Updates
July 31, 2007
I spent the last week and a bit replacing the XML parser. REXML proved to be too slow, and it’s replacement (libxml-ruby) is faring much better. For example, the searches that used to average ~4secs now take about 1! I’m also rewriting the item model and most of the javascript dealing with it. That should fix the unclickable recommendations as well as allow for a whole host of features like wishlist, Listmania, and my new feature: personal lists! But I get ahead of myself. Just a couple days more…
uncluttr 
August 2, 2007 at 6:39 pm
…Have you considered using why’s hpricot for the XML parsing?
A co-worker just started using, claiming that it blew rexml off the map, speed wise.
[shrug] Just a thought. I hadn’t considered it because it was a “web scraper”, but I guess it fares quite well.
August 3, 2007 at 12:46 am
yup, i did; our choice was either libxml-ruby or hpricot . libxml won for completely arbitrary reasons. I’m not sure which one of those two is strictly better, but so far, libxml has beat the pants off of REXML.