in reply to Efficient giant hashes

Thanks all for your comments, they have been very helpful. My colleague is sure that the slowness of the script is due to using (big) hashes of hashes of hashes, and that it would be much faster to reimplement the system in C. I'm not so sure that would be worth the effort, and would like to do some profiling first to find out what's going on. Your comments are good ammunition for that :-)

Replies are listed 'Best First'.
Re^2: Efficient giant hashes
by Avox (Sexton) on Mar 11, 2005 at 17:01 UTC
    I think you're the smarter of the two of you. Re-writes in my experience should be a last resort. There is so much history attached to legacy products that it is nearly impossible to incorporate the years of bug fixes and gotchas that previous employee's have packed in.

    I'm not saying its never warranted, but it certainly shouldn't be one of the first options IMO.