapplications will shape oursociety dramatically thiscentury. I knew I wanted tobe a part of that.”

10E18342 text_016 Page 16 29-APR-10



Faculty Notes

Michael BlackMichael Black and his collaborators re-ceived new funding from the NIH pro-gram for Exceptional, UnconventionalResearch Enabling Knowledge Accelera-tion as well as from the NSF program onCollaborative Research in ComputationalNeuroscience. Michael also received agenerous research gift from Intel and hisgroup obtained a full-body 3D laserrange scanner on loan from the Army. Inan effort to keep the airline industry inbusiness, Michael gave invited talks inSicily, Prague and Barcelona and attend-ed conferences in Kyoto and Vancouver.

Roger BlumbergRoger just completed a two-year termas Chairman of the Board of Directors,at the Rhode Island Council for the Hu-manities (RICH).

During his term, the Council madegrant-making its primary focus and thisyear RICH will make grants to humani-ties projects and programs in RI totalingapproximately $300K. In November,Roger was elected to the Board of theFederation of State Humanities Coun-cils (FSHC) in Washington, which is themembership organization that repre-sents the 56 state and territorial hu-manities councils supported by theNational Endowment for the Humani-ties. During his tenure at the FSHC, he’llcontinue to promote the idea that com-puting should play a greater role in theway the councils think about their work,their constituencies, and the meaningof the humanities in the 21st century.

Rodrigo FonsecaNew faculty member Rodrigo and hiswife Paula welcomed their daughter So-�a on December 31, 2009.

Philip KleinKlein continues to teach The Matrix inComputer Science, a course that teach-es linear algebra via CS applications.He has received funding from Google tosupport work on algorithms related toroad maps.

Sorin IstrailSorin gave a Keynote Lecture in Janu-ary at the Asia-Paci�c BioinformaticsConference in Bangalore, at the TataInstitute in India. He then lectured atthe Indian Institute of Technology inChennai. In India, he was most im-pressed by the structure of the drivingtraf�c which he calls “liquid-traf�c,”and the population substructure (sub-populations speaking different dia-lects as well as genomic substructure)with an incredible 3400 such subpop-ulations. He published, together withhis former Brown postdoctoral fellowFumei Lam, in the Chinese journal“Communication and Information inSystems,” the paper “CombinatorialAlgorithms for Protein Folding in Lat-tice Models: A Survey of MathematicalResults” in a special issue of the jour-nal dedicated to the 67th birthday an-niversary of Mike Waterman. He alsoco-authored with his Ph.D. studentRyan Tarpine and Eric Davidson ofCalifornia Institute of Technology a pa-per in the Proceedings of the NationalAcademy of Sciences. The computa-tional component of the paper wasthe result of three years of work on thedevelopment of the cisGRN-Browser, agenome browser software system builtby Ryan for gene regulatory networks.In the area of genome-wide diseaseassociations, his Ph.D. student DerekAguiar developed algorithms for thedetection of loss of heterogeneity ap-plied to the GWAS data from the Mul-tiple Sclerosis Genetic Consortium.Derek’s algorithm detected a signa-ture of missing genomic pieces, char-acteristic to mental disease; a paperdescribing the algorithm is to be pre-sented at the RECOMB 2010 Confer-ence in Lisbon. Sorin’s two postdoc-toral fellows Austin Huang and AlperUzur have focused their work on HIV

and genomic drug resistance, and re-spectively, the minimum informativesubset of genetic variants (SNPs) forpreterm labor. Work of Allan Stewart(CS) and Kyle Schutter (Bioengineer-ing) is focused on two honor theses tobe �nished this Spring: one FCC Pro-tein Folding, and the other on Func-tional Regulatory Genomics Inference.

Sorin redesigned his graduate course,now called Medical Bioinformatics. ItsFall semester organization includedfour guest lecturers of distinction: for-mer Director of the National Cancer In-stitute, Sam Broder of Celera, mathe-matician John Conway, John vonNeumann, Distinguished Professor atPrinceton, Associate Director of Ge-nomics of the Food and Drug Aministra-tion, Issam Zineh, and Jonathan Yewd-ell, renowned immunologist at NIH. Forthe Course, Alper, also a talented car-toonist contributed with artistic depic-tions of Course lessons. Sir RonaldFisher inspired one of them.

Sorin is the Chair of the May 3–7, 2010Brown University/CCMB Symposium,organized in two parts.

May 3–4, 2010: John von NeumannDistinguished Lecture Series, organizedby Leon Cooper (Physics), Stuart Ge-man (Applied Mathematics), Sorin Is-trail (Computer Science), Roberto Ser-rano (Economics)

Marina von Neumann Whitman (Profes-sor of Business Administration and Pub-lic Policy, University of Michigan), Free-man Dyson (Professor Emeritus ofPhysics, Institute for Advanced Study,Princeton University), John Conway(John von Neumann Professor of Math-ematics, Princeton University), KennethArrow (Nobel Laureate, Joan KenneyProfessor of Economics and Professorof Operations Research Emeritus, Stan-ford University), Richard Karp (Universi-ty Professor, University of California,Berkeley), David Mumford (UniversityProfessor, Department of Applied Math-ematics, Brown University), Leon Cooper(Nobel Laureate, Thomas J. Watson Pro-fessor of Science and Director, Institutefor Brain and Neural Systems, Brown

University), George Dyson (Historian ofscience, University of Washington)

May 4–7, 2010: The Genome and theComputational Sciences: The NextParadigms

Eric H. Davidson (Norman ChandlerProfessor of Cell Biology, California In-stitute of Technology), Andrew G. Clark(Jacob Gould Schurmn Professor ofPopulation Genetics, Department ofMolecular Biology and Genetics, Cor-nell University), J. Craig Venter (Found-er, Chairman and President, J. CraigVenter Institute (tentative)), Ellen V.Rothenberg (Professor of Biology, Cali-fornia Institute of Technology), RichardLewontin (Alexander Agassiz ResearchProfessor, Museum of Comparative Zo-ology, Harvard University), David Bot-stein (Anthony B. Evnin Professor ofGenomics and Director, Lewis-SiglerInstitute, Princeton University), MartinMeier-Schellersheim (Team Leader,Computational Biology Group, NationalInstitute of Allergy and Infectious Dis-eases, NIH), Sean Eddy (Group Leader,Janelia Farm, Howard Hughes MedicalInstitute), Martha Bulyk (AssociateProfessor of Medicine and Pathology,Harvard Medical School), Jonathan W.Yewdell (Chief, Biology Section, Labo-ratory of Viral Diseases, National Insti-tute of Allergy and Infectious Diseases,NIH), David Shaw (Ph.D., Chief Scien-tist and Founder, D.E. Shaw Research),Marcus Feldman (Professor of Biology,Stanford University)

Barb MeierBarb continues to teach computer ani-mation production courses. Last sum-mer, she attended a workshop on Sto-rytelling for Educators with two-timeGrammy winner and nationally re-nowned storyteller/singer Bill Harley atthe Pendle Hill retreat center outside

10E18342 text_017 Page 17 29-APR-10



Faculty Notes

Philadelphia. The attendees taught alllevels of school from kindergartenthrough high school, college, gradschool, and adult education. The com-mon thread was that, at any level, les-sons are more memorable when toldas stories. Beyond this, Barb learnedways of helping students create betterstories and then applied this to heranimation class this past fall. Iteratingthrough several drafts of scripts andstoryboards, her students createdmore complex stories with good end-ings - the holy grail of storytelling - fortheir animations. As long as they aregoing to pull several all-nighters tomake their animations, they may aswell be telling good stories! Bill Harleyis based in the Providence area, butperforms nationwide in family andadult concerts and at storytelling festi-vals. Check him out if he’s in your area.You’ll go home entertained and wiser!

Meinolf SellmannMeinolf’s 2009 started with a sabbati-cal leave which he used to visit col-leagues all over Europe and theStates. Among others, Meinolf gavetalks at the Max Planck Institute forInformatics in Saarbruecken, EcolePolytechnique Federale de Lausanne,Oxford University, University of St. An-drews, the Cork Constraint Computa-tion Centre, University of Waterloo,Georigia Tech, and the Sandia Nation-al Laboratories in Albuquerque.

Meinolf published a paper at Con-straints, two papers at CPAIOR, four atCP, and two at ICTAI. The latter two bothreceived best paper nominations,whereby one was co-authored by Mei-nolf’s graduate student Yuri Malitsky. Incollaboration with Chris Jefferson fromOxford University, Meinolf was invitedto give a tutorial on “Amortized andExpected Case Complexity for FilteringAlgorithms” at CP in Lisbon.

Meinolf is excited that his researchhas been incorporated in variouscombinatorial solvers last year. IBMhas integrated Meinolf’s and graduatestudent Serdar Kadioglu’s work on bi-ased binary search in their

optimization software. The Jacob CPsolver developed at EPFL now con-tains the ef�cient knapsack �lteringalgorithm that Meinolf co-developedin 2007. Lufthansa Systems Berlin in-vited Meinolf to talk about grammarconstraints which Meinolf invented in2006 and which they want to inte-grate into their crew scheduling sys-tem. Grammar constraints are alsobeing used by his colleagues in Mon-treal to solve real-world nurse shiftscheduling problems ef�ciently.

Finally, Meinolf secured an internation-al collaboration grant supplement fromNSF and is looking forward to visitinghis colleagues in Sweden this summer.

Andy van DamIn early December Andy van Dam at-tended the retirement of his goodfriend and fellow graphics old-timerJose Encarnacao, from the TechnicalUniversity Darmstadt (Germany) whereAndy got an honorary degree in 1995.He gave two talks on multi-touch com-puting in honor of Jose, one at the TUand the other at the nearby FraunhoferInstitute for Research in ComputerGraphics, the world’s largest graphicsresearch lab which Jose had created inthe ’90s and headed until a few yearsago, and which had a branch for nearlyten years in Providence. Andy chairedthe Technical Advisory Board of theProvidence branch during that time.

On that same trip, he, his undergradu-ate research assistant Donnie Kend-all, Brown’s head Librarian HarrietteHemmassi, and Prof. Massimo Riva,an expert in Italian Studies, met withcolleagues at the British Library to dis-cuss deployment of the “Garibaldi onthe Surface” project in a public exhibi-tion on digital research that the BritishLibrary will mount in the Fall of 2010.The project was started under Micro-soft Research sponsorship at Brown toallow interactive browsing of theGaribaldi Panorama on the MicrosoftSurface, a coffee-table sized displaythat allows multiple users to interactvia multi-touch and pen. The panora-ma is a 273 foot by four and half footlong scroll of individual paintedscenes, depicting the life and times ofthe great Italian hero that Brown hadpreviously digitized but couldn’t reallydisplay properly. This work will be thefeatured application on the Surfaceduring the approximately six monthsthe exhibition will run, and this springthe team is working on �eshing outboth the functionality and user inter-face of the browsing application andthe hypermedia content of the Garib-aldi web. The project parallels workdone on “tangible math”, manipulat-ing mathematical equations on theSurface, again using multi-touch andpen. Also good progress has beenmade on a Google-sponsored re-search project to probe the utility ofgestures for command on the Androidplatform. There currently are a groupof students including undergraduatesFerdi Adeputra, Gal Peleg, Ali Ozler,Donnie Kendall, masters student Hsu-Sheng Ko, a Ph.D. student (AndrewBragdon, leading the Google project)and a full-time researcher (Director ofResearch, Bob Zeleznik) working onthese projects.

John Savage shown with 8 of the 10Jefferson Science Fellows in front of theWhite House

Recent Ph.D.s

Alex Balan

Eric Rachlin

Will Headden

Yossi Lev

10E18342 text_018 Page 18 29-APR-10



Department Awards and Honors

Brown University/National Universityof Singapore Computational BiologyDegree ProgramBrown University and the National Universityof Singapore (NUS) have established a concur-rent computational biology degree program.Students who complete the program will receiveconcurrent degrees from NUS and Brown: abachelor’s degree in computational biologyfrom NUS and a master’s degree in computerscience with a special designation in computa-tional biology from Brown. Franco Preparata,who has been a visiting faculty at NUS forseveral years, is the primary architect of thisprogram at Brown. Tom Doeppner has alsobeen involved in setting up the program.

Brown President Ruth J. Simmons said, “Thesequencing of the human genome has openeda vast new area of research at the junction ofthe computing and biomedical sciences. Com-putational biology is growing at Brown. We areexcited by the possibilities of this new relation-ship with the National University of Singapore.”

Ben Raphael ReceivesSloan Research FellowshipBen Raphael was recently awarded the presti-gious Sloan Research Fellowship, one of theoldest and most competitive fellowship pro-grams in the United States.

Ben’s remarkable work developing novelcomputational and mathematical approaches toproblems in biology, especially cancer genomicsand comparative genomics, led to his inclusionin this elite group.

Selection procedures for the Sloan ResearchFellowships are designed to identify those whoshow the most outstanding promise for funda-mental contributions to new knowledge. “I amdeeply honored to be selected as a Sloan fellow,”said Ben. “During my career, I have beenfortunate to work in environments that promot-ed my research and to collaborate with excep-tional scientists from various disciplines. I ampleased that the Sloan Fellowship provides theopportunity to continue my ongoing work andto push my research in new directions.”

The fellowships are awarded by the Alfred P. SloanFoundation to honor and promote the science ofoutstanding researchers early in their academiccareers. The 118 winners are faculty members at55 colleges and universities in the United Statesand Canada who are conducting research at thefrontiers of physics, chemistry, computational andevolutionary molecular biology, computer science,economics, mathematics and neuroscience. Theyreceive grants of $50,000 for a two-year period topursue whatever lines of inquiry are of mostinterest to them. The funds are planned to beused to support the work of Ben and his studentsin computational cancer genomics.

Aside from the monetary aspect of the fellow-ships, less tangible bene�ts have been cited byformer Sloan fellows. The early recognition ofdistinguished performance which the fellow-ships confer, after years of arduous preparation,was said to be immensely encouraging and astimulus to personal and career development.

The Sloan Research Fellowships have beenawarded since 1955. Past recipients have goneon to win 38 Nobel prizes, 14 Fields Medals(mathematics), and eight John Bates Clarkawards (economics).

Besides his Sloan Fellowship, Ben has previouslyreceived other recognition for his work, includ-ing a Career Award at the Scienti�c Interfacefrom the Burroughs Wellcome Fund, anoutstanding accomplishment.

Ph.D. program inComputational Molecular BiologyThe Center for Computational MolecularBiology and the Department of ComputerScience are delighted to announce the newPh.D. program in Computational MolecularBiology. Applications for admission into thisprogram are being accepted now. Currentcourses supporting the degree are in ComputerScience, Applied Mathematics, Mathematics,Biological Sciences, and Chemistry.

The Center for Computational MolecularBiology (CCMB) at Brown is a world-class centerfor research and scholarship in this new disci-pline. CCMB’s central mission is to makebreakthrough discoveries in the life sciences atthe molecular and cellular level through the


Professor Franco Preparata

Ben Raphael

10E18342 text_019 Page 19 29-APR-10




creative application of existing data-analyticmethods, and to develop the novel computation-al, mathematical, and statistical technologiesrequired to exploit the opportunities emergingfrom advances in genomics and proteomics.

The Center concentrates its efforts on a smallnumber of fundamental biological themes:algorithmic methods and statistical inference ingenomics, comparative genomics and evolution,gene regulatory networks, regulatory genomics,mathematical models of genetic variation, andcancer genomics. They focus on building trulyinterdisciplinary research teams with othermulti- and transdisciplinary centers at Brown.Since sophisticated computational, mathemati-cal and statistical technologies are imperative forthe success of the center’s goals, technologieshave been developed in areas of fundamentalbiological themes with the aim of generatingtestable predictions, capitalizing on expertise inthese methodologies in the departments ofApplied Mathematics and Computer Science.

Faculty members currently associated withCCMB are Alexander Brodsky, Assistant Profes-sor of Medical Science; William Fairbrother,Assistant Professor of Biology; Sorin Istrail,Director; Julie Nguyen Brown Professor ofComputer Science; Charles Lawrence, Director2004-2006; Professor of Applied Mathematics;Franco Preparata, An Wang Professor ofComputer Science; David Rand, Professor ofBiology; Sohini Ramachandran, AssistantProfessor of Ecology and Evolutionary Biology(to join Brown in July 2010); Benjamin Raphael,Assistant Professor of Computer Science;William Suggs, Associate Professor of Chemistryand Biochemistry; Daniel Weinreich, AssistantProfessor of Biology; and Zhijin (Jean) Wu,Assistant Professor of Medical Science.

Tom Dean Named ACM FellowThe Association for Computing Machinery(ACM) recently elevated Tom Dean to Fellow forhis development of dynamic Bayes networks andanytime algorithms.

Tom is currently a staff research scientist atGoogle, Mountain View, and an AdjunctProfessor of Computer Science here at Brown.He is known in AI for his work on the role ofprediction in planning, control and decision-making where uncertainty and the limited timeavailable for deliberation complicate theproblem, particularly his work on temporalgraphical models and their application insolving robotics and decision-support prob-lems. His temporal Bayesian networks, latercalled dynamic Bayes networks, made itpossible to factor very large state spaces andtheir corresponding transition probabilitiesinto compact representations, using the toolsand theory of graphical models. He was the�rst to apply factored Markov decision process-es to robotics and, in particular, to the problemof simultaneous localization and map building(SLAM). Faced with the need to solve whatwere essentially intractable problems inreal-time, Dean coined the name “anytimealgorithm” to describe a class of approximateinference algorithms and the associated (meta)decision problem of deliberation scheduling toaddress the challenges of bounded-timedecision making. These methods have beenapplied to large-scale problems at NASA,Honeywell and elsewhere.

At Google, Tom has worked on extracting stablespatiotemporal features from video and devel-oped new, improved features for video under-standing, categorization and ranking. From1993 to 2007 he was Professor of ComputerScience and Cognitive and Linguistic Sciencesat Brown. He remains associated with theUniversity through his Adjunct Professorship.

Tom received his B.A. in mathematics fromVirginia Polytechnic Institute & State Universityin 1982 and his M.Sc. and Ph.D. in computerscience from Yale University in 1984 and 1986respectively. His research interests includeautomated planning and control, computation-al biology, machine learning, neural modeling,probabilistic inference, robotics and spatial andtemporal reasoning.

10E18342 text_020 Page 20 29-APR-10




Tom Dean

Tom was named a fellow of AAAI in 1994. Heserved as the Deputy Provost of Brown Universi-ty from 2003 to 2005, as the chair of Brown’sComputer Science Department from 1997 until2002, and as the Acting Vice President forComputing and Information Services from 2001until 2002. Tom was a founding member of theAcademic Alliance of the National Center forWomen and Information Technology and aformer member of the IJCAI Inc. Board ofTrustees. He has served on the ExecutiveCouncil of AAAI and the Computing ResearchAssociation Board of Directors. Tom was also arecipient of an NSF Presidential Young Investi-gator Award in 1989. Dean is co-author withMike Wellman of the Morgan-Kaufmann text en-titled Planning and Control which ties togethertechniques from arti�cial intelligence, opera-tions research, control theory, and the decisionsciences. He is co-author with James Allen andJohn Aloimonos of Arti�cial Intelligence:Theory and Practice, an introductory text inArti�cial Intelligence. His latest book, TalkingWith Computers is published by CambridgeUniversity Press and examines a wide range oftopics from digital logic and machine languageto arti�cial intelligence and searching the web.

Commenting on what he is currently workingon and excited about, Tom said, “More than ayear ago I became concerned that the featuresthat Google was using to catalog video were notgood at capturing the fundamental characteris-tics of biological motion. When you look at themost popular videos on YouTube, they invariablyinvolve biological motion, from sports andmusic videos to nature documentaries andmisbehaving-pet videos. The effort of taking thisinitial observation and translating it into a set ofalgorithms for �rst learning motion featuresand then extracting them from video in betterthan real time has been both satisfying andhumbling. The experience has also convincedme that a signi�cant fraction of the image, videoand audio data processing that we perform isbetter suited to a different model of computingthan what we currently support in the cloud,hence my recent work on exploiting many-corecomputing in a way that is agnostic regardingboth hardware and programming language. Formost of the time I was at Brown, I was the onlyperson working in robotics and decision making— regrettably, for much of the time that I

overlapped with Chad, Michael and LeslieKaelbling, I was up to my neck in administration.Now that I’m on the other side of the country,the department has hired two of the very best re-searchers working in the areas of computervision that I’m most excited about. I’m referringto Erik’s work on hierarchical graphical modelsand James’ work on exploiting very large imagecorpora. I really appreciate the departmentproviding me with even more incentives to visit.”

The ACM Fellows Program was established byCouncil in 1993 to recognize and honoroutstanding ACM members for their achieve-ments in computer science and informationtechnology and for their signi�cant contribu-tions to the mission of the ACM. The ACMFellows serve as distinguished colleagues towhom the ACM and its members look forguidance and leadership as the world ofinformation technology evolves.

Tom joins the department’s seven other ACMFellows: Maurice Herlihy, Franco Preparata,John Savage, Eli Upfal, Andy van Dam, PeterWegner and Stan Zdonik.

Flapjax Recognized with Best StudentPaper Award at OOPSLAA group of Brown students have worked forthree years on Flapjax, a new programminglanguage built atop JavaScript. Their workwas recognized with the Best Student Paperaward at The International Conference onObject Oriented Programming, Systems,Languages and Applications (OOPSLA 2009)in Orlando, Florida. The Flapjax group includesfour undergrads--Leo Meyerovich ’07 (currentlya third-year Ph.D. student at Berkeley), JacobBaskin ’08 (now at Google), Michael Greenberg’07 (now a third-year Ph.D. student at Universityof Pennsylvania), and Aleks Brom�eld ’08,Master’s ’09 (now at Microsoft) --and twoPh.D. students, Greg Cooper ’08 (now atGoogle) and Arjun Guha, all working withShriram Krishnamurthi.

10E18342 text_021 Page 21 29-APR-10



Department News and Happenings

People suffering from chemical dependency are sometimes fortunate to availthemselves of treatment in rehabilitation centers. Unfortunately, a one- tothree-month period of in-clinic treatment is only a precursor to, and in no wayguarantees, the real goal, which is recovery (the part that, in the case of celebrities,doesn’t make it onto nightly TV). Therefore, rehabilitation programs need tomonitor, evaluate, and encourage their patients on the road to recovery. Andone key to successful recovery is communication.

In spring 2009, our senior software engineering course (CSCI 1900) workedon the problem of automating communication with recovering patients. Weworked with a series of centers currently operating in Brazil (or Brasil, as theyspell it in Portuguese). Our contact there is John Burns, an American who hasbeen serving in various capacities (Peace Corps, etc.) in Brazil for severaldecades. John is now based in São Paulo.

About 25 years ago, John created Vila Serena, a network of chemical dependencytreatment centers around the country. Vila Serena has treated tens of thousands ofpatients during this time. Some of these centers kindly agreed to test our systemwith a group of recovering dependents.

The primary task for the system was to track patients through the recoveryprocess after in-patient care, which is known as continuing care, aftercare, or (inPortuguese) pós-tratamento. Before they are checked out, patients are made tocreate a recovery plan. These are often designed in conjunction with the familymembers, and signed in their presence, to aid recovery. A key part of the plan isto stay in constant (if brief) communication with the clinics.

Unfortunately, communication is not an easy problem. How do patients commu-nicate with the clinic? The Internet isn’t always easily available, especially in thedeveloping world (even people who have good connectivity at work might not athome). Phoning in is much easier, since virtually every patient owns a mobilephone, but calls are very resource-intensive for the clinic: handling hundreds ofcalls requires signi�cant investment in personnel; simply routing the calls to ananswering machine is not only impersonal, it still requires someone to listen tothe messages (in case someone has called in to report a personal crisis). For thisreason, though well aware of the value in such communication, John’s clinics(and others) have failed to implement it.

But these ubiquitous mobile phones have one other means of communication:SMS. This medium offers many advantages relative to phone calls. Messages canbe processed asynchronously without the negative connotation of an answeringmachine. Messages are delivered as text, so they can be eyeballed quickly.Messages are short, further reducing the time demand. And contemporaryphone users are comfortable writing short notes with SMS.

Our students worked with John and his staff to collect requirements for thesystem. We had enough students to split the class into two groups. Each groupwas given the same task—to build a system for Vila Serena—but was told to notcommunicate with the other, i.e., to behave like the other group was a corporatecompetitor. One reason for this was to see whether the groups would obtainsigni�cantly different requirements or even, for the same requirements, imple-ment very different solutions.

The two groups’ requirements were similar, but not identical. In part, this wasbecause of features—such as inter-patient communication—proposed by some ofthe group members, which were enthusiastically adopted by the Vila Serena staff.

Parenthetically SpeakingBy Shriram Krishnamurthi

IndustrialPartners Program

The primary goals of the Industrial Partners Program (IPP)are to exceed the expectations of our partner companies interms of recruiting and outreach; to provide resources andemployment opportunities to our students and to allow ourfaculty to engage in meaningful research collaborations.

The department wishes to thank our Industrial Partnersfor their support:

Premier PartnerAdobe

Af�liatesAppleFacebookGoogleGTECHMicrosoftNetAppOracleVMware

Start-Up SupporterLIPIX, Inc.

IndividualsPaul Edelman, Edelman & Associates

Robert Khoury, Worldwide Financial Industry RecruitingServices

For more information about the Industrial PartnersProgram, contact:

Amy TarboxProgram ManagerTelephone: (401) 863-7610

Maurice HerlihyProfessor and IPP DirectorTelephone: (401) 863-7600

To learn more about the IPP visit:http://www.cs.brown.edu/industry

10E18342 text_022 Page 22 29-APR-10




But even with fairly similar requirements, the twogroups proposed signi�cantly different solutions.

A minor point of difference was the choice ofimplementation technology (Java versus Rubyon Rails), dictated largely by what some of thestudents were most comfortable with or mostinterested in learning. A much more importantdifference was in how the teams chose tointerface between SMS and computer. Onegroup chose to buy a SIM card for their servermachine, effectively receiving and processingmessages locally. The other group investigatedand chose Internet-based SMS gateways, whichhandle the SIM details and offer an API forobtaining the messages. These two solutionshave numerous trade-offs (which should beevident to the reader), and at different points inthe semester, different solutions appeared tohave the upper hand. In the end, however, Ibelieve both groups felt the gateway-basedsolution was better.

These differences demonstrate that having thetwo teams work on the problem independentlywas not merely a cosmetic detail. Their diver-gence helped the class understand two verydifferent technologies in some detail, therebybetter informing them when they encounter asimilar problem in the future.

The system built by the students collected theSMS messages and collated them against thepatients’ identities. It compared the frequencyof response against what was set out in their

“contracts” at discharge, notifying clinic staffwhen a patient was off schedule. It offered anintentionally simple, usable “inbox” interfacethat clinic staff would use to check messages.The systems could also send messages to patients,such as a periodic question or “message of theday”. Most of all, staff could �ag messages thatsuggested a need for more personal follow-up.(It is important to note that at no point did thestudent-built software make any clinical decisions.Instead, it merely mimicked and automatedwhat would have been done by a horde ofhumans, per their speci�cation.)

The good news is that both groups successfullycompleted their systems, and both were de-ployed in Brazil. Along the way, I believe thegroups learned a good deal about working withreal clients, managing privacy, dealing withnon-technical users, interfacing between theInternet and SMS, using the phone (rather thana computer terminal) as the primary communi-cation device, supporting internationalization,designing user interfaces, and adapting toforeign cultures. Twice, for instance, communi-cation from Brazil seemed to halt entirely, to thestudents’ surprise. Any guesses why? (Read on.)

There is growing interest in the use of SMS inmedicine; for instance, a BBC article

http://news.bbc.co.uk/2/hi/uk_news/maga-zine/7858425.stm

describes the use of SMS communication tohelp people suffering from seasonal affectivedisorder. The system built in this class falls in asimilar vein. Indeed, this system appears to havecommercial potential, and some students havebeen in conversation with healthcare organiza-tions about adaptations of it.

CreditsAdam Emrich and Hamzah Ansari, both of whowere graduate students in Brown’s PRIME(Innovation Management and EntrepreneurshipEngineering) program, conceived of this projectand identi�ed the Brazilian clinic. The staff atVila Serena was generous with both time andknowledge. John Burns, in particular, wasprofoundly generous with his time and ideas,putting extraordinary effort into the project. Innumerous ways, it would have been impossiblewithout him.

Oh yeah, the two interruptions. One was Easter;the other, Carnival!

10E18342 text_023 Page 23 29-APR-10




Around the Department

Pascal Van Hentenryck Cha Chasin Dancing with the ProfsPascal was one of six professors who tookto the stage in a packed Alumnae Hall for

“Dancing with the Profs” on February 12, 2010.

The event, now in its third year, was created bythe Brown Ballroom Dance Team, who prac-ticed with the six professors for three months,and then competed with a 90-second routine,each choosing a different dance.

Pascal, in a fedora and white socks, and LanShiow Tsai ’10 danced the cha-cha to MichaelJackson’s “Billie Jean.” Judge Julie Strandbergpraised the pair’s “creativity and wonderfulsense of humor.” Although Pascal and Landidn’t win, the dance-off was still a CS victorywith Daniel Hackney ’12 stealing the show withhis spirited jive with partner, Assistant Professorof Sociology Nancy Luke.

HalloweenThe department celebrated Halloween withthe annual party, including a costume contestand pumpkin carving.

10E18342 text_024 Page 24 29-APR-10




Bootstrap at BrownStudents enrolled in project Bootstrap, acurriculum for middle-school students thatteaches them programming through images andanimations, gather in the CIT several times perweek and are taught by Brown CS students.Bootstrap uses algebra as the vehicle forcreating this imaginative content, resulting inmuch greater student engagement in subse-quent math classes. Historically, Bootstrapattendees have been predominantly minorityand economically disadvantaged, with about aquarter female. Shriram Krishnamurthi isresponsible for the Bootstrap effort at Brown.The Bootstrap team also includes MatthiasFelleisen at Northeastern, Kathi Fisler at WPI,and Emmanuel Schanzer at Harvard. Fundingfor the program has been provided by Google.

College Hill Programming CompetitionThe College Hill Programming Competition was held onMarch 6, 2010. The competition for high school studentshosted by the Brown CS DUG and featured a talk given byAndy van Dam. Prizes were generously donated by IndustrialPartners Adobe, Facebook, Google, Microsoft and NetApp.

Note from the CS DUG:Hey graduates – doing something interesting with CS? Whether you’re at astartup, doing research or anything else, the CS DUG would love to hearfrom you. If you’d like to give undergraduates an idea of what they can dowith a Brown CS degree, let us know by emailing .

10E18342 ibc Page 1 29-APR-10


Department of Computer ScienceBrown UniversityBox 1910Providence, RI 02912USA

Ping!

Where are you and what are you doing?Let us know what’s happening in your life! New job? Received an award?Recently engaged or married? Use this form to submit your news or e-mail .

My news:

Mail to: Conduit, Department of Computer Science, Brown University, Box 1910, Providence, RI 02912or

First Name

Last Name

Class Year

Address

City

State

Zip

E-Mail

10E18342 bc Page 1 29-APR-10


Documents

A Visit to Al Quds Universitycs.brown.edu/about/conduit/conduit_v19n1.pdfA Visit to Al Quds University ALSO INSIDE: The Kanellakis Legacy Lives On Storytelling About Lighthouses 10E18342