In answer to file sizes - I think you can split your data up into different files - so you have a file for people, products, jobs etc all linked from a sitemap file. Not sure how it works, or IF it works...
It sounded good to me so I had a go myself. I added an ror sitemap and other info to my site
www.speyside-developments.co.uk
when I first uploaded the site in April 2006. I checked just now (July 2006) on
www.rorseek.com - there was nothing there. Yet most other search engines have spidered the site. It sounds like a good idea in theory - or will it just get abused? Instead of keyword stuffing will we get ror stuffing?