- From: David Carlisle <davidc@nag.co.uk>
- Date: Fri, 24 Oct 2003 17:31:52 +0100
- To: david@allouche.net
- Cc: www-math@w3.org, vdhoeven@texmacs.org
Yes this is one of the "problem" cases where we can't win. It is flagged in the documents in http://www.w3.org/2003/entities as a problem case. It is clear that the "right" thing to do for Unicode 3.x and above is to use the new characters (this is also the right thing to do for W3C full unicode normalisation as well) However it is very hard (virtually impossible) to change the official HTML entites as html browsers don't really read the dtd they just have the definitions hard wired in the code, and that's an awful lot of applications running out there with those definitions. Currently (at least) xhtml wants to stay compatible with html... If mathml isn't compatible with xhtml then mathml fragments change meaning if they move from a mathml document to a mathml+xhtml one which is bad, so in (almost) all cases the mathml dtd uses html compatibility as an overriding rule. I suspect that the entities in http://www.w3.org/2003/entities which are a draft update to ISO/IEC 9573 won't take html compatibility as a first priority (although of course it will still be an important consideration) If an official ISO mapping can be agreed then perhaps HTML 2 could use that, and we could revision mathml to match. However I have no idea of the relative timescales involved here and so for MathML2 rec today. We have the definitions as given in the just published rec chapter 6. David -- http://www.dcarlisle.demon.co.uk/matthew ________________________________________________________________________ This e-mail has been scanned for all viruses by Star Internet. The service is powered by MessageLabs. For more information on a proactive anti-virus service working around the clock, around the globe, visit: http://www.star.net.uk ________________________________________________________________________
Received on Friday, 24 October 2003 12:36:36 UTC