Sridhar Vembu’s Blog

Philip Greenspun wanted a cell phone PC a few months ago, and I suggested a variation of it at Mobile Phone as PC. Now Microsoft is suggesting it, with the variation that it will connect to a TV. To repeat what I said

I hope I can buy this device some time soon. I am not a fan of all these less-than-PC mobile devices. I will wait for a cell phone to USB/Bluetooth to LCD display combination.

Every quack is not a scientist, but a lot of established science was once considered quackery. I am reminded of this when we deal with emerging therapies for treating my son’s autism.

Take the hypothesis that mercury (in the form of preservative Thimerosal in vaccines), combined with a certain genetic pre-disposition that makes some kids detoxify poorly, had a major role in autism. In this hypothesis, such impaired detoxification, combined with other opportunistic environmental triggers, caused it. The government establishment has dismissed it, and the mainstream press has accepted that and moved on, all based primarily on statistical evidence. What even the government establishment would not dispute is that mercury could have caused it - it has a ring of plausibility to it. That is an important point: the hypothesis that mercury caused it is not in the same league as the hypothesis that an unfavorable constellation of planets caused it. Yet, when you read the most of the media descriptions, it is dismissed with just such derisive certainty.

To add insult to that injury, doctors who proceed under that hypothesis are considered absolute quacks, and parents who try those treatments are deluded or duped by these doctors. Let me say this: myself and my wife are far from deluded and we are not easily duped. We considered all the literature calmly, we read both sides of this debate. In the end, we came to the conclusion that there is at least a case to be made against mercury. And facing the difficult choice of how to treat our son, we decided the risk of proceeding with the treatment is far less than the risk of waiting till the science is fully established. Time is an important element here - we don’t have the luxury of waiting indefinitely.

In every sense, we are no different than volunteers in a clinical trial for expeirmental medication. We know of the pros and cons, and we have consciously considered everything, but have decided to proceed. Except, of course, the government does not put its official seal of approval on this experiment, but since when is the federal government the authority on all matters scientific? If we cede that role to the federal government, in what way is that different from the middle ages when the Church had the same role?

What does liberty amount to if not the liberty to conciously proceed with such experiments, accepting the consequences of our decisions? Of course, another set of parents, exactly in the same situation, could come to the conclusion they are not persuaded by the evidence. That is as it should be. Yet, if at least a few parents like us don’t try, progress comes to a stop. After all, enrolling in a clinical trial is an act of faith too.

This brings up the broader issue: all science proceeds by hunches and trial and error. It is not a sure thing until firm evidence is at hand. Often, it is not a sure thing even after people declare it to be a sure thing (as the government has declared about mercury’s non-connection to autism) - how often do we hear controversies about established scientific truth? Subsequent discoveries often cast a new light on old accepted truths.

Yet, non-scientists, especially journalists, pretend to more certainty in these matters than there actually is. They don’t understand this process of trial and error at all.

Do I have absolute confidence to say that “Yes, mercury is it?” No, I only have enough confidence to try it, but not enough to go and persuade or evangelize anyone else, other than alerting them to read all the literature and come to their own conclusions. But I will certainly have far more confidence, if the treatments we are proceeding with yield positive results for our son. And I applaud parents who have gone down this path and reported on the outcome, almost all of them positive for their kids. They are brave pioneers.

I can understand an attitude of scepticism - a healthy scepticism is very useful in science as in life itself, but what I don’t understand is the outright contempt.

Mike Arrington of TechCrunch had a pretty critical take on Ning at Ning - R.I.P?. I think Mike has set the expectations way too high for Ning. After all, the service is barely out 3 months, and this is a tough problem to solve, as we can tell from working on Zoho Creator, which attempts to solve a similar problem.

Ning has gotten many things right. I personally thought their mash-up support is quite good, and is a role model for us. Yes, the service requires PHP to do anything serious, but I am sure over time the areas where coding will be required will be whittled down, particularly as the service gains critical mass. After all, the more code that is written on Ning, the les code a future developer has to write. So Ning is off to a very good start - and that is the main point. This whole thing is so young that it is not fair to expect anyone to get it all figured out.

And a broader point: Ning also got hit for the star power of Marc Andressen. After all, even with Marc’s magical presence, code just doesn’t drop straight from the sky for Ning developers - they still have to do all the difficult work of getting this right. I sympathize with them for having to do that under such a glare of publicity. So Mike, cut them some slack.

Keep the fight alive, Ning guys, we hope to learn from you!

I am sure most of you programmers may have experienced this: look at some code, get totally lost, curse that evil son-of-a-bitch who wrote that code, and then realize that it was something you wrote months or years ago. At that point, the narcissist among you may start to admire its beauty and elegance and brevity and style, but the more normal types may start to wonder “Should I move to another line of work?” To help you quell that inner voice, I offer my highly original 27th Rule of Software Development: All code is spaghetti, it just doesn’t taste very good when cooked by another programmer; if you cooked it, it doesn’t stay fresh for very long.

Here is a good application of the rule: when you look at sample code that alpha-geeks present with flourish, especially on internet forums where they fondly hope, against all statistical evidence, that some pretty women may be, just may be, lurking. Often the proud author would present his sample as the ultimate in brevity and style, except that you can’t make head or tail of it at first. Some of us can’t make head or tail of even at last, but that is another story.

My 27th Rule is designed to help you when faced with that situation. Armed with this rule, you can conclude with confidence that the code is the very reincarnation of the Flying Sphagetti Monster, whose ways are mysterious and strange, undecipherable to mere mortals.

As I was reading Sriram Krishnan’s Lisp is Sin it struck me why I simultaneously find myself impressed with Lisp but could never even think about using it. Sriram talks about a love-hate relationship, here is my take on it.

Well written Lisp code is poetry - it is expressive in the same way that Shankepeare is. Yet, while we may admire Shakespeare, most of us don’t routinely blog in verse, and I would never attempt it.

My experience with Tamil and English is illustrative. I did my schooling in India with Tamil as the medium of instruction until my 10th grade. English was taught as the second language. Initially, like all kids in that system, I struggled with English, but somewhere around 8th grade it “clicked”. The trick was finding the famous grammar text book Wren & Martin and learning the rules of grammar [rather odd way to learn a language, when you think about it]. After learning grammar, I took a dictionary and just randomly memorized as many English words as I could find. So grammar + dictionary database - that was how I learnt English. I still could not speak English well until I reached IIT, but my reading comprehension was quite good at that point.

Here is the interesting part: by around 10th grade, I was finding English a lot easier than Tamil. The problem for me was that as a old classical language, Tamil has a lot of polished, ornate, expressive literature, particularly poetry which is routinely inflicted on students, while English-as-second-language was mostly about easy prose. My wife, who went through an English-medium education in India finds it amusing that I “escaped” schooling without learning any English poetry at all. My written Tamil is about as prosaic as my written English, and in Tamil, that is less OK than it is in English. The literate Tamil style is more ornate and expressive. With a mathematically oriented mind, I always viewed language as a tool, a means of communication, not as art. It is both, of course, but the art part language is not for me. I get far more excited by ideas, not the form in which they are expressed.

It is not to say that Tamil poetry or literature is not beautiful. In fact, there are a lot of things in Tamil that are very hard to translate to English while preserving the original effect - something people would assert about many other “classical” languages, like Sanskrit or Latin or even French.

I view a computer language the same way as I do a human language - a tool, a means to an end, not the end itself. I love software as an end-product, but don’t particularly like to view the code, unless I have to. And if I have to, the more obvious the code the better. I am much more impressed by the genius of the finished product rather than the expressive beauty of its implementation. Contrast this view with Paul Graham’s Hackers & Painters. Different strokes …

English became standard because it is a thorougly bastardized language [that, and the little thing called the British Empire] that appeals to the pragmatic business crowd. It is telling that there is no one even to take offense when it is called a bastard language - a French person or a Tamil person would take real offense, for example. It is hard to imagine a Brit or American to be up in arms to defend the “purity” of English.

I think it is the same in computer languages. I bet that the vast majority of the world’s programmers, want an “easy” language, not a highly expressive, poetic language. In Paul Graham’s terminology, they want Blub. And the Blub programmers don’t particularly get worked up when Blub is attacked. And contrary to Graham’s implication that this says something about their IQ, I believe the reason they want Blub is that they don’t get all that excited about something as mundane [to them] as a computer programming language. After all, if you are doing wireless communications systems simulation [as I was] or writing a web word processor [as we are], what matters is the particular domain, and the language in which it is developed is not the most important issue.

So go ahead, call us Blub programmers - we don’t mind.

I read newspaper articles, watch TV programs, and hear commentators like Tom Friedman of NY Times express views about India’s economic transformation, driven primarily by the IT and outsourcing booms. Here is a different perspective, from someone who saw it happen first-hand.

Many commentators in the West, including Tom Friedman, make an implicit assumption that somehow there was or is some kind of an institutional grand plan in all this. Somehow “India” [as a collective entity] focused on education, particularly in science and engineering, and there was a planned take-off. The IITs are mentioned often [I went to IIT Madras myself] as part of that master plan. I would say there is about as much foresight and plan on the part of institutional India in the IT boom as there was foresight and plan in the emergence of the internet industry in silicon valley - both are happy accidents of a lot of not-very-related phenomena.

The origins of IT boom in India could be traced to an accidental decision of mega-star actor/politician turned Chief Minister M.G.Ramachandran, affectionately known as MGR - he was the Arnold Schwarzenegger of that time and place - in my native state of Tamil Nadu. In the late 70’s and 80’s the state was convulsed with caste politics [”identity politics” to Americans], and reservation in educational institutions and jobs [”affirmative action” to Americans] was the hot item on the agenda. By early 80’s, the socialist system that Indira Gandhi foisted on India was visibly starting to fail, and the whole political debate was how to divvy up a stagnant or shrinking economic pie - reservation was just one mechanism the political class came up with, and it mixed well with “vote bank” politics. Since government controlled all higher education, and the number of seats in professional colleges like engineering and medicine - which offered better prospects of a job in a high-unemployment economy - was woefully inadequate compared to the number of kids graduating from high school, reserving seats was the rationing mechanism the political class came up with.

MGR, due to his very successful movie career in a variety of super-hero roles [like most Tamil boys of my generation, I was a huge fan, and remember campaigning for his party in my village when I was a 9 year old!] had built-up a wide support base that cut across caste, religious and class lines in the state. He instinctively grasped that a reservation system that apportioned seats in educational institutions and public sector jobs was going to divide society permanently. After he came to power, he started musing about moving away from caste-based reservation, and instead reserve seats based on economic criteria. This was the “third rail” of Indian politics [and still remains to this day]. He got himself into political hot water quickly. In a dramatic U-turn that would make Arnold proud today, he ordered a massive increase in the reservation quota - it used to be about 30% when he came to power, and he increased it to 67%, and for good measure, he made the vast majority of the population eligible for those reserved seats by expanding the definiton of socially backward communities.

But to assuage his guilt, he decided to allow various private groups, primarily belonging to caste or minority religious organizations, to set up engineering colleges. That was about 1983-84, the time I was in high school, so I was paying close attention to this particular issue. Until that point, the state of Tamil Nadu, with about 60 million people, graduated only about 2,000 engineers a year [and it was considered one of the more educationally advanced states]. His government liberally handed out licenses to start colleges, and more important, they could charge whatever fees they wanted. It became very, very profitable to start an engineering college - due to the massive pent-up demand for education. Within a few years, the number of engineering students had crossed 10,000 a year, and today it stands close to 70,000 a year. The state now has over 250 private engineering colleges, compared to 7 government-operated ones in 1983. All this is in just one state in India, with a population of about 65 million people. Other states, particularly in the South, including Karnataka [Bangalore is the capital city], and Andhra Pradesh [Hyderabad is the capital city] got in on the act. Together, these 3 states alone graduate over 200,000 engineers a year now, by my estimation, and together, they account for only about 20% of Indian population. Northern States were slow to get their act together, and what that has meant is that the IT boom is concentrated primarily in the South.

The quality of the education, as can be imagined, was and still is all over the place. But that is not important as I have argued elsewhere, see my post Placebo Effect in College Education. What these colleges produced is a generation of people, who at least have an exposure to technical ideas and a desperate need for a job, particularly because their parents had spent a fortune on their education. Most of us in India got into engineering not because we had grand visions of invention, but because it had better prospects of a job than the alternatives.

This educational liberalization in my state happened coincidentally at around the same time that Rajiv Gandhi became Prime Minister of India, succeeding his authoritarian mother. Luckily for India, he was a very different person from his mother. He had liberal [the true classical sense of “liberal”, not the totally distorted American sense, where it is almost codeword for “socialist”] inclinations. He was fascinated with technology, and thought it held potential for India. So he liberalized computer imports [which until then attracted duties in the range of 200%, to conserve precious “foreign exchange”]. As an unintended and certainly unplanned consequence, companies like TCS and Infosys imported mainframes and started to provide services to western companies, in almost exactly the fashion of EDS a generation before in America.

It all started small. In 1989, when I graduated from IIT Madras, I do remember hearing about this small software company called Infosys that came for campus interviews. But the IITs didn’t graduate anywhere near the engineers [IIT Madras graduated 250 a year, including all branches of engineering] to satisfy the demand from these companies. So companies like Infosys went to the emerging private engineering colleges. I remember that in late 80’s, the dominant public sector companies of the era [now mostly dead or dying!] would not hire private engineering college graduates. So privatization in higher education in one state of India accidentally provided the fuel for the emerging private sector companies.

There was absolutely no coordination between the policies at the state level where MGR was backed into a political corner, so private sector education was an innovation he came up with on the spur of the moment, and those at the central level, where Rajiv Gandhi, fascinated by technology, liberalized computer imports. For those of you mercantalists, that trade liberalization, which first led to imports, eventually led to the emergence of the export industry.

So really government policy played only a “getting the hell out of the way” role. There really was no pro-active government role in the emergence of IT. Nor do I, as a libertarian, think government should play any such role.

So when well-meaning commentators like Tom Friedman use the competitive challenge from India [which is utterly meaningless because “India” doesn’t compete in any sense with “America” - these are way-too-broad entities to have any coherent meaning; only individual companies compete] to argue for more government “investment” in education, I blanch. If anything, my preference is for the government to get the hell out, and remove barriers to trade and investment.

And the one area where the American government is actively harming real industry in America [as opposed to the financial “industry”] is through faulty monetary policy, as the article by Eric Janszen that I posted about earlier explains. But Ben Bernanke looks all set to continue the harm, with his foolish [as in “only an ivory tower academic could come up with this”] monetary ideas. The credit bubble that the Greenspan/Bernanke fed has unleashed on the world is now a global monster, impacting Beijing and Bangalore as badly as Boston.

That pleasant attack-the-Fed diversion aside, the IT revolution in India really had its origins in the accidental privatization of higher education that was unleashed in one state in India. Even today, as a direct consequence, a disproportionate number of the workforce in the Indian IT industry as well as the Indian technology diaspora in the West, paricularly America, come from Tamil Nadu, though other states are catching up quite fast now. And we have my childhood hero, MGR, to thank!

Here is a beautiful one: Autism and my son by a mother.

Jambav is a kids site, with a special focus on special needs kids. It is backed by AdventNet.

Today my son achieved a breakthrough. Usually I get him ready for school in the morning. After I seat him in the school bus, I come out and wave. He would just stare at me through the window. Today, he smiled and waved back - something that he hasn’t done before.

Recovering from autism is a series of these successes. What appears so natural that we overlook it in a normal child is herculean effort for these kids. I am happy he made this breakthrough today.

Dealing with autism has brought a different perspective in life: almost everything looks like a small challenge compared to this. Having him speak fluently is the equivalent of winning the Nobel Prize for me and my wife.

I read this gem in this interview by Frederick Hayek.

In Economics, this mathematical modeling disease has gone so far that it is practically impossible to read any paper from an academic economist in a major university without finding a lot of equations. It takes a lot of intellectual courage to resist it - because there is the attitude “If you can’t do math, you are not smart enough”.

In engineering, the same disease shows up as the quest for theorem-proof style discourse. Engineering systems are far, far simpler than what you come across in the social or biological sciences, in the sense that realistic mathematical models are possible, but those realistic models often don’t yield satisfying theorems, but can only be studied through computer simulation. This doesn’t satisfy the more ambititious young men - “we must try, we will ultimately discover it” is the cry. They are free to try anything, of course, but I do mind them trying it at taxpayer expense. If they want to indulge in pleasurable mathematical speculation, let them do it on their own dime. I think once the gravy train of government funding were withdrawn, a lot fewer of those kinds of papers would be published.

The danger with such taxpayer-funded “academic” speculation is that these academics form a symbiotic relationship with the government that funds them. Ben Bernanke is classic proof - the man has never spent any time in the real world of business, has never worried about making payroll or covering costs, yet he is going to gain control of the most important institution of the free world, namely its money. For him, money is merely a variable you manipulate in some models.

As Keynes observed:

Bernanke is not that one man in a million, that is for sure.

Keynes, of course, was not very consistent in his views, as Hayek observes in the interview:

More drawings by my six year old son who [we hope] is recovering from autism. The first one “Man” [titles are given by me, because he still doesn’t have enough language to give it a title], was drawn when I requested him to draw my father, and the drawing does give the feel of an older person. The second one, “The Couple” was his own spontaneous effort; that looks better, don’t you think? Unsurprisingly, he is at his best when he grabs the magna-doodle and draws something to occupy himself; he would usually erase it just as quick as he draws it, so we have to grab it out of his hand and photograph it.

His previous drawing “Indian Lady” is at my post here