W3C home > Mailing lists > Public > www-tag@w3.org > February 2005

Re: Significant W3C Confusion over Namespace Meaning and Policy

From: Patrick Stickler <patrick.stickler@nokia.com>
Date: Fri, 11 Feb 2005 11:10:31 +0200
Message-Id: <18b06107cfe1eabd788cee6983eb4137@nokia.com>
Cc: <www-tag@w3.org>, <derhoermi@gmx.net>, <paul.downey@bt.com>
To: "ext John Boyer" <JBoyer@PureEdge.com>


On Feb 10, 2005, at 19:22, ext John Boyer wrote:

> On the other hand the namespaces rec does define a namespace
> as a collection of names plus a uri.  So whether or not the
> uri changes when the list of names changes, the namespace is
> changed when the list of names changes.
>
> So, adding base or id to the xml namespace changes the namespace.
>
> Now the separate discussion arises whether this should be allowed.
>
> Thing is, the URI "identifies" the collection of names.  Since you
> can't have two different collections of names that are identical,
> changing the collection of names should imply the obligation to
> change the URI too.  The W3C pub rules have supported this
> interpretation when it says that namespaces should not be changed
> in this way in order to "give a specification the necessary stability."
>


I tend to look at all of this from a term-centric view.

Terms are identified by URIs. Certain sets of terms share a common
URI prefix, which is serialized in XML as an XML namespace (for
syntactic convenience).

If the semantics of a term changes, then the URI of the term should
change. Whether doing so correlates to a change in namespace is
a mangement issue.

A namespace URI identifies a (potentially infinite) set of terms
which all share that namespace URI as a prefix. One might say that
from that perspective, a namespace is static and immutable set. One
may not add or remove from it. Any term that has that namespace URI
as a prefix is a member of that set.

Now, *vocabularies* are not infinite, and vocabularies may specify
particular terms that are members of that vocabulary -- and those
terms may or may not all share a common namespace. Whether they do,
is a management issue.

But the key issue is that it is the semantics of the terms that
matter, and folks who are worried about changes to namespaces
and namespace URIs are missing the fact that namespaces do not
equate to vocabularies and it is vocabularies (which may include
terms from mulitple namespaces) that software cares about.

Regards,

Patrick
Received on Friday, 11 February 2005 09:12:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:32 GMT