Re: References to SMIL 2.1 in SVG 1.2

On Thu, 2005-02-10 at 19:26 +0100, Bjoern Hoehrmann wrote:
> Dear Scalable Vector Graphics Working Group,
> Dear Synchronized Multimedia Working Group,
> Dear W3C Communications Team,
> 
>   I've already expressed my opposition [1] against the versioning
> requirements imposed by W3C's publication rules [2] and on how W3C
> manages the Technical Report URL space [3] and it seems SMIL 2.1
> provides an opportunity to repeat parts of the argument. E.g.,
> 
>   http://www.w3.org/TR/2004/WD-SVG12-20041027/animation.html
> 
> refers to SMIL 2.0 e.g. using
> 
>   http://www.w3.org/TR/smil20/smil-transitions.html

That's the mistake, if there is one: the spec should refer to a
dated (i.e., stable) version of another spec.

> which now redirects to
> 
>   http://www.w3.org/TR/SMIL2/smil-transitions.html
> 
> which is not the SMIL 2.0 Recommendation but rather the SMIL 2.1
> Working Draft. Obviously such a policy change comes unexpected
> and introduces all sorts of confusion. In particular because
> SMIL 2.1 states
> 
>   If this specification is approved as a W3C Recommendation, it
>   will supersede the 07 January 2005 version of the SMIL 2.0
>   Recommendation (Second Edition) [SMIL20].
> 
> The term "supersede" is not defined in the Process document and
> there does not seem to be a reasonable interpretation for it...

I'm not sure we have to define supersede.

> The Process document should be updated to include a definition
> for this term if W3C intends to continue using it outside of the
> well-understood SotD note.
> 
> There is further no reasonable way to replace the links to SMIL
> 2.0 in the SVG 1.2 Working Draft, it would have to use the "dated"
> URLs which refer to SMIL 2.0 Second Edition. That's not a very
> useful reference should there ever be a SMIL 2.0 Third Edition.

There shouldn't be one. SMIL 2.0 2E is the end of the line for 2.0.
SMIL 2.1 takes over from here. That was what the SYMM WG committed to.

> The idea behind linking to the "latest" version is exactly that
> there is no confusion about the status of the reference if the
> referenced document gets updated and to aid readers who prefer
> to read the latest normative text than old, incorrect text plus
> the errata (should they remember to look into it).

"Latest version" is an ambiguous phrase when there are multiple versions
of a technology at various maturity levels all co-existing. We are
working to disambiguate the phrase.

I do not think the proposed semantics (two different "latest thing" 
references) is the problem, if there is one. If there is a problem
as you have described, it is that a specification refers to a moving
target, which is almost always problematic.

 _ Ian

> I thus request, again, that the publication rules are changed
> such that they do not contradict W3C's publication practise and
> that (consequently, as that would mean that revised editions of
> a specification do not get new version numbers) each Technical
> Report's latest edition be made referencable through a URL, for
> example, SMIL 2.1 should be at http://www.w3.org/TR/smil21/ not
> <http://www.w3.org/TR/SMIL2/>.
> 
> I would further like to request, though that's of much less con-
> cern to me, that no "latest foo version" URLs are published, I
> think http://www.w3.org/TR/soap/ and http://www.w3.org/TR/html/
> have already caused sufficient confusion and there is very little
> use to link to such documents, in particular, if there is no
> consistent maintenance for such documents, the "soap" and "html"
> documents here are quite different indeed.
> 
> The SVG 1.2 Working Draft should be updated to refer to the right
> version of the document (one that actually includes the text that
> is referred to and/or corresponds to the version the document
> cites; ideally both, of course, but I'll address "changes-only
> documents" separately).
> 
>   [1] http://lists.w3.org/Archives/Member/w3c-archive/2004Aug/thread.html#12
>   [2] http://www.w3.org/2004/02/02-pubrules.html#head
>   [3] http://lists.w3.org/Archives/Public/site-comments/2004Mar/thread.html#21
> 
> Thanks,
-- 
Ian Jacobs (ij@w3.org)   http://www.w3.org/People/Jacobs
Tel:                     +1 718 260-9447