As much as I admire Jim Hines and his
wonderful penchant for coming up with words intended to make us
feel better, I do think we should stop using SMILE. XMILE is the
standard and it encompasses both the equations and the display.
I do think it makes sense to separate out the computational
portion of the model structure from the display information. I
also think the computational portion deserves substantial
discussion before the Schema implementation is described.
Given that I would suggest the following layout (though it may
be that the standards for standards writing dictate something
else):
1. Short background as is
2. Container Schema (version codes, producer...)
3. Model equation structure
4. Model equation schema
5. Display and widget structure
6. Display and widget schema