Perrin Harkins wrote:
Several people have brought up benchmarking in reference to the pet store. I don't think it will possible to do a good benchmark of this application, partly because it's so big (it's a reference app that uses lots of functionality just to demonstrate it) and partly because it's well known that the J2EE pet store performs badly. It does not represent anyone's best efforts to make a high-performance Java store.
An excellent point.
If people are more concerned with seeing something that would dispel myths about Perl performance, rather than a talk on feature portability from J2EE to Perl, I could look at implementing something that really can be benchmarked like the TPC-W spec or the Doculabs Nile Bookstore benchmark. These would be more comparable to existing Java and .NET performance tests.
The saliva begins to leak from my lips...
Personally it would warm my heart to help enable a press release saying something like "Perl blows away previous price/performance leaders on TPC-W benchmark", but I don't know if hearing about that would be as interesting to people as the other things I proposed.
Oh yes, now this is more like it.
Regardless, I think that posting a good reference implementation of one of these specs might get mod_perl some good attention from the business-oriented mags that usually focus on Java, and would be a valuable marketing tool.I think I've just had an orgasm. ;-)
Perrin, you've probably gathered by now that IMHO you've struck gold here. I honestly don't know why this hasn't been done before. Obviously it would be great for all mod_perl programmers to be able to direct their PHBs and/or clients to a paper that validates and justifies the use of mod_perl. I, for one, really hope you pursue this.
--
Jonathan M. Hollin
Technical Director: Digital-Word Co. (http://digital-word.com/)
Co-ordinator: WYPUG (http://wypug.pm.org/)
