If you are relying on the sender to do the right thing, then why not force them to create proper message-ids?

I think Barry's proposal is essentially a numbers game - e.g. he's hoping for significantly better results using "Date" in the calculation than not using it.

http://wiki.list.org/display/DEV/Stable+URLs

I'll try to tease out some more useful stats from some large datasets this weekend. (I can't just run the python scripts as is because I don't have python 2.5 in the same place as the data, I don't keep raw message in mbox format, blah blah blah, but we'll figure it out).

My hypothesis is "Date" doesn't really buy much, but that's in part because I have a vested interest in that outcome. We'll see how the data plays out. And I still think RFC2369 headers are needed in the calculation if cross posted messages are to be handled correctly.

Jeff