Re: [xml-dev] Pragmatic namespaces

[Michael Ludwig <mlu@as-guides.com>]

Richard Salz schrieb:
> My proposal: only allow setting the default prefix, as in
>       xmlns='....'

> The drawbacks are that you cannot mix attributes from a different
> namespace

Isn't that pragmatic to the point of being a showstopper by breaking
compatibility?

> and that if you embed clumps from xmlns A inside B, a lot, you might
> get more verbose, with the emphasis on might.

Some people, including myself, tend to do this quite a lot, and hence
feel inclined to apply s/might/will/.

In fact, from a usability point of view, I find the current xmlns/prefix
approach not too bad, especially if you declare all prefixes at the top
and don't mess around by reassigning prefix horses in mid-stream (pun
attempted and intended).

My reason for complaint about the current situation, which is apparently
shared by some other people, as users of XML, is just that the namespace
URIs are more complicated than necessary without providing any
discernible benefit as compensation.

More authoritative voices, however, presumably dwelling on their
experience as implementors of standards, have pointed out that it's
actually those innocuous-looking prefixes that cause a lot of problems
in the data model; and while not being anything of an implementor
myself, merely by reading a document where prefix horses are assigned
and reassigned in mid-stream while my eyeballs and brains have trouble
not submerging in the sea of tags, I can imagine that implementing a
high-speed boat for such treacherous waters isn't exactly kindergarten.

http://lists.xml.org/archives/xml-dev/200907/msg00055.html

http://lists.xml.org/archives/xml-dev/200907/msg00104.html

-- 
Michael Ludwig