Hi Trees,
Did the VGA experiment reveal anything?
The disconnected-connected sounds Emre is hearing when powering on the monitor makes me wonder if there's something besides just EDID polling going on.
Apparently DVI has a dedicated pin for hot plug detection, so perhaps it could somehow be involved. It could explain why the EDID override trick is not working. Perhaps the EDID is always correct, but the hot plug pin state changes when powering on the monitor, making Windows think the monitor is temporarily unplugged. The "Plug and Play Specification" over here may hold some more clues:
http://www.ddwg.org/lib/dvi_10.pdf
My guess is the MS support staff is just as or even more clueless than we are. They are only allowed to use approved replies from the database, and they've already done that. The software engineers that would instantly understand what we are talking about don't read these forums, and even if they did they would not be allowed to comment.