--- Ronald J Kimball <[EMAIL PROTECTED]> wrote: > > Their purpose was performance. Perl was a bit far down the list of results, > > IMHO. However, I thought that it should all be squeezed onto one line at the > > same time as making it faster through regexp trickery perhaps. > > For example, it took half a second of thought to work out how to double > > the speed of the "hash" perl program, for example, so they've really not > > ever tried to make the perl fast. Shameful distrespect for the language! > > "For this test, each program should be implemented in the same way." > > http://www.bagley.org/~doug/shootout/method.shtml#sameway > > Tests That Call For Implementations To Be Written The Same Way > > For some tests, I will specify that they be written using the same logic > and data structures. The goal of this kind of test is to try to measure > languages doing the same operations, as closely as possible. (Since > functional languages have such a different mode of expression, I allow > them more leeway). > > I find this kind of test useful when I am considering questions like "Is > array subscripting faster in Perl or Python?", or "Are hash table > insert/lookup operations faster in Tcl or Ruby?". An example of this > kind is the Sieve of Eratosthenes test.
Such that one set of languages uses fixed length vector with a true sieve (stepping forward at regular intervals to cross elements off), and another set of languages uses a list of integers and performs trial-division on each one to prune the list. If that's in /any way/ "the same way", then I'm a Dutchman. > Since the purpose of the same way tests is to try to compare, side by > side, the same kind of operation in one language as in another, they > often use code that is naive and unidiomatic by design. Show me any other language creates a temporary list of all numbers it will look for, then reverses it, and then iterates down it in the "hash" test. That's a _unique_ feature of the Perl version, and slows it down by a factor of two. If for(;;) is good enough for other modern algoloids, then it should be good enough for Perl too. Deliberately crippling languages should not be in their mandate. Also note that the regexp ones had _different_ regexps in some of the different languages. Therefore if this test was supposed to be "the same way", then I see some regexp flexibility should being permissible. I think there's an extra 'un' and 'ma' in their third word from the end, and am convinced that the final word has nothing to do with what has taken place. Just because someone writes a section called "methodology", doesn't mean that it contains sense or anything. I just think that Perl has been given a very short shrift by some of these tests. Phil ===== Given that Dubya has control of a such vast arsenal, I'm sure the most pressing issue on his mind is : Which bombs would Jesus drop? (-- "mm") __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com
