discussion of the Creative Commons Metadata work

[Lucas Gonze <lgonze AT panix.com>]

> A regular expression like /\s*(\S+)\s*$/ would very forgivingly extract
> the license claim URL from TCOP.

It's worth worrying about how to parse out the URL, given that URLs are
fairly flexible objects.  It seems like a bad thing to have programs doing
logic like:

  Pick up last chunk of the TCP string that resembles a URL.  Test out
  whether this is a URL by trying to fetch it.  If there's nothing that
  resembles a URL or the URL fetch fails, assume there's no license claim
  URL.

There should be flag bytes to eliminate the fuzziness.  For example if
"Verify at " exists in TCOP, everything after it must be the license claim
URL, and if there is a license claim URL then it must be preceded by
"Verify at ".

- Lucas