Author: Sune Fischer
Date: 13:19:51 01/16/03
Go up one level in this thread
On January 16, 2003 at 12:10:07, David Rasmussen wrote: >On January 16, 2003 at 03:29:43, Sune Fischer wrote: > >>On January 15, 2003 at 21:41:19, Robert Hyatt wrote: >> >>>Saying "vincent exaggerates all the time" is like calling a cup of sand a >>>"desert". It is a gross understatement... exaggerate means to expand >>>something beyond its normal boundary. Most of what he says is not based on >>>any sort of factual evidence whatsoever, which makes it more fiction and >>>less exaggeration. >> >>I know there has been many discussions, the last one I remember was about >>functional languages. IIRC he said something about them being slow and gave an >>example with a program he wrote that was 2000 times slower than in C. >> >>Everyone disagreed, but from my (granted limited) experience he is right. >>It's like a hiarcy: asm, C, C++, java,... the more advanced the slower it is. > >What about everybody elses experiences? That's exactly the problem with Vincent. >He thinks nothing of other people or their experiences. > >More abstract languages tend to be slower in practice, sure. That's my >experience too (I would place C and C++ on the same level, speedwise, though). >But that was not what Vincent said. He hinted that functional languages are 2000 >times slower than C, which is nonsense, unless you're a really bad programmer. Maybe it's all a communication problem then :) I didn't read the same into it, to me it sounded like it was just an extreme example. >Given some C implementation of a checkers program, one can make a functional >equivalent, in a given functional language. But for the functional languages I >know (SML,Lisp,Scheme and others), this would not mean a 2000 time slow down. It >would mean something on the order of 10 times slower. For many calculation >intensive programs, 10 is even high. With Caml or other good functional language >implementations, it would mean a 2-4 factor slowdown. At the most. So what is >the reason of writing 2000? 2000 seems only to indicate that Vincent is a very >poor programmer of functional languages. Could be, I still think a factor 2-4 is huge, certainly it's too much when speed is of the essence, as in chess. IIRC there was also something about no one would write 10000 lines of code in a functional language. I agree with Vincent there too, I can't imagine anyone trying to write an OS is Lisp, for instance :) >>I guess there is not much of a theoretical reason why that should be true, but >>it just happens to be a fact most of the time. >> >>Besides that, I don't think anyone really talked about what a functional >>language is, so isn't it possible that too was a source of disagreement? >> > >What do you mean? Without a definition how do you sort the languages into functional and non-functional subsets? I lack a clear definition of "functional". Is C functional? I guess no (or else the discussion wouldn't make much sense). Is that the abstraction level is higher in some way, in what way? -S. >/David
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.