[REBOL] Re: RFC: Cross-language benchmark proposal
From: greggirwin:mindspring at: 9-Nov-2002 12:27
Hi Joel,
>> >Option 1
>> >>> >
>> >Goal: Demonstrate to the world that REBOL is a viable,
>> > competitive language for many common programming tasks
>> > (where "competitive" is defined in terms of run time
>> > performance).
>> >
>> >Audience: Developers at large.
>> >
I'm probably just not saying what I'm thinking very clearly, or my
mind was on a different track (read - derailed :).
In any case, yes, the above goal is stated clearly, and your other
replies have helped also. My question was targeted at the
implementation of the test (custom versus standard) with this overall
goal in mind. I.e. we have to sub-goals.
JN> Where do you see a "disadvantage"? Of the languages I proposed...
Sorry, I was thinking in more general terms than just the languages
currently proposed. However, Java is at least byte-code compiled in
all cases, correct? It may be JIT'd as well. C is the reference point,
so it is a special case IMO (i.e. how fast *could* something run).
JN> I don't have any firm preconceived notions about what the results will
JN> look like (except for C usually being fastest), but let's not be too
JN> quick to wave the white flag just because REBOL isn't the fastest
JN> language for all tests! Brute speed is only one of many factors that
JN> go into a language decision, and for 90% of what I do, "fast enough"
JN> is the key issue, rather than "fastest".
Agreed. No white flags here.
JN> It's interesting that I would have reversed the labels. My original
JN> intent was for the tasks to be small but realistic ("applied") in
JN> nature, and the "standard" solution to be the kind of thing that a
JN> reasonable programmer (though not a Grand Master of any particular
JN> language) would think to write. The custom solutions, where the
JN> esoterica of a particular language's world view emerge, seems much
JN> more "academic" to me.
My thinking is that the "acadaemic" version refers to something that is worth
knowing (how it performs against a particular algorithm) but isn't
what you would actually use on a day-to-day basis. "applied" is what
you do (e.g. calling SORT rather than writing your own -
notwithstanding special cases where SORT just won't do). Maybe the
esoteric solutions are a third category, though I don't know how much
we want to fragment things. I don't know that esoteric solutions are
really useful in this context. If you're talking about solving
problems in unique and interesting ways, yes, but not here.
--Gregg
