You'll still need to put the formatting bit in to the SMIL to indicate it's video. That's something you can test now by getting a couple smaller videos and using them to test from phone to phone. I'm curious, is there an indicator for the type of video? Right now I know there are at least 2 popular formats for video on cells (3pg and mp4), and a ton of audio formats (acc, amr, mp3, mid, etc). Is there a SMIL spec for each type, just a generic "video" flag, or both?

Rescaling can be done lots of places (and honestly people are probably going to want to recale and clip a segment from a larger video). I think the bigger concern is getting the flagging right to be able to transmit a file after it's been massaged into the right format.

Great work as always, and looking forward to the 1.0 release! Let me know when it's ready and I can test/thumb-up it!