Lists Home |
Date Index |
- From: Peter Murray-Rust <firstname.lastname@example.org>
- To: email@example.com
- Date: Mon, 03 Aug 1998 23:32:24
At 13:36 03/08/98 -0700, Jon Bosak wrote:
>A major revision of the namespaces draft is now publicly available at
>This draft incorporates a new attribute-based syntax for namespace
>declarations as well as new mechanisms for defaulting and scoping.
>The XML Working Group actively solicits feedback from early
>implementors of the revised draft and has set up a special mailing
>list to gather input for an editorial team that will review early
>implementation experiences. See the section titled "Status of this
>Document" near the beginning of the WD for details.
>Chairman, W3C XML WG
Many thanks for posting this, Jon,
I have been privileged to observe process by which this spec was produced and
when it was near completion made an offer to the XML-WG that XML-DEV might
be able to assist in informally helping early implementations of the spec.
I append what I wrote (about a week ago]:
The XML-WG is releasing the next draft of XML-names and as this is of
particular interest to XML-DEV members, I'll try to set the scene. As you
will recall, XML-DEV has no formal standing in the W3C process but
unofficially its presence has been of value to the process. I suggested on
the XML-SIG list that XML-DEV could have a useful role to play in helping
the active development of namespace implementations and received broad
agreement to post this message.
[NOTE: the W3C has a number of Activities, of which XML is one (others are
DOM, RDF, P3P, MathML, etc.). Each has a chair and editorial board/team
and/or WGs which produce WDs, and PRs. The XML-WG also oversees the
XML-names development (which is central to several other activities). The
XML-WG has chosen to have a SIG of about 100 members to which many problems
are referred for discussion. The deliberations of the XML-WG and XML-SIG
are confidential to W3C members and invited experts (which includes me and
several other members of XML-DEV).]
The XML-WG, XML-SIG and other W3C groups have spent (literally) thousands
of e-mails discussing XML-names. It has turned out to be much more
challenging than was originally anticipated. I have been privileged to see
some of these discussions and I believe the people involved have shown
great commitment and technical ability.
Fairly recently it was decided to address the whole namespace problem (not
just the deliberately limited version of 1998-05). The questions of
scoping, global attributes and many other features have been very
challenging but are now included in the draft. I ask you to take on trust
that what is presented is the product of a huge amount of labour by
You may feel that it could have been done otherwise, but I suggest that we
don't pursue those thoughts but try to make the current draft work. I am
very confident that *a* good solution has been reached. With this tool we
shall be able to do some remarkable things. When/if we hit problems the WG
will pick them up, and will take a view as to whether the spec needs any
revision. There will be no need (or point) in lobbying them.
As with other XML-DEV activities we are doing this as an experiment and
because it's worth doing. It may or may not lead to protocols, software or
other resources. It will almost certainly be a good idea to start small and
expand, just as with SAX and XSchema. I'd suggest that we might:
- collate experience from those who have actually implemented namespaces
(somewhere in the 1997..1998-05 period). For example I discovered while
browsing XML4J today that it has considerable namespace support, and I'm
sure there are other tools (I'm happy to make the latest JUMBO2 available
shortly - it has a simple namespace approach).
- collate DTDs which have been developed to be namespace-aware. These will
be essential for testing systems later. Examples of (I think)
single-namespace DTDs include:
- XML-data (if simplified to primitives)
we can appeal for other examples (for example, I'd be happy to contribute
- collect documents which adhere to these DTDs.
As you can see I am assuming that DTDs (or schemas) are likely to be
important for namespaces. I think we should be wary of addressing
"well-formed tag soup" at this stage - that can come later, if at all. The
issues that we'll need to consider include:
- how can we manage prefixed and non-prefixed versions of a DTD? With and
- how can we combine more than one DTD? With and without validation.
There are also questions of the interfaces to namespace-aware:
In particular the XML and XLL specs define a number of XML:* attributes. At
what stage should these be processed? Should there be a special module for
processing xml-reserved elements (if any) and attributes? At what stage
should software need to know the identity of the prefix, the URI, etc.
Should we define a Name class/interface?
These are challenging enough. Beyond that we have the questions of scoping
- how to implement it, what its semantics are, how to treat default and
global scopes. I'd suggest we take a day or two to think about this before
commenting, and that we may be wise to take the simplest issues first. It
won't help if our discussion is so broad that we get lost.
This is an issue of great interest to W3C members and non-members alike.
Offers of help will be extremely welcome, but don't be disappointed if
other solutions or groups appear to be more appropriate for a particular
I am really looking forward to seeing this move forward. It will require
patience and forbearance. We are moving into an area where semantics and
ontology are important and this is why it is so tough, but exciting. In
essence we are developing the first global mechanism for supporting
ontology and if we can crack that we can do anything...
xml-dev: A list for W3C XML Developers. To post, mailto:firstname.lastname@example.org
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:email@example.com the following message;
To subscribe to the digests, mailto:firstname.lastname@example.org the following message;
List coordinator, Henry Rzepa (mailto:email@example.com)