Dear B.Tech/DD student,

When it comes to meta courses (no I am not talking about *that* dept :-) I mean a seminar, a mini project, or a project), the relationship between a student and a faculty member needs to be addressed. This document is about choosing guides based on my past experience as a b.tech/dd facad and my several years here at IIT. Important: Please correct any mistakes/mistaken impressions I have.

First of all, remember that both of us work together, and so it is important for the match to be a mutual consent.

In an ideal situation, only the student with the right (whatever that means) background will approach me, and that will exactly be equal to the number of students I can handle. And this will happen to every potential guide.

About the right background. For several years, I have been suggesting to the students that we should have an undergraduate course in computer graphics in the fifth semester. You should note a couple of points in this regard.

  1. Virtually every engg. college cse dept (other than the IITs) in India have an UG course in graphics (usually compulsory).
  2. If you look internationally then the right place to look is ACM (acm.org). They have literally dissected various CSE programs across the world. If you believe you should earn a computer science (or cs and engineering degree), then you should demand that you learn what the world expects you to learn -- I have extracted relevant pages from their latest curriculum recommendations. A new one is in the pipeline.
    See CS version INTRO . Intermediate courses.
    Also see CSE version courses. Note that a semester is about 17 weeks, and a six credit course is about 50 hours.

    Now compare your program at IITB. You will find that you have a lot more emphasis on theory in your first six semesters (DS, Formal Methods: second year), POPL, Algorithms, Theory of Computation, Compiler: third year) than what the ACM recommends.
    (Let me also list the systems courses. Databases, AI, OS in addition to the basic courses CS101, CS152 Abstractions (though this might also be theoretical), Advanced programming and Data Structures.)

So what's this? IITB seems to do things differently? Is that necessarily bad?

Hold on. Every local situation demands a re-orientation based on available faculty. And yes, as a theory Ph.D, I love the theory courses. However, should all of you be doing this?

Anyway, now that you have NOT done an UG course in graphics, what's the best bet? Well not much. You'll have to pick up the material on your own which is why you need to spend more time with me (if you are going to work with me). But how are you going to do it given that in summer you plan to go outside IIT (I strongly urge this!)?

Well that is why I'm going to insist this year (2006) that

  1. If you are a b.tech student, you must have an extremely good academic record, or some evidence of work in the area I do, AND agree to do a b.tech project with me in the 7th sem (by default).
  2. If you are a DD student, you agree to do a miniproject (either 6 credit in your 7th sem or 18 credit, preferably the former) by default.
What do I mean "by default"? I mean that if you want to opt out after seminar, you can opt out -- no compulsions. But by default you will be in, and I will proceed under this assumption. If you opt out, please let me know no later than a week AFTER you receive your seminar grades. Also, note that under normal situations, I do not expect you to come for a project unless you have done a seminar with me. I will also expect you to be taking my classes in image processing (odd sem), and Graphics II/computer vision (even semester). You can find a description of these courses on my web page.

There is one more practical matter to consider.

But I don't know what I am interested in, you say. I don't know the professor's interests and whether it will all work out if I commit. Right. Nobody can argue that you should not do your homework; that you just plunge in. In fact DD students have exactly this luxury.

The key problem is that there are limited resources. Let's say you figured out after a seminar, many courses, and halfway through the next semester that Area Z is what you want to spend time. Great. You might have done this on various considerations. How tough the area is. What the job prospects are. How bad the prof in Area X is. And so on.

Now ask yourself. What's the chance that your friend will not have come to the same conclusion? Is he or she likely to be too different from you in terms of job prospects, good/bad prof and so on. Since both of you came through JEE with marginal difference in quality, the chances are that you will think alike on this topic rather than differently

And your friend might have done a seminar in Area Z. Who do you think the prof in Area Z will pick?

Or you might have done in Area Z and she might not have :-)

Of course individual skills, interests might differ. So do the homework but remember that changing next semester might be too late. There is of course the risk that you might not have enough time to do the homework AND you made a bad choice. You could of course opt out in this case.

Thanks for reading

Sharat