Monday, September 19, 2005

Lugging books on treks has its rewards

On the trek to Chandertal, that I had to give up on due to bad weather, I carried 3 books with me. One of them was a trekking guide to the area, a novel and the third one was a collection of papers by Knuth. I had to think about carrying these -- I did not want to lug around extra weight unless I was pretty sure I was going to need it. And surely enough, on the last mile before Hampta pass, I felt the weight of every single ounce (that seemed to become heavier with every step). What clinched the argument was the fact that I was getting horses to carry most of the stuff, so they could easily carry these books too.

However, when I had to stop hours at end waiting for the storm to go away, these books came in very handy. I devoured the trekking guide book many times over, and every time I found it lacking the detail I was looking for. The novel was the next to go. For the last two days, I went through the book by Knuth and spent quite a bit of time trying to solve some of the problems I encountered in it. All said, I was very, very happy to have got those books with me -- they kept me good company in an otherwise lonely tent. :-)

The trekking guide I used was Trekking in the Indian Himalaya by Harish Kapadia. When I first leafed through it I found the book quite good. It helped me choose the trek and plan for the number of days. But that's where the usefulness of the book ended. Once I started on the trek, I realized that the plan that the author laid out, similar to the one charted by my guides, was for a lazy stroll and not a trek. I mean, we started around 9am the first day, and we had already put up tents by 1pm. So I pushed the guide into squeezing two days' worth of trekking into the second day, and things looked much better. The book was full of errors, had the order of campsites wrong, had elevations wrong (surprisingly, the text didn't match the figures as well). I think the author probably spent some time in the town of Manali, sat down with some guides and in a week had slapped some stuff together and sent it for publishing. Quite a letdown.

The novel I was carrying with me was the very fat Cryptonomicon. And fat, of course, also means heavy. But this book was what made passing time real easy. It's a very geeky book, what with encryption algorithms and perl scripts in the middle, talk of Riemann zeta functions and an appendix by Bruce Schneier. It was a blast, though. I thought it turned into a bit too much of fantasy towards the end, but till then it was a great ride. It did make me more curious about some of the Maths involved.

Once I was done with Cryptonomicon, I turned to Selected Papers on Computer Science by Knuth. Knuth is sometimes really difficult to read, just because of the sheer density of material in his text. These papers were at times much easier (some of them were almost like stories), and at times just as difficult to follow. But they were almost always interesting. In one of his papers, he discusses whether toy problems are useful. After recapping some older mathematicians' point of view, he jumps into some actual problems. One of the problems was given by a Stanford professor to his class in 1975 (or some such year): Write a program to divide the square roots of numbers 1 through 50 into two sets such that the sum of numbers in the one set is as equal as possible to the second set. Further the program should not take more than 10 seconds of computing time. Of course ten seconds of computing time was a lot less than what it is today, but the problem was interesting. I didn't have a computer at hand, but I tried solving it without the use of one. The result was a good amount of time doing simple algebra and some simple insights. I had a solution but I knew it was not the best. It did help me spend some time sitting in a dhaba, consuming glasses of milk tea and scribbling on a letter pad, while a snow storm brewed outside. Didn't someone say mathematicians are a machine for convertng coffee into theorems? The other interesting problem that I remember is the discussion about the hash tables and average cost of insertion with open hashing -- just the mathematics parts were interesting and sometimes difficult to follow.

Next time onwards, I am definitely going to carry some books on treks with me. I might consider, however, tearing a novel the size of Cryptonomicon into 3 or 4 parts and dividing the load across the horses. ;-)

No comments: