Re: The "request for an element" nonsense...

On Wednesday, August 21, 2002, at 08:56 , Philip TAYLOR [PC335/O-XP] 
wrote:

> Henri Sivonen wrote:
>>
>> Being able to import one's own tags is no more useful that being able
>> to say <div class="foo"> for the purpose of publishing on the Web.
>
> I disagree, for reasons which I will explain.

Disagree about what? About adding elements or about the utility of 
private namespaces compared to the class attribute?

An XHTML 2.0 user agent doesn't know anything more about the meaning 
of <foo xmlns="http://www.iki.fi/hsivonen/FooML"> than it knows about 
the meaning of <div class="foo">. (Even if I had privately and 
unilaterally defined the meaning of "foo".)

> Whether this can be reconciled with your other requirements
> is an open question, but it is surely more productive to
> explore a genuinely /extensible/ route than to continue to
> define tag-and-attribute sets which in hindsight are
> invariably sub-optimal.

Extending XHTML to XHTML + MathML is OK, because MathML is also has a 
public spec that is produced by an organization whose specs have a 
chance to get implemented in Web browsers. However, if J. Random Web 
Author defines his own extensions, the chance that widely available 
software will be extended to support those extensions is minimal. (By 
support I mean that eg. user agents provide a meaningful default 
rendering.)

--
Henri Sivonen
hsivonen@iki.fi
http://www.hut.fi/u/hsivonen/

Received on Wednesday, 21 August 2002 16:16:58 UTC