RE: Comments on Section 2.6 of XML-Namespaces

Andrew n marshall (amarshal@usc.edu)
Wed, 1 Apr 1998 00:21:53 -0800

On Tuesday, March 31, 1998 7:26 PM, Andrew Layman
[SMTP:andrewl@microsoft.com] wrote:
> Thanks for your question, and for reading the document closely.
>
> I think you are expecting namespaces to do more than it in fact does.
> Namespaces simply allows you to distinguish the T.Heat:Temp attribute
from
> the T.Color.Temp attribute. It does not take on the job of expressing
> grammatical rules such as that the T.Heat.Temp attribute must only be
used
> in conjunction with another attribute, e.g. T.Heat.Units. Namespaces are
> just a named set of names. Keeping the names distinct is the goal of
> namespaces. Defining the grammar, semantics, etc. of the named things is
> beyond the pale.

But aren't attribute names already unique? From the XML spec [3.3], a DTD
may contain "at most one attribute definition for a given attribute name
[in an attribute-list declaration]". As long as the element names unique,
which it seems you have done successfully, then there is no need for your
syntax. I believe the syntax you propose introduces more problems than it
is worth for the previously given reasons, even if it does not intend to
describe semantics.

I will make one exception, which it seems you are already aware of:
processor specific information that may apply to every element and
therefore does not specify a particular element as the source of the
attribute. A good example of this usage can be found in the new XLink
notation 'xml:link'. Another possible use may be in XML architectures, as
in 'Arc:URL' ( realize this isn't the currently recommended form, but I
think it makes sense).

I guess my complaint boils down to allowing QNames of the form
'namespace:ELEMENT.attribute' as the name of an Attribute. While I don't
see any reason for this notation at all, it should at least include a
validation constraint to check that the STag uses the QName
'namespace.ELEMENT'.

Andrew n marshall
student - artist - programmer
http://www.media-electronica.com/anm-bin/anm
"Everyone a mentor, Everyone a pupil"