Re: Reservations about <mchar>
David Carlisle wrote:
> If however the implementation costs on mchar turn out to be too high
> I agree that we might need to reconsider this.......
Yes, it is going to be expensive (and verbose too :-). For example,
if implemented in Mozilla now, the MathML code will also be keeping
its own data structures to resolve the names, thereby duplicating what
the XML parser is already doing. Plus, there will be the overhead in
the creation and rendering of associated "frames". Since <mchar> is
an alternative to entities that are widely-accepted and supported
very well with a simpler <!DOCTYPE SYSTEM "mathml.dtd", it didn't
seem okay to me that it should have such overhead.
I should perhaps also note that with Mozilla, the list of entities
"mathml.dtd" is auto-installed (when the browser is installed) in the
bin/dtd directory (pretty-much like the bin/pluggins directory). And
since this resides in the client-side, the overhead to load it in
<!DOCTYPE SYSTEM "mathml.dtd" is not too much.
Therefore authors can copy-paste hand-written fragments that use
the MathML's REC standard entity names without much worry. I understand
other renderers have built-in knowledge of these entities (or at least,
some part that could be extended when the REC settles down). That's why
I was inclined toward not deprecating the entities, and waiting a little
bit before considering introducing <mchar>.