[
Lists Home |
Date Index |
Thread Index
]
Danny Ayers wrote:
> Some cleavage :
> http://www.intertwingly.net/wiki/pie/NamespaceDiscussion
Is that cleavage a joining or a coming apart?
> Note that the namespace cleavage has only appeared in this "simple" branch.
It seems that elsewhere technical judgement was suspended insofar as
namespaces were accepted without much thought (that wiki being
another place where I'm waiting for a good argument pro namespaces).
At least I haven't seen any explicit consensus pro XML Namespaces
doe Atom. So I guess that page doesn't really have any influence on
Atom syntax. What matters are the current test feeds (which to the
man are namespaced), the blogs that filter the wiki noise, if you
can call it a wiki, and the opinion of maybe 20-40 people, yourself
and a few other on this list included.
I may yet have to hack Perl...
> The RSS 1.0 branch uses namespaces extensively even for relatively simple
> feeds, using standard terms like those of Dublin Core. There it just
> works...
RSS1.0 uses XML Namespaces to tunnel URIs through XML for the
benefit of RDF. Does the XML serialization really 'just work'?
> I wouldn't characterise those as much aspects of RSS practice, rather of
> certain practitioners. I don't think the CDATA stuff is quite as bizarre as
> it seems - the motivation is to use HTML markup for content, but without
> namespaces and XHTML this leads to a bit of a mess.
It sanctions sloppy production of markup Danny. That's the real
world use case for RSS CDATA. The lack of namespaces and XHTML has
nothing to do with it. You can make the effort tidy your HTML to XML
- it's not that hard and cheapest overall when the producer does it
instead of the consumer. Worst of all CDATA tunneling plays into the
hands of legacy HTML engines with code bases dedicated to rendering
gorp no matter - gorp rendition raises the bar greatly for an RSS
client. Saying producing sloppy syntax isn't a bizarre need is no
different to saying producing sloppy semantics isn't a bizarre need.
> there has been a continuous tug-of-war between developers that want
> to do things as 'properly' as possible and those that want things to be as
> 'simple' as possible.
But we shouldn't confuse simple with simplistic.
> There probably wouldn't have been any conflict if
> 'simple' hadn't used underspecification as a tool.
+1
> The 'properly's have
> often frightened the 'simple's when talking about standards, namespaces and
> so on, and now the 'properly's are moving on with Atom. Meanwhile the
> 'simple' RSS 2.0 wagons have circled around some good stuff but also quite a
> lot of garbage.
Not really. RSS users are understanding that 'simple' should mean
simple - not technical debt, not deferred costs, not crapping on the
next guy.
Bill de hÓra
|
