strangest systems I've sent email from

Raymond Wiker rwiker at gmail.com
Fri Apr 29 12:11:46 CDT 2016


> On 29 Apr 2016, at 19:03 , Swift Griggs <swiftgriggs at gmail.com> wrote:
> 
> On Fri, 29 Apr 2016, Raymond Wiker wrote:
>> The regular expression support in Perl is implemented in C, and are 
>> supposedly fairly fast.
> 
> They are faster than some, like Ruby and slower than others like 
> (apparently) LISP. 


It's not *generally* the case that cl-ppcre is faster than PCRE - it depends
completely on the (Common) Lisp implementation that it is running in.
> 
>> That didn't stop a Lisp programmer from implementing PCREs in Lisp (that 
>> supposedly slow and inefficient language),
> 
> Cool. Which LISP ? CL ?

The original benchmark was run using CMUCL, which is generally considered to
be a high-quality, fast implementation of Common Lisp. The benchmarks are not 
part of the cl-ppcre homepage anymore, but an old version can be found at 
the Wayback Machine <http://web.archive.org/web/20080624164217/http://weitz.de/cl-ppcre/#bench>.

>> and getting better performance than Perl :-)
> 
> Hehe, well, right on then.
> 
> My opinion is that benchmarking and subsequent proclamations using 
> scripting languages is like racing snails vs slime molds (my money is on 
> the snails, BTW). It's all fun until someone shows you a graph of the same 
> algorithm in C and puts a quarter-horse in the race. Then your saying to 
> yourself things like "Should I be 10x or 15x slower?" :-P
> 
> http://benchmarksgame.alioth.debian.org/u64q/which-programs-are-fastest.html
> 
> -Swift
> 



More information about the cctalk mailing list