Archive.org has a better cache to use than Google, in most cases.
I would put the new site into Google Webmaster Tools and review the External Links report. That should tell you which pages have inbound links. You can ignore all the other pages.
You can either use the existing filenames or 301 them to new content. That depends a lot on how bad the old filenames were.
How quickly you redevelop the domain isn't as important as how long the domain was parked. But, the honest truth is that no one really knows what time frames are involved. I've redeveloped on domains that were parked for years and still kept their PR.
Optimum amount of content for PageRank? I'd say one page for every page which shows an external inbound link in Google Webmaster Tools.