Re: writing entity references.

Subject: Re: writing entity references.
From: David Carlisle <davidc@xxxxxxxxx>
Date: Tue, 5 Jan 1999 11:53:32 GMT
James Clark writes

> I would say it was an abuse of XML to attach semantic significance to
> the entity names

Yes, I suspected that that would be the official answer, and really
I do agree with that point of view.

As a practical problem though, you go on to say

>  With an appropriate font, you should be able to see the
> right glyphs, so readability needn't be lost.

which isn't really true for the example that I am currently stumbling
over, namely InvisibleApply as you can't see this glyph, it is zero size
character put between the f and ( in `f(x)' to help flag this as
function application rather than multiplication, It may be treated
specially if the MathMl is being rendered to speech, or converted to a
computer algebra system.

Another practical (but hopefully short lived) problem is that Unicode
does not really have a sufficient repertoire of characters for
Mathematics. This means that almost all the MathML entities point to the
private area. However just before Christmas the `STIX' proposal was
presented to the Unicode consortium and if eventually accepted would
result in several hundred new characters being assigned `official'
unicode slots. Given this situation, the definitions of all the
character entities may well change in the future, as any officially
adopted unicode set of math characters would be assigned into a
different range, and so using _names_ such as &curvearrowleft;
produces MathML that has a better chance of being long lived than
using whatever `temporary' slot is used in the STIX proposal for this
character, given that it may be allocated an official slot in unicode at
a future date.

While this is something of a special situation, I would have thought
that if XSL is being used as a transformation language (only) to
write out XML instances, it would be quite a common request to write
out the instances in a `natural' manner (using entities defined in some
dtd) rather than necessarily writing all characters as numeric character
entities, or directly` as unicode characters.

So, perhaps, something to consider in a future draft?

David


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


Current Thread