[whatwg] Table integrity and conformance

[Henri Sivonen]

On Nov 9, 2006, at 16:07, Lachlan Hunt wrote:

> Since (a & b) is equivalent to (a, b)|(b, a), aren't both of those  
> equivalent as well?

That's one of the limitations of DTDs. :-)

In RELAX NG & means a real interleave, so the above equivalence holds  
if a and b are terminals but doesn't hold if a and b are non- 
terminals that expand with + or *.

-- 
Henri Sivonen
hsivonen at iki.fi
http://hsivonen.iki.fi/