W3C home > Mailing lists > Public > www-math@w3.org > August 1999

Re: MathML 1999-07-07 Errata

From: David Carlisle <davidc@nag.co.uk>
Date: Sat, 7 Aug 1999 02:07:41 +0100 (BST)
Message-Id: <199908070107.CAA02178@nag.co.uk>
To: altheim@eng.sun.com
CC: www-math@w3.org

I did worry about using SYSTEM, but 
but it makes it a bit simpler to set up for the average user (no catalog
files  to set up) and more importantly the entity files  don't match
the original iso setsn
>      "ISO 8879:1986//ENTITIES Added Math Symbols: Arrow Relations"
as far as I know that fpi matches the original SDATA entities doesn't it?
but for XML need mappings to unicode, and since unicode doesn't
(currently) have the relevant characters most of the mapping ends up in
the private use area, also, in the case of isoamsa and some of the
others the basis was the extended set from the  iso report 9573-13

As you say we could  define fpi's for the mathml entity files but 
still at the current time at least I think the end result would be the
same, that a copy of these particular files would still need to be used,
you couldn't expect to find equivalent files existing on the system
in existing SGML catalogs.

we have a face to face meeting  in a couple of days, perhaps we could
look at this again (especially for what to do if/when unicode adds the
math characters)

  <!ENTITY % ent-isoamsa SYSTEM "entities/isoamsa.ent" >

Could do that, it would be nice if we could have a param entity
controlling where they were, but I don't think the restrictive rules
in XML allow this to be easily customiable with a paramentity for the
directory path, will think about that....

David


(personal response not checked with the WG)
Received on Friday, 6 August 1999 21:10:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:48 GMT