Lists Home |
Date Index |
From: "Michael Kay" <email@example.com>
> The fact that the model is retrofitted rather than being a
> normative part of XML means that questions like "are comments
> significant" have never been satisfactorily answered.
How can that question ever have a single answer?
> the confusion over marginally-significant stuff like CDATA sections,
> namespace prefixes, and inter-element whitespace continues to cause
> interoperability nightmares. If people had defined the model before
> defining the syntax we wouldn't be in this mess.
For marked sections, ISO 8879 says that marked sections
- give the significance of a portion of data (i.e. should it be ignored,
is it temporary, should it be included) and/or
- set a recognition mode for parsing.
So a marked section was the SGML equivalent of cpp's #IFDEF.
(XML does not have IGNORE marked sections in content, which
I suppose may be one reason for the LMNL stuff, though
presumably marked sections and LMNL would have different
optimal grain sizes. )
If we ask about C "Is #IFDEF part of the C program's 'program model'"
or does it disappear, then we probably give ourselves the an answer
"well, its effects are, and probably a fancy IDE would maintain and
understand the #IFDEF structure, but the lion's share of tools shouldn't
care." Which is exactly the same as the XML answer.
Java has replaced #IFDEF but not completely satifactorily
(as anyone who has to resort to reflection to get a 1.4 program
to compile on 1.2 tools, or prepare stubs to get MRJ APIs
useful in non Mac tools can attest.) Having a dumb macro
mechanism with a different syntax to fall back on as a last resort
is a great thing. At the moment, XML has no standard way
of marking up the significance of parts, coping with variants
or change. The nearest thing seems to be html:del!