review of TAG work on versioning

ACTION-241 / ISSUE-41

I've briefly reviewed many of the web pages and blog posts and
draft TAG findings on versioning. It seems like an enormous amount
of good work and careful thought has been put into the topic,
and yet we're left with a difficult and uncomfortable situation
with respect to HTML and XHTML and various extensions which
seem to revolve around issues that the current writings don't
address.

Since these are still issues that face the HTML WG today and
are of importance to the evolution of the web in the real world
and the issues about control, extensibility, role of the W3C,
ownership of namespaces, tag soup and so forth, I think it
is important for the TAG to keep trying to find some way of
approaching the topic, but in a way that is productive and
sheds light on the situation.

I'm impressed by the careful thought of the various frameworks
for thinking about extensibility "in general", but I wonder if
we might be able to make progress by focusing more precisely
on the ways in which HTML *is* being extended and examining the 
current "draft standards" in light of their applicability to
the draft findings.

Some things I haven't found readily in 
*  the use of "Modes" ("quirks mode", "near standards
mode", etc.) in the browser seems like it would have some
correlation to "versions", but the connection isn't clear.

* The ownership of the MIME type and the way in which 
the application/xhtml+xml migration might or might not
be assigned to one or another development path is unclear.

* The use of <!DOCTYPE HTML> and the relationship of DOCTYPE
to versioning isn't clear.

I think doing so would be useful to the HTML community and
focus the TAG discussion in a way that I hope can be productive.

I think this isn't exactly ISSUE-41, because we should expand
the scope to cover "good practices for extensible languages",
with specific attention to the issues around designing extensibility
to languages (and protocols!) which have widespread deployment
and massive implementation bases which are not all necessarily
compliant with previous specifications.


[1] http://lists.w3.org/Archives/Public/www-tag/2009Mar/0146.html 
[2] http://lists.w3.org/Archives/Public/www-tag/2008May/0155.html 
[3] http://www.w3.org/TR/webarch/#versioning 
[4] http://www.w3.org/QA/2007/12/version_identifiers_reconsider.html  
[5] http://lists.w3.org/Archives/Public/www-tag/2009Feb/0147.html 
[6] http://www.w3.org/2001/tag/doc/versioning-compatibility-strategies 
[7] http://www.w3.org/html/wg/tracker/actions/108

Received on Tuesday, 7 April 2009 23:24:15 UTC