- From: Roger B. Sidje <rbs@maths.uq.edu.au>
- Date: Sat, 29 Apr 2000 12:00:56 +1000
- To: www-math@w3.org
I have a number reservations about <mchar>, and I feel its introduction is premature. I understand the motivations in introducing it, but from an implementor perspective, I see that at this stage, the disadvantages of <mchar> outweigh its benefits. I think it is best to wait for a general solution from the XML language itself. Perhaps XML should support a different class of inline-replaced "&&Entity;" that are not reported as an error if they are unresolved, I don't know. That's a question for the XML experts. But using <mchar> as a work-around raises other serious problems. That's why I thought I should bring these reservations to the consideration of the W3C MathML Panel. Here are some of the problems with <mchar>: * foster incompatibilities since the same char be known under different names depending on the usage and field of interest, it is hard to set a definitive list. (cf. the listing at http://www.mozilla.org/projects/mathml/fonts/chars/table.html which is best viewed with a Mozilla MathML-enabled browser). The trick of quickly aliasing in the &Entity; list provides a satisfactory solution to this problem in a standard-compliant manner. With <mchar>, there wouldn't be a standard-compliant way for the users to resort to that trick and use the most relevant names w.r.t. to their context. * poor extensibility - have to wait for next releases of renderers to see built-in support for a new name. - or have to suffer incompatibilities if each renderer provides its own way to extend its client-side list. * consume memory and slow to process - since it is a <tag>, it is part of the persistent XML/DOM tree, therefore a document with a lot of <mchar>, as it will typically be the case with scientific documents, will consume more memory, and will furthermore be considerably slower to render - gone is the speedy substitution that was done at parsing time by the (generic) XML parser on the &Entity;. - some implementors may think that, since it is a <tag>, it would have to be exposed to DOM events, meaning even more trouble given that hooks/resources will be allocated in anticipation of users' actions on the tag. --- RBS
Received on Friday, 28 April 2000 22:11:09 UTC