Rick Jelliffe wrote:
> >I am maintaining the position
> >that constraining syntax is pretty well orthogonal to naming characters
> >and it is only due to the accident of DTD history that we have this idea
> >that it's inevitable that "schemas" do both.
>
> But, for example, MathML defines both elements and entities.
> Would you regard it as a special case, pioneering?
Well, MathML is clearly a special case, it seems very unlikely that any
other application domain in the history of the universe will have a
comparable requirement for the use of characters that are not on any
keyboard and do not have any input method.
Having said that, let's grant that some languages will want to have
their own names for funny characters; it is still not clear that when
you're defining a language, the mechanism you use to constrain the
logical structure is the right one to use to give names to such
characters. -Tim
