The intention is to stick with the $frame stuff because it makes things so much easier for integrating with an existing codebase (both QC and engine side). The caveat is - of course - that frames in the IQM will have to match with frames in the original MDL if the IQM is replacing an original MDL. If it's a new model that's not replacing existing content you can do what you want, of course, but $frame will keep things consistent (and mean that QC folks don't have to remember if they're setting up frames for an IQM or an MDL). I like consistent (witness my rants about new text file formats that don't use COM_Parse...)
Bottom line - I honestly don't
care if an alternative may be technically superior. If it can't be integrated easily into an existing codebase, it will die. If something is to stand even a halfway decent chance of ever becoming a new standard, it must be reasonably painless to implement. And that includes for the copy 'n' paste folks as well as for more serious implementers (in fact it's the copy 'n' paste folks who do most to make something a standard by facilitating fast and widespread uptake).
So that's the intention (and the thinking behind it). How it's going to work out in practice - we'll see.
