Let's make sure we've got nails and not screws... (Re: Let's get some principles nailed down)

[Patrick Stickler, Nokia/Finland, (+358 40) 801 9690, patrick.stickler@nokia.com]


----- Original Message ----- 
From: "ext Miles Sabin" <miles@milessabin.com>
To: "WWW-Tag" <www-tag@w3.org>
Sent: 18 November, 2002 17:24
Subject: Re: Let's get some principles nailed down


> 
> Paul Prescod wrote,
> > An XML namespace is just a set of names. According to the abstract
> > model they have no semantic significance beyond that. By analogy, a
> > function in a programming language is just a unit of code that takes
> > inputs and generates outputs. Nevertheless, one SHOULD document
> > functions in a programming language and SHOULD document one's
> > namespaces.
> 
> The documentation for a function in a programming language should tell 
> you what it does and what it's for. A namespace doesn't _do_ anything 
> and it's _for_ disambiguating names ... which is documented perfectly 
> adequately in the Namespaces REC.
> 
> I _might_ choose to impose additional semantics on my namespaces, in 
> which case I agree that I SHOULD document those semantics. But if I 
> don't impose additional semantics then there's nothing to be 
> documented.

Furthermore, as has been pointed out before, there can be different
semantics attributed to the same qname by variant XML content models. 
Just take xhtml:html for example, which has different definitions depending
on the content model, e.g. strict/transitional versus frameset.

The namespace simply partitions terms. Any semantics attributed to
those terms is not a function of the namespace, but of something
else. Any correlation between a namespace and a model is purely
coincidental, or a matter of convenience, etc. but is not inherent
in the namespace itself and a given model may encorporate terms
from any arbitrary number of namespaces and different models may
employ terms from the same namespace.

C.f. http://lists.w3.org/Archives/Public/www-tag/2002Feb/0022.html

Should the semantics attributed to a given term be consistent across
different models? I certainly think so. But it is not the namespace
that constrains any term to have consistent semantics. A qualified
name has no semantics insofar as the namespace is concerned. The
semantics comes from elsewhere. Namespaces are just punctuation.

And because a namespace does not assert any semantics, having a
"namespace document" as some kind of representation of the namespace
is simply not coherent. Because a namespace is simply a set of names,
all that a representation of a namespace could provide is a list
of those names. Since the semantics of the terms are not attributed
by the namespace they cannot be part of any representation of the
namespace and therefore the inclusion of links to schemas, etc. in
a "representation" of the namespace is simply wrong. It is external
information not inherent in the namespace. Simply calling it a
representation is not going to make it correct.

I'd like to see some RDDL examples that demonstrate taking into
account the variant definitions of xhtml:html per the strict, transitional,
and frameset models as the qname itself does not differentiate which
is indicated and thus querying a "namespace document" per the namespace
of the term is not going to tell an application the intended semantics
of the element.

Regards,

Patrick

Received on Tuesday, 19 November 2002 05:02:04 UTC