OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] Re: URIs, concrete (was Re: [xml-dev] Un-ask the question)

[ Lists Home | Date Index | Thread Index ]

Tim Bray wrote:

> The namespace rec describes a formalism which includes an abstraction called a
> "namespace" and a condition called "being in a namespace".  It provides
> syntactic rules for determining when an element type or attribute name is in a
> namespace.  Within that formalism, it is clear that unprefixed attributes
> cannot be in any namespace.  It is the formalism on which the many
> highly-interoperable namespaces implementations are based on.

Granted, but my question  was something different.

> And yes, it is obvious that an attribute can draw the bulk of its semantic
> import from the namespace it's in, rather than the context of the containing
> element.

I didn't say the context of the containing element; I said the dependence of the
attribute upon the element. That is precisely my point. Adjectives are parts of
speech but may not of themselves form what are understood as the parts of
sentences or predications. They nevertheless have real meaning, which can be
demonstrated in their weight on the semantics elaborated from any particular
processing of the substantives upon which they depend. I am wondering if XML
attributes are not accurately, and usefully, to be understood in much the same

I think that your example below serves my point better than yours (even if we
dispense with any special pleading for the xml:lang attribute, which is not
exactly a routine example of attribute namespacing). This particular attribute in
this particular namespace is intended as, effectively, a constant:  to impart the
semantics of 'being in the vocabulary of the French language' wherever it is
used. But in the instance which you construct--and perhaps in any other
instance--it is at least as likely that the conflation of content Bonjour! with
element <anyOldElementAtAllIDontCareWhich>tests the assertion imparted by
xml:lang="fr" as it is that the use of the attribute definitively colors the
French-ness of that conflation of element and content. And if that is so, we are
back to my original question:  when the instance is processed, what semantics are
elaborated specifically because of the namespace of the attribute which color the
outcome beyond what results from the dependence of the particular attribute upon
the particular element in the particular context?

>  Consider
> <anyOldElementAtAllIDontCareWhich xml:lang="fr">Bonjour!  ...
> There's a technical term for Simon's argument in this case: "wrong".

> Unless of course you're talking about some general philosophical notion of
> namespaces, outside of the formalism in the spec, in which case I freely grant
> your point, and Simon's.

I'm not, and I wouldn't want such a concession. Simon can speak for himself. I'm
not really asking about namespaces. I'm wondering about the weight of an
attribute's namespace upon the semantics elaborated from processing that
attribute and the element upon which it depends in a given context.


Walter Perry


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS