Friday 22 January 2010

The next killer app?

Written in Thrissur under the influence of severe jet lag, Kerala masala chicken, a round of Mindflex in Sunnyvale and an overdose of TED videos.

Name: MISCRE

Executive summary: As you sit and daydream, or just think, this wi-fi nanoimplant will read the signals in your image forming cortex, access all your image databases (cellphone, Laptop, Picasa, etc), retrieve the most relevant images and play a slide/video-show for you, SCREened in your MInd's eye (aka visual cortex). Animation, seguing, cartoonization and other image-processing effects will be available as add-on apps priced separately. The killer amongst the killers will be an app that will allow you, via a Pranav Mistry (SixthSense)-type headband, to "project your innermost images" on a wall near you. The catch is that, like Mindflex, you will have to learn how to turn this app on and exercise parental "control" on what it shows!

Technology: It's all here already, but if you want a new flashy name, you could call it neuroimag(in)ing.

Caveat: With all this global body search imaging technology around, you will have to watch out for MISCREants.

Dateline: Thrissur, 5:30 a.m., 4th January, 2010.

Thursday 21 January 2010

Je vous salue, Maurice

La chose essentielle, c’était que Serre à chaque fois sentait fortement la riche substance derrière un énoncé qui, de but em blanc, ne m’aurait sans doute fait ni chaud ni froid – et qu’il arrivait à “faire passer” cette perception d’une substance riche, tangible, mystérieuse – cette perception qui est en même temps désir de connaître cette substance, d’y pénétrer.—Recoltes et Semailles, page 556.

These words were written by Alexandre Grothendieck, the brilliant French mathematican, about his colleague Jean-Pierre Serre, whom he called a “detonator”, one who provided a spark that set the fuse burning for an explosion of ideas.

I find these words and the term detonator particularly apposite of our beloved colleague Maurice Bazin, taken so suddenly from us, because they describe his insatiable curiosity and desire to get to the bottom of things, to understand and “feel” things, whether about science, society or his fellow human beings. And it was never idle curiosity, but rather one with the intention of demystifying, reducing to the lowest common and comprehensible denominator, always maintaining rigor and rationality and involving everyone around him to discover together.

Maurice was always closing the gap between ideology and praxis, in his personal life as well as in the lives of people he came in contact with, determinedly following a path of mutual dialogue and inspiration. He taught us to be obstinate in order to achieve intellectual goals, but also to be flexible when practicality demanded it. He was never one to avoid dirtying his hands or shirk hard work, whether physical or mental. Whitewashing walls in the Espaço Ciência Viva hangar, doing carpentry to make exhibits, discussing the finer points of ethnomathematics and then going and living it in a tribe near São Gabriel da Cachoeira: all this he took in his stride, never hesitating to plunge in headfirst. Generosity with his time and affection also marked Maurice out as a very special individual.

We are bereft but privileged to have had the chance to be illuminated by someone so warm and human and yet so rational and rigorous, a mix that we will always miss and spend our lives trying to emulate.

I close with another quote from Grothendieck, that Maurice would have appreciated: Et toute science, quand nous l’entendons non comme un instrument de pouvoir et de domination, mais comme aventure de connaisance de notre espèce à travers les âges, n’est autre chose que cette harmonie, plus ou moins vaste et plus ou moins riche d’une époque à l’autre, qui se déploie au cours des générations et des siècles, par le délicat contrepoint de tous les thèmes apparus tour à tour, comme appelés du néant – Récoltes et Semailles, page P20.

Viva Ciência, Viva o Espaço de onde Maurice agora nos espia, com aquele olhar maroto.

20th October, 2009

Reflexões sobre ensino, pesquisa, revisão crítica e avaliação

As reflexões do título foram feitas no contexto da Sociedade Brasileira de Automática, da área de controle e automação e da revista Controle e Automação da SBA, e estão baseadas nas seguintes hipóteses/querelas:

1. A maioria de nós, pesquisadores das áreas de controle e automação, não envia os melhores trabalhos para esta revista.

2. Mesmo que enviasse, não temos massa crítica para conseguir revisões de alta qualidade para tais artigos no Brasil, levando em consideração que a participação de revisores estrangeiros está limitada por vários motivos [nem todos os artigos são escritos em inglês, a revista ainda não tem fama suficiente para atrair mão de obra voluntária, ainda mais estrangeira, etc.]

3. Ao longo dos anos, a revista vai perdendo o foco: hoje publica muitos artigos de outras áreas (potência, eletrônica de potência etc.) que têm pouco ou nada a ver com a missão original evocado pelo título da revista. Mesmo com essa abrangência mais ampla, artigos publicados nela recebem pouquíssimas citações (44 citações ao todo, para todos os anos e todos os artigos, desde que pertence à base SCIELO, utilizando a ferramenta de SCIELO. 36 dessas citações são oriundos de artigos da própria revista).

4. Os melhores artigos (pelo menos os premiados) geralmente são surveys e/ou didáticos.

5. As universidades federais, e a própria CAPES (cuja missão é avaliar Educação e Ensino) valorizam o quesito pesquisa muito mais do que o quesito ensino. Como prova disso, podem-se citar muitos fatos. Vou me contentar com apenas dois. A maioria das universidades federais não possui avaliação séria do ensino, não premia bom ensino (embora haja premiação das melhores teses). A progressão na carreira universitária é baseada maciçamente na produtividade em pesquisa; quanto ao ensino, há uma simulação de que está sendo levado em consideração, mas sabe-se que na realidade, não há mecanismos para fazê-lo.

Dadas essas hipóteses, e o gosto que temos pelo ensino (pessoalmente, pelo menos!), a pergunta que se coloca é a seguinte: como podemos agir para melhorar o quadro desenhado acima?

Antes de entrar no mérito desta questão, queria colocar mais algumas observações sobre revistas técnicas e científicas. Em minha opinião, os formatos atuais das revistas bem como o processo atual de revisão estragam bons artigos. Especificamente, mesmo com versões eletrônicas de quase todas as revistas importantes, parece que limitações de papel ainda são impostas aos autores. As seqüelas disso são retiradas de material didático [considerado elementar], supressão de abordagens e exemplos que não dão certo, etc. [muitas vezes quase tão importantes quanto os que dão certo]: processos que tornam o artigo menos compreensível e certamente menos didático. Quanto ao processo de revisão, deparamos com a perversidade de revisores anônimos. Justifico minha escolha de palavras. Hoje, com o acesso fácil a quase toda a produção científica do mundo, uma revista se tornou nada mais do que uma lista de leitura indicada. Indicada por quem? Os revisores, é claro. Leríamos um romance cuja contracapa traz recomendações anônimas, por melhores que fossem? Creio que não. Creio, portanto, que os revisores devem se identificar. Isto traria outros benefícios. Por exemplo, levaria a uma redução drástica de comentários levianos, mal-justificados ou simplesmente mal-educados, que, certamente, todos nos recebemos algumas vezes nas nossas carreiras.

Aonde quero chegar? Quero sugerir a criação de uma “magazine” da revista Controle e Automação, que teria existência apenas eletrônica, com forte foco didático. Seria no formato Wiki, contendo o artigo, os comentários e contribuições dos revisores, assim como as contribuições e comentários dos leitores (que incluiriam texto, imagens, áudio, vídeo, código [em MATLAB, Scilab, Java, you name it]). Poderíamos imaginar um contador de “hits” relevantes (isso teria que ter a intervenção de um ou mais editores), e no final do ano, os 10 artigos mais acessados e que receberam mais contribuições relevantes da comunidade receberiam o grande prêmio CAPES de ENSINO, com conseqüências positivas para a avaliação do departamento/programa que abriga o(s) autor(es). O acesso seria gratuito, mas potenciais autores e leitores teriam que se cadastrar (login e senha). Seriam aceitos artigos didáticos, notas que sugerem como abordar um assunto em sala de aula (um vídeo de uma aula sobre um determinado assunto, por exemplo, no formato YouTube). Enfim, our imagination is the limit! Eu imaginaria o dia em que esse magazine engoliria a eminência parda (revista atual)...

Em tom mais sério, vejo algumas vantagens. Enquanto sabemos que há falta de revisores qualificados para assuntos mais técnicos de pesquisa, temos plena consciência de que há muitos docentes que possuem excelente didática, geram ótimo material didático que se perde quando se aposentam, e que, em geral, estes docentes se sentem pouco prestigiados pelo sistema vigente. Ou seja, temos massa crítica de sobra, e, mais ainda, uma massa cinzenta nesta área que quer mostrar seu valor e, atualmente, não possui instrumentos para isso.

Para finalizar, se esta proposta funcionasse, poderíamos imaginar um futuro no qual a CAPES avaliaria apenas ensino, usando mecanismos como esse, ao passo que o CNPq faria o que ele já faz, avalia pesquisa. Como um programa é um conjunto de docentes que são avaliados em relação aos quesitos de pesquisa e ensino, só restaria discutir como agregar os dados individuais de ensino (CAPES) e de pesquisa (CNPq) dos membros de um programa para gerar uma avaliação completa do conjunto. Radical? Muito distante? Pode ser, mas creio que vale a pena pensar em começar, ainda que fosse com um passo inicial pequeno.

Reflections on reaching 50

First I’d like to register the fact that, even though I seem to have fallen for the half-century mystique by choosing this day to write on, I don’t really think that 50 years of life on this earth is worth such a big deal. Some reasons that I could give for this follow.

Given that my life expectancy, based on both maternal and paternal relatives and immediate family, is around 90 years, the half way mark is around 45 and not 50. Then again, I don’t think that there is much to celebrate at a half way mark, whether it’s at 45 or 50, since you have just completed the better half of your life. Your powers, both physical and intellectual, have now been on the wane for quite some time, and, as the mathematician Hardy pointed out, you are now at the age when you can start writing books, since innovative research is no longer a real option.

Although I still have most of my teeth, and all my hair, which is salt and pepper, with the salt winning handsomely, signs of wear and tear are everywhere. My children’s friends all call me “coroa” and “tio” (geezer and uncle, respectively, if not respectfully) and, recently, reflecting my new status, even “vovô” (grandfather). I am ten kilos heavier than I should be and my knees hurt after a week of running that is intended to shake off the circumferential excess adipose tissue.

So what should I celebrate? Perhaps the fact that I’m still alive and that, despite all manner of maintenance, plumbing, hardware and software problems that will undoubtedly arise, I will eventually have plenty of leisure time to do all sorts of interesting things.

Optimistic interlude: Perhaps a good retirement occupation for me, even a dream occupation, would be to become a culture jockey for some kind of multimedia “site” which has a music section where I can podcast my current “world” music favorites, a documentary & film site which is also mainly on an exotic axis, a mathematical “gems” corner to show Proofs from The Book, and various other tabs to showcase poetry, linguistics, literature. Finally, an e-library, which embodies the motto “Borrow what you want, forever”, but also has a drop-box where borrowers can contribute their favorites, in any form and where I can go and curate the “droppings”. All this to be seamlessly and shamelessly (i.e., no copyright hassles) integrated into something that ensnares an unsuspecting surfer who rides into this “site” to at least tarry awhile and contribute some gems. End of interlude.

Maybe I should also celebrate the fact that, so far, I have been luckier than many of my counterparts and contemporaries, although, given the large variability in their luck quotas, this is not saying very much.

In any case, I hope to have convinced you that, except for the privileged few who rise above the sea of mediocrity that engulfs us all (myself included as engulfee!), there is really not much to celebrate at 50. Perhaps the picture changes at 70 or 80, but I don’t really believe this.

Dotting your ı’s in İstanbul

Keyboards, which belong to a bygone era of typewriters and other mechanical and electromechanical “business machines” are still well represented in today’s touch screen world, which has got to the “touchy” phase, but still needs to evolve to the “feely” phase.

Keyboards are still the most common man-machine interface and still cause most people grief of various kinds, ranging from inability to use all ten fingers to type fast and without mistakes, all the way to tendinitis. But it would be unfair to suggest that they haven’t evolved at all. They now represent their countries of origin in all their linguistic and typographic complexity.

Here are some examples. The Turkish language has two i’s: one with a dot over it, which has the sound of the long e (as in the English word feet) and one without the dot, which is a schwa, to use the linguistic term for it. The catch is that the undotted i occupies the spot where you would expect to find the ordinary dotted i on a “normal” QWERTY keyboard used in English speaking countries. The dotted i is just to the left of the enter key (see the figure below).

Figure 1: Keyboard layout for the Turkish language (Image source: http://www.turkishlanguage.co.uk/pics/smallqwerty.jpg )

This figure also shows the umlauts on the u and the o, the cedillas or hooks beneath the c (in common with French and Portuguese), and the hacek or inverted hat over the g, which softens its sound to the point that it almost vanishes.

And, staying with Continental languages that use Roman scripts, there is Polish, which uses cedillas on other letters such as e and a, as well as acute accents, dots and bars, both horizontal and inclined, as diacritic marks (see figure 2).

Figure 2: Keyboard layout for the Polish language. (Image source: http://www.terena.org/activities/multiling/ml-mua/test/img/kbd_polish.gif )

So, you might legitimately ask, how do keyboard designers accommodate these extra letters, phonetic wiggles, curlicues and so on. The figures above would lead you to deduce, correctly, that although the basic QWERTY layout of “normal English” keyboards is maintained, all the other special characters are moved around as the designer sees fit, motivated in part by the frequency of occurrence of the nonstandard symbols. Frequently used symbols can be typed using just one keystroke, while less frequent symbols are mapped to “chords” or sequences of keys that must be pressed simultaneously. The chord solution usually involves the use of the Alt (from Alternate) key, which when pressed, endows each of the usual QWERTY keys with an alternate symbolic value that appears on the key in a smaller font or sometimes in a different color.

In all this diversity, there is, however, one surprisingly consistent choice of “displaced” symbol: the all important character @, which you need for e-mail addresses that your mail program has not stored. The figures above illustrate this well: in the Turkish keyboard, the @ has migrated from its usual place above the numeral 2, to a spot below Q, while in the Polish keyboard, it has been demoted further, to a spot below V. And that’s not all. It is usually not obvious how (i.e., using what combination of keystrokes) you type @. In fact, in my travels, I have sometimes had to resort to copying and pasting the @ from a mail header into the address box. There is, of course, always an official way to do things. For example, on the Turkish keyboard you should type the chord Alt Gr + Q to get @. And what pray is Alt Gr? Alas, nothing to do with a disgruntled Turk by the name of Altug: the abbreviation stands for Alternate Graphics, a fact that you can read more about in the excellent Wikipedia article on the Alt Gr key.

You might be wondering, at this point, why anyone would go to the trouble of using local (nonstandard) keyboards when you could just use your own laptop, from the comfort of your hotel room. At different times, I have had different reasons. Traveling light and not wanting to carry a laptop, not wanting to deal with the complications of getting or paying for a wireless connection (for example, you could get a free wireless connection in most airports today, but you usually need to have a cell phone and a large amount of patience, neither of which I have), and finally, on the positive side, wanting to see the grungy but democratic cyber hotspots (not cafes!), where ordinary people who are on the poorer side of the digital divide go to apply for jobs, type CVs, chat with boyfriends or girlfriends in unattainable first world or Eastern European locations, play war games online, do research on the history of Ukrainian tractors and such.

All of this doesn’t even begin to get into the complexities of using Roman QWERTY keyboards to type those Asian languages that are phonetic (hence approximately transliteratable into Roman script) but use a different character set. And then, of course, there are the languages based on ideograms, such as Chinese and Japanese, which would require a separate essay and one that I am not qualified to write (but watch this), except to mention that an even more painful popup menu-driven technique, called Input Method Editor is used. But I will say a little about typing Bengali on QWERTY keyboards, using True Type Fonts for Word, designed by intrepid Bangladeshis and Indians, in order to be able to write in the beautiful Devanagiri script.

The difficulties in typing are many and I will describe the main ones. The character set is much larger (32 consonants and 11 vowels). In addition, the Bengali alphabet is a syllabic alphabet in which vowels can be written as independent letters, or by using a variety of diacritical marks which are written above or below, and before or after the consonant they belong to. Finally, when consonants occur together in clusters, special conjunct letters are used, and the symbols or letters for the consonants other than the final one are truncated or modified in some way. (see examples here.) Let me explain the particular subtlety of the diacritical mark corresponding to a vowel being written before the consonant it belongs to. In most languages written using the Roman alphabet, letters are written in exactly the order that they are pronounced. Thus LI represents a sound in which the letter L is pronounced before the letter I. However, in Bengali, the letter I could represent a short vowel and, in this case, would be written before the consonant L in the Devanagiri script. In short, this means that if you were using a QWERTY keyboard mapped to the Bengali script, you would type IL instead of LI.

After this elaborate description, you might imagine that the complexities are such that no one would really be able to use such a keyboard. Wrong! My father was already into his eighties when he mastered one such keyboard system for Bengali and this year, just a few months short of his ninetieth birthday, has just completed a translation of Rabindranath Tagore’s two volume, 600 page plus, opus called Shobdo Tattvo, which explains the intricacies of the Bengali language, from linguistic, phonetic and cultural points of view. The manuscript, by its nature, even in translation, calls for the insertion of phrases and words in the Bangla script. My father did all this using software called iLEAP, created by CDAC (India’s Centre for Development of Advanced Computing), and billed as an intelligent, Internet ready, Indian language word processor. Of course, as one might expect, even though the look and feel of iLEAP is like that of Microsoft’s Word, the two are not compatible. So the latest mission, in order to publish this translation on the web under a Creative Commons License [1], is to produce a PDF version of the two volume opus. An intermediate step in doing this is to translate the parts of the manuscript which are in Bengali script to a font known as Bangla 21st February font. The date, proclaimed in 1999 as the International Mother Language Day by UNESCO, originated as the international recognition of Language Movement Day, which has been commemorated in Bangladesh (formerly East Pakistan) since 1952, when a number of Dhaka university students were killed by the Pakistani police and army in Dhaka during the Bengali Language Movement. The date became a symbol of the Bangladeshi struggle for linguistic and political independence. The significance of this is that it is perhaps the only example of a language so venerated by its speakers that they actually fought and won a war of independence against an occupying power, in order to have the right to continue speaking it. The slogan of the times was Joy Bangla, meaning Victory to Bangla, used for both the country and the language, and this is as good a slogan and story to end on as any.


[1] The reason for this is that, sadly, no Indian publisher or university, including Visva Bharati University, is willing to publish this translation.