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] Why associate an attribute with a namespace ?

[ Lists Home | Date Index | Thread Index ]

On Thu, 17 Feb 2005 00:26:04 +0200, mihaiu@mihaiu.name
<mihaiu@mihaiu.name> wrote:
> >On Wed, 16 Feb 2005 23:36:03 +0200, mihaiu@mihaiu.name
> ><mihaiu@mihaiu.name> wrote:
> >
> >>>The rationale is to allow attributes whose semantics are independent of the
> >>>element they appear on: examples are xml:lang, xml:base, xsi:type,
> >>>xsl:version.
> >>>
> >>       The purpose of a namespace is to make it easier to avoid name
> collisions. The
> >>special attributes that you speak about only need some kind to special syntax
> >>in order to distinguish them from the others.
> >
> >
> >Yeah, that special syntax is called namespaces...
> >
> ><snip>new form of special case syntax adding yet more semantic confusion</snip>
> >
>    I am not an adept of any new form. In order to be able to group attributes
> with special meaning you only need syntax changes and not namespaces which in
> my understanding have a "deeper" meaning.

You seem to want to invent some new special case syntax in order to 
provide "special meaning".  That is, precisely what namespaces already
do.  There's no "deeper" meaning then what any specific name space
aware application wants to attach any random element it might

> >>       I hope that you understand me: there is no need for namespaces in this
> case
> >>because namespaces are designed to solve a *completely* different set of
> >>problems.
> >
> >
> >Prey tell, exactly what would that be?
> >
>    To avoid name collision as much as possible. What do you use namespaces for,
> Peter ?

On the rare occasions I see a need for name spaces it's in order to
transparently intermingle separate sets of semantics. I'm already in a
position to manage the names I use and avoid collisions, I don't need
name spaces for that.

> >>       One more thing: you said "independent". For mode clarity I will quote
> you
> >>again:
> >>
> >>>The rationale is to allow attributes whose semantics are independent of the
> >>>element they appear on:
> >>
> >>The attributes are not independent from the element containing them; they
> apply
> >>to the containing element inclusively. For example the xml:lang attribute: if
> >>this attribute is applied to a mixed content element then you expect that the
> >>#PCDATA from the current element will respect the xml:lang attribute (not only
> >>its children).
> >
> >
> >You appear to be confusing semantics with scope.
> >
>    I may confuse a lot of things. Please explain what you mean.

In your example the xml:lang attribute might tell some application to
expect any  PCDATA within the scope of the element containing that
attribute to be in some particular form (then again they might not). 
However, the semantics of the attribute xml:lang are independent of
the semantics of the rest of the element. For example, they tell me
nothing about what a attribute named "foo" might mean.

Peter Hunsberger


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

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