Re: Consensus on alternate prefixing mechanism

Mark Birbeck wrote:
>> It's because of this point that I think that "DEFAULTNS" would be
>> preferable to a blank prefix identifier for changing the default namespace.
>>
>> I believe that we end up with a non-deterministic case if we allow
>> arbitrary spaces between mappings and the equal sign AND try and use a
>> blank prefix identifier to change the default namespace.

Just to clarify - I was responding to Ivan's point and using the same
terminology as to not confuse the issue further. I still believe that
Ivan's point (how to set the default prefix) is a somewhat separate
discussion from the point that you and Shane are making (the default
namespace is not to be confused with the default prefix).

I did make a mistake in calling it a namespace on the wiki page and in
hindsight I probably should have straightened this out sooner on the
mailing list (which, thankfully, Shane did).

> The point that both Shane and I were trying to stress is that these
> are *not* namespaces.
> 
> Even when using @xmlns they are not namespaces. :)

I agree that prefixes are not namespaces per the definition of
'namespace' in the XML sense, but I have a feeling that this community's
decision to use xmlns is going to haunt us for a very long time.
Specifically because we are ambiguous in the spec on the link between
xmlns and prefix mappings.

Here's our argument:

RDFa uses prefix mappings to express bindings for CURIE prefixes. These
are not namespaces as defined by the XML specification.

and then we go on to state:

XHTML1.1+RDFa uses xmlns to express these prefix mappings.

The above statement is where we screwed up - we failed to see that
people would assume that anything placed in xmlns would not actually
have anything to do with XML namespaces. We abused the tag.

Furthermore, the CURIE Syntax Specification even states[1]:

"""
When CURIES are used in an XML-based host language, and that host
language supports XML Namespaces, prefix values MUST be able to be
defined using the 'xmlns:' syntax specified in [XMLNAMES].
"""

We're certainly setting up an environment where one might assume that
there is a stronger link between xmlns and CURIE prefixes than what we
intended.

> They are prefix mappings, that just happen to be resident in the
> @xmlns attribute.

...and people will just happen to assume a stronger link between xmlns
and prefixes since xmlns is used to express this relationship.

> But if we use @prefix (or @token, as I hope), then they lose even that
> superficial relationship to namespaces.
> 
> So we should drop that term completely from our discussions.

There is much more than a "superficial" relationship to namespaces that
"x:y" produces. We've re-hashed this discussion numerous times on the
Microformats mailing list (but from the opposite direction), prompting
an entire wiki page on the subject:

http://microformats.org/wiki/namespaces-inconsistency-issue

Personally, and I believe that I've made this clear during the telecons,
is that we should run away from xmlns for defining prefix mappings as
quickly as possible.

We will not, however, be able to drop the term completely from our
discussions. "x:y" does have a good bit to do with namespaces, in the
traditional Computer Science sense of the word. Yes, it does not have
anything to do with namespaces as is defined in [XMLNAMES], but it
certainly does have something to do with the general concept of
namespacing, which is described in [2].

Again, I agree with the concept that prefix declarations are not the
same as namespace declarations and prefix mappings are not the same as
XML namespaces. Prefix mappings do, however, have a relationship to the
traditional Computer Science term "namespaces" - and I do think it is
going to be very hard to teach the nuances to the web at large.

So, I don't think we'll be able to drop the term completely from our
discussions, and even if we could, I don't think that it would be the
right way to go forward.

-- manu

[1] http://www.w3.org/TR/curie/#s_syntax
[2] http://microformats.org/wiki/namespaces-inconsistency-issue
[XMLNAMES] http://www.w3.org/TR/1999/REC-xml-names-19990114

-- 
Manu Sporny
President/CEO - Digital Bazaar, Inc.
blog: A Collaborative Distribution Model for Music
http://blog.digitalbazaar.com/2009/04/04/collaborative-music-model/

Received on Saturday, 2 May 2009 16:14:53 UTC