At 02:43 PM 3/18/99 +0800, James Tauber wrote:
>>Am I being bitten by a no-entities-in-attribute-values beast?
>Actually, the no *external* entities in attribute values beast. Because your
>entity declaration is in the external subset, it isn't in the document
>entity which means it's an *external* entity.
>>If so, is there another way to achieve the same thing -- to point users of
>>my DTD to a standard set of logos without requiring that the hrefs be
>Just use elements.
Is the prohibition against external general entities in attribute values
because a non-validating parser isn't required to read the DTD? I guess
that makes sense -- sort of -- but then why allow external general entities
*anywhere*, including element content (as in your solution, which does work)?
So much of XML makes good sense to me, seems so clearly to be the product
of wisdom and years of experience dealing with text and document processing
and, latterly, the Web. When I run into some of the darker mildewed corners
of the spec -- like "external general entities are fine, just don't use
them in attributes," or like the bizarre SYSTEM attribute in notation
declarations -- well, I can't help but be surprised. The failure to makes
sense itself does not make sense.
Sigh. I know this is another of those "What does it mean, O burning bush,
when you say 'I am that I am'?" sorts of questions.
John E. Simpson | It's no disgrace t'be poor,
[log in to unmask] | but it might as well be.
| -- "Kin" Hubbard