Re: To Alex



[some of my posts seem to be going to /dev/nul - apologies if this shows
up twice!]

Betov wrote:

....
You might miss the other significative point, that is
that, since one week, or so, you are talking about
optimizing a trivial Chunk of Source, that, much probably,
nobody will ever have any use for, and that, when you will
be over with having a version that could execute, say,
- let's be "optimist"... - twice as fast as the most
trivial way... that is, taking 1 second, instead of 2...
you will have saved one second, in... one week.

That's very true, but perhaps you misunderstand my motivation. I'm not in a hurry to find out how many upper case vowels are in my file. I want to find out what's fast and what's slow.

I prefer saving one week in one second.

I prefer to think of it as an investment. As we both know, we disagree about what's "worthwhile" in program optimization. What's "worthwhile" depends, of course, on cost and benefit. The benefit, in any one program, is quite small. My view is that if I can learn to write small, fast code as a matter of "habit", the cost will eventually become near zero - and I'll get the small benefit in *every* program I write.

It might be worth noting that while it's been a week or more that I've
been fooling with this, I mess with it for a half-hour or an hour when
I'm in the mood. It ain't a full time job. If I had more
self-discipline, and stuck with it for hours at a time, I'd be much
farther along towards being able to write small, fast code as a matter
of habit.

Best,
Frank


.