Thursday, February 23, 2012

My Computing - 5

========================================================================

So there I was at IIT KGP at a tender age of innocence trying to learn Fortran II to help my ex-colleagues at AU. With a live problem in hand as an example, I wanted a bread and butter start-up.

Friends told me that Dr N of the Math Dept is a wizard. I knew him as a nice man, our age; so one fine Sunday morning I knocked at his home-door and told him the purpose of my visit. He asked me to come in and explain my problem (now I know it was essentially summing a series and diagonalizing the resultant 3x3 matrix...but not then). So I beat around the bush and showed him the piece of paper on which I wrote the steps of algebra.

He took a pen and a paper and wrote down his Fortran II program in ten minutes and handed it to me asking me to go to the IBM 1620 Palace and they will tell me how to proceed further. I left his place dissatisfied since he was so curt and I was afraid to ask him to explain what all that Jabberwocky of DIM and DO and SQRT was all about.

And after a couple of days of hunting for someone more openfisted, I was told to go to Prof Venkataratnam (KV). I asked them if he too belonged to the Math Department. Oh, no, they said, "He is from EE and a fantastic teacher."

Now that we are on this topic let me state the "3 Golden Rules of gps" learned from a lifetime at AU & KGP:

Golden Rule # 1:

If you are a beginner trying to learn bread & butter Mathematics, go to a Physicist.

For instance, I wanted to learn the Vector Calculus used in Physics. I picked up a couple of books on the subject stacked in the Math Section (515.112). And threw them on the desk...they were all abstruse, like if you ask the simplest path from hand to mouth they will start from the definition of geodesic and affine connection. And picked up Reitz & Milford's and Panofsky & Phillips' EM books and their Appendixes gave me all I needed to know about all those vector calculus theorems and the expressions for grad, div, curl in all the coordinate systems that a sensible chap wants.

Likewise I wanted to learn the Tensor Calculus needed in Physics and went to my good friend Dr SGH of the Math Dept. He took his used packet of Charminar cigarettes, turned it inside out and scribbled an A with ten superscripts and eleven subscripts and wrote down its transformation formula using twenty one or so wicked partial derivatives with sub- and super-scripts and discovered that there is no more space left on the wrapper. And he wrote the rest in the 3-D space around us. He looked forlorn that there were only 26 or so measly alphabets in English or even Greek and would have loved to use his Marathi alphabets (my pristine Telugu had 60 at one time before they omitted ten or more). He would have loved E J Post's system where he invented a tensor notation with superscripts and subscripts on superscripts and vice virtusa till the APS Journals' printers gheraoed their editors and they had to banish them in their Instructions to Prospective Authors.

I learned my tensors from Landau & Lifshitz's Classical Theory of Fields (1st edition, now out of print...the subsequent editions are no good for beginners).

Golden Rule # 2:

If you are a beginner trying to learn bread & butter Physics, go to a Chemist.

Our teachers of QM at AU were beginners in the subject and stayed so. They must have been too scared of Schiff and so they prescribed the book by two Chemists: Pauling & Wilson. That was good enough for them and us because it avoided abstract things like kets and bras of various brands and stuff like Hidden Variables...if they are hidden they should stay hidden, like if a singularity is removable, it should be removed...

And when they had to teach us Classical and Quantum Stat Mech, they preferred Samuel Glasstone's Theoretical Chemistry...although it had all those grand and not so grand canonical ensembles and the wicked ergodic hypothesis, they were clearly winking at them.

Golden Rule #3:

If you are a beginner trying to learn bread & butter Chemistry, go to a Psychiatrist ;-)

=========================================================================

1 comment:

Varun N. Achar said...

Varun's Golden Rule: If you are a beginner trying to learn bread-and-butter (or even bacon-and-eggs) x, where x is one of:

Vector calculus
Classical field theory
Boundary condition tricks and other skulduggery such as the image method
Electromagnetism (including such advanced topics as the quantization of electric charge)
Tensor calculus
Numerical methods for PDEs

... then go to John David Jackson!