Mostly falls right in line with my experence. I have some different opinions on internal and external (proxy) cashing methods then they express here. And I think their 'single big ha db' strategy is totally the wrong way to go, but I also don't know what their db structure looked like and why it was like that.
Good info for folks intrested in how the big-boy systems are structured.
Infact, I'm going to be looking at the modules they recommended for DB searching. Might be better then my current approach.
perl.com | Large-scale E-commerce Site with Apache and mod_perl