Search This Blog

Thursday, September 21, 2017

Completing 24 years at IITK

Today I complete 24 years as a faculty member of CSE Department at IIT Kanpur. It has been a fascinating journey, and I have learnt a lot. The student interaction has been particularly satisfying. So here are some incidents/events that I recall on this day. I realize that people write these blogs on 20 years or 25 years, and not at 24 years, but I don't want to assume anything about the future, and I feel like writing today.

It is long. So take your call whether you indeed want to read this.

The first thing I remember about the department is how poor it was when I joined. As a graduate student, I had two workstations on my desk, each with 1GB hard disk, both full of network monitoring data. When I was coming to India, I copied that data onto 20 tapes of 60MB each. My advisor told me that it would be useless since I won't have so much storage/compute power to do anything with this data. But I was like, if I can have two workstations as a student here, the premium computer science department in India would at least give a workstation to its faculty. How wrong I was.

When I joined, I was pampered by the department, and I was given the PC with the best specifications. A giant 40 MB disk. Out of this, 20 MB was consumed by Windows, and I was left with 20 MB of diskspace for my files. The Computer Center used to give a disk quota of 10 MB, and the CSE lab gave me yet another 5MB. I pleaded with both CC and CSE to give me 60 MB of space for a few hours. There were two tapes which had the most important data with which I could start working with. I need to copy the entire 60 MB on a disk, and then delete 40MB out of it, and move the other 20 MB to my PC and then I could do some post processing. But for the next six months, I would not get 60MB. I was promised that if CC or CSE buy any new server then before making that server public, I would get access to it for a day. And that took six months.

CSE Department had a total annual hardware budget of Rs. 1 lakh, and an entry level server would cost more than that. So CSE department could never buy a server unless in some year, the administration was kind enough to give us a special grant to do so. Most of the servers and infrastructure in the  lab, including PCs, were bought through projects by faculty members. (Can you imagine today faculty members placing their resources in a common lab in any department. But that was the CSE department culture at that time. And it was this culture that had attracted me to IITK to begin with. The labs are considered personal fiefdoms of the faculty in most departments even at IITK, not to mention the other institutes.)

It was a very tough period. And frankly, CSE departments in other IITs were not this poor. Our budget on a per capita basis was the lowest compared to all other departments in IITK. I had other options, and I thought about them. But the person who mentored me throughout this difficult period was Gautam Barua. But for him, I would have left IITK soon after joining.

The department was a very cohesive unit, and it took care of everyone. For one whole year, before I got married, I only cooked on Saturdays and Sundays. Lunch was in some hostel mess. Dinner was fixed. Every Monday at Sanjeev Kumar's place, every Tuesday at Ajai Jain's place, and other days by rotation among several other faculty members. The social life was absolutely fantastic. So many kids' birthday parties, potlucks, picnics - sometimes with families and sometimes with students, and sometimes both. Today, we are more professional and such get togethers happen within small groups, that too infrequently.

Pankaj Jalote realized that the only way to build the department was to get more money, and the primary source of that money could be industry. So he started something called Industry Affiliates Program (IAP). In those days, any faculty member wanted to initiate anything, they would be encouraged to do so. In IAP, we will invite industry to become our partners by paying a small annual fee. In return, we will inform them of every project, thesis, etc., happening in the department. We will invite them once a year for discussions on various issues, including joint projects and curriculum, etc. After a year, this responsibility was given to me. I was much more direct in asking for funds. I never had shame in asking for money. And we received a large donation from Verifone. It was around Rs. 35 lakhs. Let that sink in. A department with an annual budget of Rs. 1 lakhs suddenly get a check for Rs. 35 lakhs. That was the largest donation that IITK had received till then, and was more than all the donations that IITK had received in the year through Dean of Alumni Affairs office (called DPRG at that time).

And the way it happened was interesting. Verifone decided that they wanted to donate Rs. 1 crore to Indian academia, and to ensure that this had some impact, they also decided that they will not spread this too thin, but divide this money into only three departments. So they wrote to who ever they knew. And thanks to Industry Affiliates Program, we were on their radar. Prof. Phatak from IITB immediately called them, arranged a meeting and flew to Bangalore. I did not have budget to fly, but I called them, sent them a presentation, invited them over to Kanpur, and there was no third department who approached them. They told me that since Prof. Phatak was most pro-active, IITB will get Rs. 50 lakhs, since I was next, we will get Rs. 35 lakhs, and some third department will get Rs. 15 lakhs.

This was really the turning point. We had a "Verifone Lab" in the department for the next few years, which could take care of all our requirements. And within that time frame, thanks for Pankaj's stint in Infosys and knowing Mr. Narayan Murthy, we received a very generous donation from Mr. Murthy.

But before that, other problems had to be faced. We were suddenly told that there is very little money for MTech and PhD fellowships. So we could only admit so many PG students with assistantship. If we wanted to admit more students, they would have to be on "self financing" basis, which means that they won't get any assistantship and their tuition was also higher. But, if we could get industry to sponsor their assistantship, they will continue to pay lower tuition, and the department would get a matching grant for another MTech student. With my credentials as the fund-raiser of the department, I was put on the job. And we got more industry fellowships than all other departments combined. So much so, we did not have to reduce our admissions by even one student, and the Institute said that they did not have enough money to give matching grant.

Another major funding happened by IBM. They wanted to get into the training business which was completely dominated by NIIT and ApTech at that time. So they prepared a curriculum and approached us to develop coursework for them with fairly generous terms, both in terms of initial money, and royalty over the next few years. Normally, we wouldn't accept such an assignment. But the money was so much that it could transform our labs. More than half the faculty was involved in developing course material for IBM. (They started this business, but couldn't succeed, and later used those notes outside India, and still gave us royalty from the money they earned in other countries.) Now, after taking care of all department costs, and Institute overheads, there was still a lot of money left. As per the norms, this was personal money. All of us who had build that course work sat together and decided that we will take a small amount of honorarium and donate the rest to the department. What we donated to the department was more than 6 months' salary for me (I think it was closer to a year). And we did that because our labs and other infrastructure despite Verifone lab was still considered inadequate, and we weren't going to get any money from the Institute.

Let that sink in too. More than half the faculty members donate several months' salaries to build department labs. I don't know if this has ever happened anywhere else in the world. That was the commitment of the faculty towards academics, and I was and still am proud to be part of such a group.

Of course, once we had the gift by Mr. Murthy, there was no looking back. Once we did not have to worry about small moneys, the focus shifted to research. Not that we didn't do research earlier. My group was the first one in the world to have a minimal working implementation of IPv6 for Linux. Unfortunately, we couldn't carry that forward for lack of resources.

We were housed in the limited space on the first floor of Computer Center, since a separate program was created in beginning of 70s. As the department grew, there was very little space for anything. I shared my office with another faculty member for a few years. I don't think any other department had that problem. Very limited space for keeping lab PCs. No space for PhD students. No meeting room. Only one lecture room.

We were overjoyed when we heard that some of Mr. Murthy's donation would be used for a new building that will house CSE Department. Pankaj Jalote was the head, and he involved each one of us in deciding our specifications, interacting with architect, and later interacting with the works department, and also for the interiors. Of course, he wanted this to be a modest building, a modern building but no luxuries. And what came out was one of the best designed in terms of usability of every inch of space. This allowed us to expand PG programs, faculty, new labs, projects, and everything else. PhD students had their own personal spaces now.

 I was extremely lucky that Pravin Bhagwat and Bhaskar Raman decided to join our department. Together, we received a huge Media Lab Asia grant to build the world's longest multi-hop network based on WiFi. This was my first time to work with poverty and rural India. It was so surprising that there were places near Kanpur-Lucknow highway where the nearest public telephone booth (PCO) would be a few KM away. And, of course, the nearest Internet cafe was perhaps 10-20 KM away, or even more. Poor persons were willing to pay double or triple the BSNL charges for making a phone call to their son in Mumbai, since their only other option was to lose half a days' wages by going to that PCO in the other village. We also did research on how poor people consumed Internet by setting up an Internet cafe in a village not too far from IITK campus. This was probably the most exciting times in the department.

Later, me and Deepak Gupta worked with some not-to-be-named agencies on monitoring Internet. This was a self-inflicted disaster. We worked on it for 6 years and at any point in time had a product far superior to what the Indian intelligence organizations were using at that time. But in every review, they would tell us that if we could add one more feature, they would start using our stuff. This was the first time I was getting exposed to the politics. Product development does not result in publications, and in any case, they strongly discouraged us from publishing whatever we could potentially publish. After 6 years, we realized that they had no intentions of ever using our work, and we called it off, but the damage was done.

One day, we were sitting in the department and thinking how to improve PhD program. The outcome was that there aren't enough students applying for PhD not just in IITK but across IITs. So we need to tell students about the advantages of doing Phd. Just the previous year, IITB had arranged such talks in many colleges, and we felt that we should do something similar. With the department now very rich could easily afford to go around and give talks. I volunteered. I was given one semester off. Every week, I will choose three colleges where I could visit on Thursday, Friday and Saturday. The criteria was that I should be able to go from 1st to 2nd and 2nd to 3rd college either by a couple of hours of car journey, or by an overnight train. 12 weeks, 36 colleges. I thought what I should do for the whole day besides giving a lecture on why they should do PhD. I came up with a plan that included one technical talk, one career talk, meeting with students, meeting with faculty, meeting with leadership, etc. I used this opportunity to understand technical education. What are the issues our colleges are facing. What kind of curriculum they follow. What is the pedagogy. How do they hire faculty. And so on. I realized that while resources were a big issue in many colleges, autonomy was an even bigger problem, and in many cases, it was just the lack of exposure to best practices. We created a program under which any faculty member of any CSE department could spend a month at IIT Kanpur. We will take care of local hospitality and the person could sit in classes, sit in some of our meetings, meet any faculty and find out how, in general, we did what we did. Unfortunately, never took off, and we had less than just a few visitors from one college from South India. The colleges did not want their faculty to be away for even a day, and certainly not when they had to pay for it.

What has always been interesting in CSE department is that I could go around and make commitments, even those commitments which will have financial implications, and the department would always support those commitments. They supported me not just when we had good bit of money but when we had very little money. The trust between the colleagues was the hallmark of the department.

After this semester of traveling, I realized that very small things can improve the quality of education in India, at least CS education at college level. And I started writing articles in media, my website, and later started my blog. The reaction to most of what I wrote was that all this is good in theory but won't work in practice. That despite that semester on travel, I still haven't understood Indian higher education. So when I received the offer of leading LNMIIT, I grabbed it with both hands, and the institute became my lab. In 2 years, we did a lot of interesting things and looking at the success, I got even more convinced of my ideas.

One of the big part of department culture is its faculty tea room, a place we could go to any time and help ourselves with a cup of tea or coffee, all at the department cost. This was the place where we could discuss our ideas, discuss issues related to courses, students, and what not. It was a comfortable place to have discussions with a guest. But alas, one of the recent Heads stopped free tea and now we need to pay. This has ensured that only a few coffee addicts like me go there, and none of the department issues get discussed there (because what is the point of discussing with a small group, when a larger group would take a decision later on).

24 years is a long time. Lots of memories. I have not written about students at all. But it is already too long. May be if I am around next year, will write about students. Thanks for reading.

Wednesday, September 20, 2017

Should we pamper our best institutions

This is not my typical blog where I write my views and we can have some discussion through comments (or outside the blog like on FaceBook). This is an issue that I am very confused about and am seeking clarity from my readers.

I attended a meeting of BRICS university admins in Brazil last summer. In the Brazilian presentation, they raised an issue. They pointed out that most countries were raising the budgets of their top few universities substantially so that they have greater representation in top 100 ranked universities in the world. Certainly China, Russia and South Africa had done that. India was thinking about it, and so was Brazil. And of course, lots of other countries have used this strategy to have more of their universities in top 100 or top 500. The Brazilian delegation was concerned that a very large amount of money will be used for very small number of students and faculty for a questionable goal like getting into a top 100 rank, while the same money could be used to improve the quality of education across a much larger set of institutions.

Why am I remembering this. Well, MHRD has a couple of schemes of this nature, which were designed to encourage the better institutes to get into top 100 or 200 ranks. And one of them, Vishwajeet, which was designed to help the older 7 IITs to get into top 100 ranking, has been disallowed by the Finance Ministry. This news a couple of weeks ago reignited my thought. Of course, another scheme called "Institutes of Eminence" is being continued where the Government will select 10 private and 10 government institutes. The government institutes will get about Rs. 1000 crores each over the next 10 years. And all these 20 institutes will get a lot of autonomy.

Of course, the debate in India may not be very important since the money involved is very small. Vishwajeet scheme was to cost Rs. 8700 crores over 7 years, and Institutes of Eminence scheme will cost Rs. 10,000 crores over 10 years. The first one is anyway discarded, and the second one has a tiny budget compared to overall education budget of the country. But the debate is still important because IITs will continue to push the Government to keep giving them additional funds under one pretext or the other.

So, here are the arguments that I have heard.

First, the additional funds is a mirage. The year that such schemes start, the government may budget higher amounts, but then not all the money is released/spent and the increase every year is less than inflation, so after a few years, the top institutions who were anyway getting a higher share of the budget start getting even higher share of the budget and the poorer institutions get their budgets reduced in real terms.

Second, while there can be substantial gaps in the research infrastructure of different universities, there should not be very large gaps in the teaching and learning quality of universities. If reduction of budget of the second level universities start affecting their teaching programs, it would be bad for the society, since most citizens get trained in these institutions and not in the top institutes. Also gap between top institutes and the next set (as it exists in India) causes too much stress for admission to the top institutes.

Third, the ranking is not a worthwhile goal to spend lots of money on. Ranking, by itself, does not give any benefits. There should be a mechanism to continuously improve the quality of our educational institutions and ranking should only be a by product. So even if you have more money, think of how you would improve the quality of whatever is important for that institute.

The opposite viewpoint that I have heard is:

Ranking is important. Higher ranks attract foreign students, foreign faculty, which allows a country to project its soft power. Having higher ranked universities give more prestige to the country, and it becomes easier to attract investments particularly in high technology areas where trained manpower is a critical input.

Now, my questions for the readers:

1. Are there other reasons for or against substantial hike in funding of top institutions.

2. In light of these arguments, should Indian government come up with another version of Vishwajeet (perhaps not restricted to IITs, choose 5-7 institutions across the board, and perhaps an even higher budget than what was proposed).

Tuesday, September 5, 2017

Discipline and Security in a campus

Last week has been terrible at IIT Kanpur. (For account of just one incident, read this blogpost.) I am constantly hearing of students being referred to as criminals and thieves (by those responsible for security policies on campus) and it pains me. I was a student here myself a few decades ago. I can't imagine that boy who lives in the room in which I lived to be very different from me. What has gone wrong?

To understand what has gone wrong, one has to look at the annual reports of our disciplinary committee (SSAC), particularly since 2006. For many of the years, you would find that SSAC (Senate Student Affairs Committee) did not have to recommend any punishment other than a warning to any student at all, and in some years, you may find 1-2 students being punished mildly. In some instance, if Senate took a harsh stand in one meeting, it reconsidered that decision a few meetings later. These annual reports are telling us that we have a fantastic student body, and even though we may have 6000 students on campus, not even 6 of them did anything wrong of serious nature.

If our students are so well disciplined, how come the narrative of security establishment is so different.

What has really happened is that SSAC has acted irresponsibly all these years, and not given appropriate punishments (or given no punishments in most serious cases). Take an example of copying in exams. Any student in the Institute will tell you that copying is a serious problem in IIT Kanpur. Most faculty members also believe the same, though many of them would argue that checking for copying isn't their duty. Fair enough. But what about cases in which someone is caught red handed.

Since faculty cannot really wash its hands off from this, many of them come up with different strategies for "preventing" copying. It is said that unless we have done enough on prevention, we have no business punishing anyone. How do you prevent? Well, let there be CCTV cameras in all lecture halls. Let the number of invigilators be increased. Let the question paper consists of questions whose answers will be difficult to find by searching on Google through the hidden mobile phone. We should use jammers and disconnect WiFi access points during the exams. There should be a security guard outside each toilet who should ensure that only one person goes in at a time. And many more. Basically convert the campus into a police zone for the duration of the exams.

If you did all this, would you reduce copying, and would you be able to dismantle the security apparatus after the exam. Of course, not. If a student knows that the chances of his getting caught have gone up slightly, but he also knows that even if he is caught, there will be no action, the copying isn't going to reduce at all. Prevention can work only in conjunction with a justice system.

If we have a complaint of a girl of a boy misbehaving with her, and you are able to catch the boy, establish the incident and yet give him just the warning, what signal are you sending. Now, prevention would mean that you place cameras at every intersection, in every lane, in every floor of every building. The number of guards is increased. But does it reduce harassment. Of course, not. You have only increased the probability of catching the boy. But if he knows that nothing will ever happen to him even if caught, what does prevention strategy achieve.

The problem with this strategy is that we are putting most of our eggs in one basket - the security forces.

Since we have "prevention" as the only strategy, we start criticizing security the moment some incident happens. What does an average guard do in response. Well, as soon as there is slightest suspicion, he calls the control room and asks for additional security. When two security persons go out to the location, the guard on the location has to exaggerate the situation to justify why he called the control room, and the situation goes out of hand from there.

Also, the security guards are typically illiterate. They are not able to or willing to argue with students or community members. Whenever someone argues, their only option is to inform the control room and bring more forces. Arguing has now become a crime for which you will be taken to the control room, abused, threatened, asked to write a letter of apology which the security establishment can later show as a proof that the student was indeed a "criminal." Who wouldn't write such a letter after being in solitary confinement for a couple of hours.

There is a relatively recent rule that you need to show an I-card to enter a hostel. If someone has forgotten one, he is harassed. I asked a security committee member, if we can have a more humane approach. Can we not let some other student identify this student. Can we not have some small fine for this. Can we not have a biometric based entry system if someone loses the I card. (By asking this, I am not supporting the rule to carry I card all the time, nor do I want a large scale collection of biometrics, which has its own privacy implications.) And the response was that none of this is security's business. Security will only do what is easy to implement. And what is easy to implement is that if a student does not have an I-card, abuse him for an hour or two.

Why has Security Raj become more visible in the last few months. Well, it is a reaction to the student agitation of August 2016. A weak administration can not find out if there were students who crossed a line, and punish them. Even if you believe that it was a very emotional period and everything that happened was due to emotional outbursts, a weak person can not forget and forgive. So if you don't have guts to punish administratively, and if you are not strong enough to forgive, what will you do. Well, when the new security committee came about, you would put those who are well known for their harsh attitudes and let them take a revenge on the students. Again, a proper investigation, and a proportionate punishment against a few specific students who put Institute property on fire would have been far better than harassing a much larger number of students who have done very little (like forgetting an I card).

Establishment of security raj on IIT Kanpur campus has followed the destruction of administrative mechanisms of maintaining discipline. And the only way to dismantle this Raj is to strengthen the civil administration. The civilian processes are more democratic. The disciplinary committee that I mentioned above has students as almost half the members and has representation of women and other groups. There is much greater trust in such a system, and if someone is punished by this system, it acts as a deterrent for many others. But when the security Raj punishes anyone harshly, it is seen as unfair and arbitrary. It can lead to more people fearing security in general, but it does not lead to a linkage between the crime and punishment in people's mind. And in that sense, it does not act as a deterrent for that specific act.

In those rare instances where, primarily due to community pressure, SSAC has recommended some punishment, students have generally opposed any punishment. They are so used to SSAC being ineffective for so long. But they don't realize that their opposition to SSAC recommendations is leading to security Raj in the campus which is worse for them and everyone else in the community. The security Raj is loved by the Director. For any administrative mechanism, the buck stops with him. He does not want to be seen as approving any recommendation of SSAC which gives a real punishment. On the other hand, in cases of harassment by security, no one believes that Director could have prevented it. So he can continue to present an image of a student-friendly Director, while at the same time, ensuring harassment for all "criminal" students.

So while we can all criticize the specific examples of excessive use of force, and we must ask for heads to roll when such things happen, but we must not lose sight of the reasons why Security Raj has come about on the campus, and we must strengthen democratic administrative processes for ensuring discipline, something the last two directors have undermined.