Our education problems.

So, we finally have a HRD minister who is more interested in education than (de-)saffronisation, a government that understands the seriousness of our education problem, and a bunch of reports that tell the government what to do. Why then, am I putting my skeptic hat on as I write this post?

Because there is nothing to suggest that the government, the minister, or even the reports have moved beyond the command/control mode of operation. Take the Yashpal committee report for instance. It recognizes that the current system of regulators, consisting of the UGC, AICTE, and the MCI amongst others hasn't worked. There is endemic corruption, the latest being this case against the AICTE chiarman and a senior member, inefficiency, and a lackadaisical attitude. But look at what the committee has recommended: a uber-regulator that encompasses all these! How, in the lord's good world is this going to solve the problem? Not only that, the report suggests that this uber-regulator will also solve the problem of multi-disciplinary education! Sixty years after independence, and eighteen years after liberalization, we still haven't understood the fundamental difference between regulation, and control. Our administrators, and report-makers don't get the fact that the best regulation is the market and full disclosure. For instance, instead of having corrupt bureaucrats decide which institute should function and which one shouldn't, have full disclosure of every institute - the intake, the aggregate scores that the students got, the placements that they got, and the types of companies they got placed into, the numbers who went to higher studies, the number of working computers, the facilities in the lab, whether the hostels have enough clean toilets - disclose disclose and disclose. Then let parents and students decide which institutes should survive and which ones shouldn't. Let foreign universities in, set stiff criteria that these institutes should satisfy, but let them teach what they want to. Let every institute pay what it wants to, charge what it wants to, maybe subject to a range that the government can specify. Let there be merit-based salaries for teachers, and scholarships for those students who cannot afford high fees. Let go, but keep a watchful eye.

I can give three instances of why my prescription will work. Today, PESIT is probably the best engineering college in Bangalore. Management and NRI seats in this institute are auctioned off, with waiting lists spanning a few years. Few people know that just ten years ago, few people joined this institute. In fact, at that time, the institute offered to pay the fees of any student below rank 1000 who joined it! The dedication of the founders has led to this institute becoming the top institute in Bangalore. They poached professors from other institutes, got people from abroad to join it, and of course, made a lot of money in the process. But what the city got was a good institute.

My next example is DAIICT, an institute from where my team has recruited many interns. A large number of them have been fantastic and have gone on (or will be going on) to graduate programs in UCB, UWash, and other universities. Our experience has been that students from this institute were really well-rounded, however, I was told recently that the AICTE had refused to recognize the Info. and Comm program that these students had graduated from!

Then, ISB. The International School of Business in Hyderabad has the most expensive MBA programs in India. It wasn't recognized by the AICTE, whose diktats it couldn't live by. However, as this news article points out, it was the only Indian B-school to be in the FT top 20 B-schools in the world.

To conclude, I want to bring in Game theory, of course in a very unscientific way. When you have two competing players each one making a decision knowing fully well the other's strengths and weakness, you usually reach an equillibrium that may not be the best for either player individually, but is a good bet overall. This is what the parent/student v/s institute game would do. On the other hand, cartels break this equillibrium in favour of one party - which is what the nexus between institutes and our current regulators is.

I told you so! :)

Some time ago, I’d predicted that Google’s next step towards world domination would be to produce a free OS for PCs. And as I predicted, they first made the mobile OS and now, they have the same running on PCs produced by Acer.

PS: This post, like the rest of the blog, only reflects my opinions and has NOTHING to do with my past, present or future employer(s).

Nixon-Kissinger II

If you are a student of contemporary Indian history, you'd be familiar with the Nixon-Kissinger duo, and their attitudes towards India. While a lot of water has flown down the Mississippi and the Ganges since then, it looks like the wheel of time has come full circle again, giving us Nixon-Kissinger Part 2, also known as Obama-Clinton.

First, some history. Nixon and Kissinger, with no little help from Indira Gandhi, took Indo-US relations to their nadir. In their quest to get China on their side, against the Soviets, the duo looked the other side when Pakistan committed some of the worst abuses of human rights in the sub-continent, in what was then East-Pakistan. They formalized the US policy of preferring dictatorships over democracies, even when those dictatorships brutally supressed the mandate of democratic elections. In fact, Nixon offered to send the "Seventh fleet" carrier group into the Indian Ocean to pressurize India which intervened on the side of democracy. So bad was the deterioration in relations that the leaders of the countries couldn't even talk to each other without one calling the other a b***h.

The intervening years, the Kargil war, India's economic recovery and the Dubya-presidency had all contributed to healing the rift, but it looks like this is one wound the great black healer is going to rip open in his quest to heal other wounds.

Now, it is not wrong that the US has its own national interests in mind. But Obama in his nearly 6-month presidency hasn't made a single serious comment (and I'm ignoring platitudes like the praise he heaped on Moron) towards furthering Indo-US relations. He's made noises about getting India to sign the NPT and CTBT, has completely ignored the nuclear deal, brought back the hyphen between India and Pakistan, put pressure on India to start talking with Pakistan, made Bangalore the enemy in the outsourcing debate and in the latest of his antagonising statements, has opened the Kashmir bogey again. Not since the Nixon-Kissinger era have we seen such 'attacks'.

Will the relationship survive the Obama-Clinton foreign policy administration? I doubt it.

(Postscript: BTW, we had Ramachandra Guha, the author of "India after Gandhi" visit our lab and deliver a talk recently. Next in line are Sudha Murthy and Ramesh Ramanathan!)

Post = Random.NextPost();

Last week, we had a “Bring Your Child To Work” day, when parents are allowed to bring their kids to work so that the kids can get an idea of what their parents do during the day. We had a wide range of kids – from age 2 to age 14, and boy, was it fun!?

I couldn’t help noticing in all the revelry, that a disproportionate number of the children were girls. Even more peculiar, it seemed that it software fathers had a higher chance of having girl children than software mothers. Non-software folks had either male children or had an equal number.

This couldn’t be statistically true, I thought. But after a quick recap of my friends’ families, I think I’m ready to say that it is indeed statistically possible that male employees of software companies tend to have more girls than their female counterparts, and their male counterparts in other occupations. This needs more evidence, of course, and there needs to be a scientific reason for why it is so, but I think there is enough merit in trying to investigate the case.

And it was last week that I took part in my first protest march. We were protesting against the illegal felling of trees at CNR Rao circle, near IISc, which was ostensibly being done to build an underpass there. Now, I’m not a tree-hugger, and probably will never be, but this was something illegal being done. So, a few friends from MSRI and I went together, shouted slogans against the BBMP, had our photos taken, and came back. I have more to write on why such protests aren’t successful, but I’ll save that for another post.

BTW, here are a couple of photos of me at the protest:

http://www.deccanherald.com/content/5075/iisc-staff-students-protest-tree.html (I'm the one leaning on the tree)
http://www.expressbuzz.com/edition/story.aspx?Title=These+cuts+run+deep&artid=9tgoy95Efds=&SectionID=Qz/kHVp9tEs=&MainSectionID=wIcBMLGbUJI=&SectionName=UOaHCPTTmuP3XGzZRCAUTQ==&SEO=

Anyhoo, another thing that happened last week was that a submission a few of us from MSRI made to FSE 2009 was accepted and it’ll be part of the proceedings. The conference itself is in Amsterdam in August. I’ll see if I can post the paper somewhere.

All for a single post.

So, what do I write about? Well, there is the election and the mess the BJP finds itself in. There is the Lankan Tamil issue. There is a bunch of stuff happening at work. There is the realization that I’ve forgotten to read, and there is a lot of economics that I’ve been ‘reading’. There is also a nice report in "ವಿಜಯ ಕರ್ನಾಟಕ" on the measures colleges in Bangalore have taken to ensure a ‘safe’, educating environment for college boys and girls.

Must get out of this mode.

The greatness of Google

For some time now, friends and blog-readers have accused me of having an anti-Google bias, thanks to many posts on my blog that didn’t share their unbridled enthusiasm about the company. However, for a knowledge-o-phile like me, Google has come as such a blessing that I can’t help but write this post.

As a child, I devoured books. I read everything I could lay my eyes on, and that included “Children’s knowledge bank”, Wisdom, Misha, “Baala vijnana” and many others. Still, so many questions remained unanswered. I had to wait for a book or a magazine article or for someone to answer the questions I had. Many an argument remained unresolved because there was no authoritative source for the subject.

Web 2.0 led by Google and Wikipedia and so many content providers has changed all that. Wondering about tectonic shifts? Go to Google. Thinking why the Afghans have had such a history? Read Wikipedia (but mind you, only as a starting point to other articles.) In an argument with a close friend over some obscure scientific fact? Type your question in Google.

Few companies have brought information to our finger tips like Google has. For that, thank you, Google.

Of course, I have a strong opinion on the Google-Wikipedia duopoly over information, but I’ll let someone else take potshots at it, here.

(PS: Before some of you start wondering, let me disclose: NO, I have not been offered employment at Google, nor am I in a race for working at that firm.)

Technology Notes Vol 1, Issue 8: Lessons at work and play

This edition of Technology notes is dedicated to some lessons I learnt at work and at play.

* Conceptual Abstractions

If I were to stick my neck out on things I don’t fully understand, I’d do so and say that Abstraction and Recursion are the most fundamental principles in Computer Science. There is something magical about the way these two concepts solve so many problems that we encounter 'in code’. In fact, in the OO design community, “an additional layer of abstraction” is almost a silver bullet for most design problems.

In the last month though, I got first-hand exposure to some conceptual abstractions. In my definition, a conceptual abstraction is one that solves a complete class of problems while trying to solve a single instance of the class. Let me give an example. Over the last year or so, I have been working on a tool that uses some modified IR techniques to solve a problem faced by developers in our organization. We read some algorithms, coded them up, created a service abstraction for folks to use, and made it available to our user community. We decided to publish this work, and I had detailed discussions with my manager and Sriram Rajamani, who is a principal researcher in our organization. In the space of a few hours, we (truth be told, they) created a conceptual framework around our tool, in such a manner that we solved the general IR problem for our problem space! I wish I could write more about it, but I’ll have to defer it till we submit the paper.

* The beauty of Windows

When I got my first PC, nearly 14 years ago, one of the games I loved playing was Xargon, a 1993 game created for DOS. I couldn’t afford to buy the game, so I played a shareware version that was available on a PCQuest CD. 15 years and five OSes later (Win3.1, Win95, Win98, Windows ME, Windows 2000, Windows XP)  I installed the now-free retail version of the game on my XP box, and lo-and-behold, it worked! I had a few problems with the sound card which I was able to fix by some minor changes to the game’s default settings. I don’t think any company makes software that preserves backward compatibility to such an extent. In a world where different versions of Linux are not compatible, where Mac software bought two years ago for the 68x processors won’t work properly on the new Macs, the extent to which Windows preserves backward compatibility is amazing. You can read the 68 posts Raymond Chen has about it here.

* The best feature of Java

The one feature of Java I miss in C#/.NET is that of checked exceptions. You know, you declare a function as

public static void X( ) throws A, B, C

and the function can only throw A, B or C. Clients of this function must handle these exceptions or declare them in their “throws” clause. Further, if a derived class overrides a base class method, it can only throw a subset of the exceptions that the base class method has declared in its “throws” clause. Not only is this a great example of documentation enforced by the compiler, what this also does is takes Java one step closer to the ideal of LSP (Liskov Substitution Principle).

(PS: C++ fanatics will now argue that it too has a “throw” clause. All I’ll say is “don’t use it”.)