If MS were to open W7 - or as much of it as they can bearing in mind third-party rights, what would the FSF and others look for in it?
There's a chance a community might coalesce around the goal of supporting W7 and keeping it going. But that would be existing Windows users: noone from the Linux community is going to be motivated to take on a task that size when they don't even have a use for it!
Much more likely is that it may be raided for ideas: things other projects could usefully adopt. Some libraries (or parts) may be re-usable with minimum adaptation; other code might be far more trouble than it's worth to try and re-use, but still provoke valuable thoughts.
There are plenty of precedents. For example, in code Sun and IBM have opened. I was involved in snarfing ideas from Sun's webserver to enhance Apache capabiities: generally it was nothing mindblowing, just getting a new perspective begged the question "why aren't we doing this?" Similarly look at filesystems: jfs, xfs, zfs all originate with once-corporate Unixes. Or dtrace.
I expect there are valuable spare parts in Win7, too. As well as a lot of dross unlikely to interest anyone.
MS may indeed be discussing it internally. They'll want to think it through in rather more depth than the peanut gallery. FSF is just giving those at MS who like the idea a thumbs-up.