a couple of years ago I had a tool that would scan any page of text for domain names and produce a quick report ... expired, expiring, taken, for-sale, etc. It was written by a guy in Eastern Europe and I have looked hard for it. Standard story: new hard disk, can't remember the name of it, terabytes of data to deal with.
KEY to this thing: You didn't have to format the page, and if the page was HTML it would dig into the code for you - it did the parsing, and would remove everything but the TLDs. It would find all objects that were domains - this was the cool part (and what made it better than say, a regular multi-domain whois tool.) You could just paste a load of crap at it and get out good info. I guess it wouldn't be too hard to duplicate but I'd rather type gunzip than start coding at the moment.
The best time to use this was when you were on an old edu site and started finding broken links.
Does anyone remember this? Is there something else like it now?
Scott