WG4's handling of DR-09-0248 - General: Removing the need for qualifiers on attributes in Strict

[Francis Cave]

I had a similar discussion some weeks ago with Rex on a related topic - how
a large number of identical minor editorial corrections would be handled in
Corrigenda - and I understood from what Rex told me that there had already
been discussion with ITTF and that it was clear that the way that ITTF
expects corrections to be presented is one instruction for each change in
the text. The reason for this (as I understand it) is that there needs to be
a consistent way of presenting all corrections so that there is no ambiguity
of what text has changed and what text has not changed. Someone implementing
the standard should be able to determine whether any corrections apply to a
specific Clause by looking up that Clause number in all the Corrigenda that
are to be applied to the full text, and systematically applying any that are
found. Blanket revisions, such as have been proposed for attribute names in
examples, don't work well in this scenario, because it requires the
implementer to apply two (or more) different techniques for determining the
current state of the text at any point. If more than one technique is used,
it may be difficult or even impossible to determine exactly the current
state of the text. I don't believe that there is any difference between the
handling of changes to normative and informative text.

An XML expert may feel comfortable with an instruction "please ignore
qualifying prefixes on attribute names in examples - they shouldn't be
there, and we may eventually get around to removing them", but personally I
don't think this is at all satisfactory in a standard of any size, let alone
one the size of Part 1 of IS 29500.

Francis Cave



> -----Original Message-----
> From: Jirka Kosek [mailto:jirka at kosek.cz]
> Sent: 03 July 2009 15:01
> To: Rex Jaeschke
> Cc: e-SC34-WG4 at ecma-international.org
> Subject: Re: WG4's handling of DR-09-0248 - General: Removing the need
> for qualifiers on attributes in Strict
> 
> Rex Jaeschke wrote:
> 
> > 1. In 25 years of working on 9 standards, my experience has been that
> > most non-trivial changes/additions made at the last minute turn out
> to
> > range in quality from problematic to disastrous.
> 
> This change is in principle very trivial, it "just" affect too many
> code listings shown in examples.
> 
> > 2. A week ago in Copenhagen, we pushed back hard on this specific
> proposal.
> >>From the minutes:
> >
> > "There was broad support for adopting the proposed solution. After
> > some discussion, it was agreed that the solution involved changes to
> > narrative, examples and schemas covering at least 800 pages spread
> > through Parts 1 and 4. And qualified versions of some examples from
> > Part 1 will need to be added to Part 4. The Project Editor estimated
> > that the effort needed to implement this solution was on the order of
> > that for all the other DR resolutions combined. Given the time
> > available before the planned start of the ballots, members saw no way
> > that such a big editing task and WG4 review can be accomplished. As
> > such, resolution of this DR will be considered after the closure of
> the COR1 and AMD1 sets."
> >
> > I remain unconvinced that the task has gotten any simpler since then.
> 
> If you assume that AMD1 should literally contain all changes made in
> code listings in *informative* examples, then I agree that amount of
> editorial work is so large that it is unreasonable to do it in AMD1.
> But I'm not convinced that amendment should literally contain changes
> in those code listing, especially because of the sheer volume of
> changes.
> 
> > 6. Finally, we have some non-trivial implementations out there living
> > just fine with this "wart". Does it really need "fixing"? It's mighty
> > tempting during standards making to "just do it right" and clean up
> > certain unpleasant artifacts. However, to a very large extent, we are
> > consolidating prior art here, we're not crafting a new spec from
> scratch.
> 
> This is true for Transitional. But as namespace for Strict was changed
> and compatibility with existing applications was already broken it is
> chance to do little bit more cleanup.
> 
> If we want to have cleaner Strict and use unqualified attributes then
> we should make this change now or never. Postponing the change to a
> next set of CORs/AMDs would mean that Strict is not stable but
> radically changing beastie -- something what people are not expecting
> from IS and our WG.
> 
> It would be interesting to know what others think about this issue and
> sufficiency of editing instructions I have proposed.
> 
> 			Jirka
> 
> 
> --
> ------------------------------------------------------------------
>   Jirka Kosek      e-mail: jirka at kosek.cz      http://xmlguru.cz
> ------------------------------------------------------------------
>        Professional XML consulting and training services
>   DocBook customization, custom XSLT/XSL-FO document processing
> ------------------------------------------------------------------
>  OASIS DocBook TC member, W3C Invited Expert, ISO JTC1/SC34 member
> ------------------------------------------------------------------