RE: Infoset consistency between versions

>So my question is whether there is any W3C guideline that says 
>"revisions that have cascading impact should be a major version 
>not a minor version"?

Any specification that is normatively dependent on "X", should refer to
the specific major/minor release of "X" that it is dependent on.  If a
new major or minor release of "X" becomes available then it is up to the
maintainers of the specifications that are normatively dependent on "X"
to determine if they must revise their specification to take into
consideration the changes in the Infoset introduced by the revision of
"X".

I am not sure we need an administrative rule to force a major version
changed when there is a "cascading impact" since the determination of
the "cascading impact" in most cases can only be determined by the
authors of the dependent specifications.  For example, the authors of
"X" could decide to delete a specific part of the Infoset it describes
and most people would assume this would have a "cascading impact" but
there might be a specification that did NOT depend on that part of the
Infoset.

But your question is a good one especially given the increasing number
of W3C specifications that have been a Recommendation for sometime and
need to be revised.  

BTW, another interesting question is when should a revised specification
introduce a new namespace?  From my view if a specification is being
revised and its namespace does not change, then I would assume that
there or no "cascading impacts" in the revision and the only changes
being included are editorial errata.

/paulc  

Paul Cotton, Microsoft Canada 
17 Eleanor Drive, Nepean, Ontario K2E 6A3 
Tel: (613) 225-5445 Fax: (425) 936-7329 
<mailto:pcotton@microsoft.com> 


> -----Original Message-----
> From: Rick Jelliffe [mailto:ricko@topologi.com]
> Sent: Tuesday, July 09, 2002 7:13 AM
> To: www-tag@w3.org
> Subject: Infoset consistency between versions
> 
> 
> This is question that probably can be answered in this forum without
> needing any TAG discussion: it may be that there are some W3C
guidelines
> in place, in which case I apologise for my ignorance.
> 
> Some W3C recommendation, let us call it "X", has an associated
> infoset, and "X" and its infoset are used as the building blocks for
many
> subsequent specifications.  (I guess X could be HTML, XML, XPath
> or WXS at the moment, but of course I am primarily thinking of XML
1.0.)
> 
> Let us make a distinction between classes of infoset items and the
rules
> for parsing and deriving an infoset.  It seems to me that the rules
> for parsing and deriving an infoset are nicely layered: they can
> be changed without really affecting other subsequent layers.
> 
> But adding (or subtracting) classes of infoset items seems a different
> kettle of fish. The new class of info item will require a cascaded
> version up throughout the specs that use it.
> 
> So my question is whether there is any W3C guideline that says
> "revisions that have cascading impact should be a major version
> not a minor version"?
> 
> Would that be useful?
> 
> Cheers
> Rick Jelliffe

Received on Tuesday, 9 July 2002 08:33:23 UTC