Objection to resolution of LC comment [was: Answer to Paul Grosso about Classes of Products]

Karl, QA WG,

The XML Core WG discussed this once more today.  Other 
WG members share some of my concerns and confusion, but 
the WG decided not to make a WG statement.  Therefore,
this statement is coming just from me (though other
WG members may also make follow up comments on their on).
Please consider this email to be my rejection of your
response [1] to my comment [2] on the QA Framework Last Call.

As far as your statement that you are happy with what xml:id
says now, that doesn't really help.  We put in some wording,
but many of us either don't understand what it really says, 
why it satisfies you, or how it is really helpful to anyone.  
So despite the fact that xml:id may be in the clear, we still 
don't really understand how to satisfy what the QA framework 
document requires us to do in the general case.

The XML Core WG writes specs that define the semantics of XML
language constructs and/or the virtual information elements
represented by a parsed XML document.  That is, we are defining
something that is somewhat theoretical, we are not standardizing
the behavior of a UA.  As such, I don't see how it is either
possible or fruitful to provide a list of classes of products
that are supposed to conform to the specs.  

Furthermore, many of the core XML specs are, in fact, referenced 
by other specs that may well be applicable to other classes of 
products not mentioned in the core spec.  

If we list a bunch of classes of products in a spec, we risk
not listing a product class that should in fact conform to
the spec in some way at some time.  If we play it safe and
just say that "all products that may ever have anything to
do with any aspect of XML processing", then that is useless.

Specifically, in the intro to xml:id to which you refer, all
I read that might be relevant is a reference to "conformant 
XML processors".  Is that the wording that you feel defines
the class of products?  If so, then does this say that xml:id
is relevant to XPath implementations or CSS implementations?
I suspect few people would think of calling either an XPath
implementation or a CSS implementation an "XML processor",
yet xml:id is relevant to both of them.

In summary, I remain uncertain as to what you could possibly
mean by "class of product" in any useful sense, and I remain
unconvinced that it is feasible or beneficial to attempt to
include any such statement in most of the specifications of
the type developed by the XML Core WG.  Therefore, I object
to the requirements to do so imposed by the wording in the
QA Framework document.

paul

[1] http://lists.w3.org/Archives/Public/www-qa/2005May/0041
[2] http://lists.w3.org/Archives/Public/www-qa/2005Jan/0025


> -----Original Message-----
> From: Karl Dubost [mailto:karl@w3.org] 
> Sent: Wednesday, 18 May, 2005 13:19
> To: Paul Grosso; Norman Walsh
> Cc: www-qa@w3.org; public-xml-core-wg@w3.org
> Subject: Re: Answer to Paul Grosso about Classes of Products
> 
> Paul,
> 
> Le 05-05-18 à 11:57, Paul Grosso a écrit :
> > Thank you for your response.  The XML Core WG has been evaluating
> > the QA WG response for a couple weeks, and I know you have 
> personally
> > been cc-ed on some of the email, but as of this date, the 
> XML Core WG
> > still doesn't have an official WG response.
> >
> > At least some individual members are unable to accept your 
> resolution,
> > though at this time I cannot say if the XML Core WG will lodge an
> > official WG objection, but I'm sending this email today to 
> comply with
> > your deadline to provide you with the status.
> 
> 
> I have difficulty to answer. I am a bit confused because you 
> disagree  
> with something you did with success.
> 
>   I will try to explain why. The issue if I remember well has been  
> raised specifically because we said at the start that you didn't  
> comply with one of the guidelines about Classes of Products 
> for xml:id.
> I have worked closely with Norman Walsh and the document which has  
> been finally produced by the WG is conformant for this guideline.
> 
> When we filled the ICS for xml:id
> http://www.w3.org/QA/WG/2005/04/specgl-implementation- 
> report.html#xmlid_table
> 
> For the xml:id Version 1.0 Specification located at
> http://www.w3.org/TR/2005/CR-xml-id-20050208/
> 
> The resulting ICS for class of Products is:
> 
> Requirement 03: Identify who or what will implement the specification.
> Status: Yes
> Comment:  In the introduction explain that the xml:id
>                       specification gives a uniform mechanism
>                       for XML processors and XML document to
>                       create identifiers for XML Schema and DTD.
> 
> The introduction of xml:id explained your class of products.
>      http://www.w3.org/TR/2005/CR-xml-id-20050208/#intro
> 
> This was the excellent work of Norman Walsh. Kudos to him.
> 
> 
> Maybe the XML Core WG has in memory a previous discussion 
> when it had  
> difficulty to understand how to proceed, but Norman Walsh did it well.
> 
> 
> Best
> 
> -- 
> Karl Dubost - http://www.w3.org/People/karl/
> W3C Conformance Manager
> *** Be Strict To Be Cool ***
> 
> 
>

Received on Wednesday, 25 May 2005 19:23:30 UTC