action 209 and namespace documents for the SML namespaces

Last week I took an action (209, at the URI
http://www.w3.org/2005/06/tracker/sml/actions/209) to "Investigate
source of the url pointing to sml-schema.xsd", specifically
http://www.w3.org/2008/09/sml-schema.xsd

This is to report what I've learned so far.

The file was created by manually, by the Webmaster, so we don't have a
rogue process in the editorial production system producing documents
we don't know about.  The comment on the original checkin is:

     namespaces for
     http://www.w3.org/TR/2008/WD-sml-if-20080912/
     http://www.w3.org/TR/2008/WD-sml-20080912/

so the conjecture that the document was put in place in connection
with the namespace of the last WD seems to be correct.

Further inspection shows that at the same time, the Webmaster checked
in a number of namespace documents in (X)HTML+RDDL form, each pointing
either to the schema document mentioned above or to a schema document
for SML IF:

     http://www.w3.org/2008/09/sml.xhtml
     http://www.w3.org/2008/09/sml.html
     http://www.w3.org/2008/09/sml-if.xhtml
     http://www.w3.org/2008/09/sml-if.html
     http://www.w3.org/2008/09/sml-function.xhtml
     http://www.w3.org/2008/09/sml-function.html
     http://www.w3.org/2008/09/sml-err.xhtml
     http://www.w3.org/2008/09/sml-err.html

Note that the two sml-if namespace documents and the two sml-function
namespace documents point erroneously to
http://www.w3.org/2008/12/sml-if-schema.xsd rather than to
http://www.w3.org/2008/09/sml-if-schema.xsd

Doing some background research in connection with this question, I
learn that I may have misinformed the WG on the subject of namespace
documents; if so, I apologize for screwing up.  I believe I told you
that it was good practice to provide namespace documents and that we
should definitely do so, but that it was not something that we
absolutely had to do before going to CR.  That turns out to be false.
The W3C Namespace policy at http://www.w3.org/2005/07/13-nsuri says in
section 3 that

     When a namespace URI appears in a Recommendation Track document,
     the responsible group must publish a corresponding Namespace
     Document.

This means that strictly speaking we ought to have published namespace
documents along with each of our drafts, since each draft defined a
new namespace.  And as your staff contact I screwed up by not
insisting that we do so.

What has happened is that since we didn't provide namespace documents
for our WDs, the Webmaster appears to have helped us out by providing
them for us.  If he mentioned it to anyone, it wasn't to me.  But I see
namespace documents and schema documents at (for example):

     http://www.w3.org/2008/03/sml
     http://www.w3.org/2008/03/sml-if
     http://www.w3.org/2008/03/sml-err
     http://www.w3.org/2008/03/sml-function
     http://www.w3.org/2008/03/sml-schema.xsd
     http://www.w3.org/2008/03/sml-if-schema.xsd
     http://www.w3.org/2007/08/sml
     http://www.w3.org/2007/08/smlif
     http://www.w3.org/2007/08/sml-err
     http://www.w3.org/2007/08/sml-function

And (of course) the current and final namespaces:

     http://www.w3.org/ns/sml.html
     http://www.w3.org/ns/sml-if.html
     http://www.w3.org/ns/sml-err.html
     http://www.w3.org/ns/sml-function.html

It's not entirely clear to me whether the creation of namespace
documents on behalf of WGs has become a routine part of publication of
working drafts, or whether this particular Webmaster has just been
more accommodating and energetic than usual.  Instead of bouncing back
our drafts with the observation that we hadn't provided the necessary
namespace documents, he seems to have provided them himself.  It's
very helpful, on the whole, though I sort of wish he had told us.

We may want to edit at least the older documents in place (we should
check with the webmaster to make sure this isn't a faux pas, but I
think it's not, given that they are not in /TR space) to add a comment
saying something like

    N.B. this namespace (http://www.w3.org/2008/09/sml), is specified
    by the 12 September 2008 working draft of SML (the Service Modeling
    Language).  That draft, and this namespace, have now been
    superseded by later drafts of that specification.  The namespace
    http://www.w3.org/ns/sml.html should be used in preference to
    this one.

Or for the schema documents themselves:

    <!--*
        * N.B. the target namespace of this schema document, namely
        * http://www.w3.org/2008/09/sml, is specified by the 12
        * September 2008 working draft of SML (the Service Modeling
        * Language).  That draft, and this namespace, have now been
        * superseded by later drafts of that specification.  The
        * namespace http://www.w3.org/ns/sml.html and the schema
        * document provided at http://www.w3.org/ns/sml-schema.xsd
        * should be used in preference to this namespace.
        *-->

We should also correct the schema-document pointers in the existing
namespace documents for /ns/sml and /ns/sml-if, which are bad links.

And it's my personal view that it's good practice for a namespace
document to list the names currently defined in the namespace; for XML
vocabularies, many people like just to refer to a schema document, but
for sml-function and sml-err that doesn't really help much.

I think that action 209 is discharged by this message.


-- C. M. Sperberg-McQueen, W3C XML Activity
    cmsmcq@acm.org
    http://www.w3.org/People/cmsmcq
    http://cmsmcq.com/mib/
    http://www.balisage.net/

Received on Thursday, 18 December 2008 19:08:42 UTC