It occurred to me today that bzip2 seems to have lost the VHS vs BETA battle. Not enough tools to work with it, not enough benefits from using it... Seems that VRML may be another example of this, and I suppose we could add IGES to that list, along with possibly STEP (though the steel industry's use of this in SDS2 seems to offer some hope.) It all just got me thinking that we should be aware of rules of the game as we're inventing our new format (though a new format may not even be the end result of all of this.) The tools have got to be many and attractive. I know I've mentioned it before, but I don't mind repeating it: I'm out to slay the entire notion of a file format. Instead of thinking about files and formats, users just need to think about programs communicating. Sure, that is currently possible, but it almost always involves clicking file->import, file->export, start new program, click file->import, ... cilck, click, re-label information. This is what we get from file formats. I don't really care how we get there, whether it involves creating a new format and tools to go with it, encapsulating old formats, just creating tools for existing formats, ... Now the trick is, no matter how we go about it, creating this new way of working is guaranteed to drag us into the VHS vs BETA race. Look at Linux vs Windows right now (this is a little different, because Linux isn't going to die out) and there are some other historical precedents (Intel vs Motorola.) It's possible that we won't know which is which until the dust clears, but just remember: quality isn't the only selling point. I'm not trying to be discouraging here, but I think everyone should know and respect the dangers of the cheap plastic competition, advertising saturation, defacto-standards and one-button mice. --Eric