I've just finished watching the video up on channel9.msdn.com about Sparkle and must say I'm extremely impressed, but who would've expected any less with people like Sam Wan and Manuel Clement on the team! :) Its obvious that these guys have put in a heck of a lot of work in the IDE and it really shows. They've no doubt also looked at Flash very closely and identified a number of key areas that could be improved. The timeline with all the properties looks pretty much like the one you see in Adobe AfterEffects (and if I was a gambling man I'd put money on us seeing that in future Flash versions as well).
What was funny is the way the camera guy reacted with amazement to some things we're all so familiar with for years now: "so you mean, you can just scrub the timeline and it updates the screen in real-time!?"
Great to see they've got 3D rotation and a camera view in there, also a lot of things like scaling parts of the layout is handled really well (see scale9 in Flash 8 but with a much nicer way to set it in the IDE). The data binding with wizards and live previews was also very cool. Again nothing you can't do in Flash in one way or another but it sure beats writing custom cellrenderer classes.
Now for the bad news, from the demo it looked like you actually need to deploy multiple files to your server (at least three, plus any graphic libraries you have compiled) if you compile it as a web application. It should run on all types of servers but they only talked about playing it back in IE7 -- I should hope they'll develop a plugin for other browsers as well or I can't see it taking off at all.
If I were to describe Sparkle in terms of Macromedia products, I'd say it was Flex Builder for traditional Flash content -- a solid IDE for designing a code-based interface with the expressive freedom you get in Flash.
The one area I think Sparkle has the edge is deploying to the desktop, the Macromedia Central initiative doesn't appear to be very active at the moment though it certainly had its merits. I think the defining factor in this epic battle will be desktop integration and the future of MXML.
Now for the question everyone is asking, will it kill Flash? I don't believe it will by a long shot, though they may manage to take on a considerable chunk of the enterprise market depending on how they play their cards and how Flex will develop as a technology. If anything a bit of competition will do a world of good for the Flash platform. Can't wait to see what Zorn will bring us and if all else fails we've still got a very active open source community to fall back on :D