[
Lists Home |
Date Index |
Thread Index
]
"Champion, Mike" wrote:
> I'd guess this discussion has the predictable conclusion: be liberal in what
> you accept, and conservative in what you produce. Work to find and fix code
> that makes assumptions about attribute order, but don't change attribute
> order if at all possible in case a downstream application -- perhaps written
> by a non-SGML veteran to a literal reading of the original XML spec --
> cares.
Even someone who is minimally proficient in HTML knows that attribute
order means nothing in the processing of the document. If the original
XML spec literally said "attribute order is important, and must be
preserved", or if there were DTD constructs for defining ordered
attribute models, just as there are for defining element content models,
I could see somebody making the assumption that order matters, but
otherwise I don't buy this argument.
Yes, it's good to preserve order... I try to do it in the processing
code that I write, just as a favor, but it shouldn't be necessary, and
it shouldn't be taken for granted. Maybe by writing processors that
preserve attribute order, we've painted ourselves into a corner in which
the user believes that it's important, and if so, that's our own fault.
I propose that writers of XML processing tools begin to forcefully
randomize attribute order after processing documents to teach these
ungrateful users a lesson!
--
Tom Bradford - http://www.tbradford.org
Developer - Apache Xindice (formerly dbXML)
Maintainer - jEdit-Syntax Java Editing Bean
|
