SMIL for professional video production

The usual method of video production is "destructive", which means that when you're done editing, compositing, "effecting", and synchronizing your media elements, you haven't preserved continuity with  the original source materials.  You've made something new and separate from the original source. What's more, the only descriptive path between the two is, at best, the proprietary EDL format from your editing system of choice. 

This isn't a huge problem in itself - we're used to that just being the way it is.  But with all the media increasingly available on the network, is there a way to do more editing, compositing, "effecting", and synchronizing by reference rather than destuctively?  What efficiencies will this enable in the creation, delivery and consumer usage of complex media content?

Well, Sony is driving an effort to examine whether SMIL has a role in professional video production.  From the W3C Note on SMIL 2.0 Extension for Professional Multimedia Authoring: