RE: [office] Re: Draft I-7-13 Review: Section 18.334.19 <draw:page draw:name> Optionality

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

Patrick,

Thanks for your questions.  I have a couple of clarifications.

1. I am not saying that every attribute requires required or optional in the text, although when an attribute is optional and has a default, there must be something as already agreed by the TC.  In the specific case of <draw:page> draw:name, the text discusses the attribute being optional already, all the way back to ODF 1.0.  It wasn't my idea.

2. With the absence of schema fragments in the text, it is now more difficult to confirm whether an attribute is required or not, and the use of "may" in the boilerplate text creates confusion with normative language.  If you want to be purely descriptive about what is possible, the proper (JTC1) normative word is "can" as in "<draw:layer> elements can have attributes draw:display 18.232, draw:name 18.334 and draw:protected 18.348."

3. I agree that "must be unique within the document instance" appears to be more precise, but these are not values of type ID where the requirement that each value of any attribute of such type must be unique within the [XML] document instance has a very sharp meaning.  I am concerned over UNIQUE AMONG WHAT within the document instance.  If we mean the ODF document instance (for a document represented in a package) rather than an individual XML document instance, I agree with you that what is customary among XML folk cannot be assumed to provide an answer.  Not knowing the origin of the "within the document instance" addition, I chose to revert to the earlier phrasing in my recommended change to the text just to avoid inadvertently making a bad problem worse.  I was operating on the "do no harm" principle along "don't fix it by breaking it.

4. With regard to dropping the comment about automatic generation of the draw:name attribute and illustrating a case where an attribute is needed, I refer you to this discussion thread where this became a problem in interpreting the specification:
Hamilton, 2009-01-05: http://lists.oasis-open.org/archives/office/200901/msg00025.html
Faure, 2009-01-05: http://lists.oasis-open.org/archives/office/200901/msg00021.html
Zachman, 2008-12-13: http://lists.oasis-open.org/archives/office/200812/msg00117.html
Hamilton, 2008-12-11: http://lists.oasis-open.org/archives/office/200812/msg00102.html
and continuing back to January, 2006 as summarized in the 2008-12-11 note.

 - Dennis

-----Original Message-----
From: Patrick Durusau [mailto:patrick@durusau.net] 
http://lists.oasis-open.org/archives/office/200901/msg00032.html
Sent: Monday, January 05, 2009 16:09
To: dennis.hamilton@acm.org
Cc: Rob Weir; Michael Brauer; ODF TC List
Subject: [office] Re: Draft I-7-13 Review: Section 18.334.19 <draw:page draw:name>

Dennis,

Dennis E. Hamilton wrote:
http://lists.oasis-open.org/archives/office/200901/msg00027.html
[ ... ]
> Current Text:
> [
> The draw:name attribute specifies a name by which this element can be referenced. It is optional but if present, must be unique within the document instance. If not present, an application may generate a unique name. 
>
> attribute-draw:name_element-draw:page
>
> The draw:name attribute may be used with the following element: <draw:page> 9.2.4.
> ]
>
> Replacement Text:
> [
> The optional draw:name attribute specifies a name by which this element can be referenced.  When present, the attribute value must be unique.  Note that for the <draw:page> element to be referenced from a presentation:pages attribute, the draw:name attribute must be present.
> ]
>
>   
Are you saying that every attribute must have optional or required in 
the prose?
[ ... ]

I am not sure why you added the "referenced from a presentation:pages 
attribute" since that information is already given under the definition 
of presentation:pages, thus:

[ ... ]
>
> **NOTE 2: The ODF 1.2 draft adds "must be unique within the document instance."  This is still imprecise.  However, determining the set of values within which a particular draw:name value must be unique is a substantive matter that needs to be addressed separately.
>   
Well, it isn't any less precise than the original and I would argue is 
commonly understood among XML practitioners.

But, you are correct, what does "unique" mean for any attribute value 
that is said to be unique where a document is composed of other 
documents isn't answered by ODF 1.2 nor is it likely to be. Recall that 
we can't be more restrictive or at least that is my understanding, that 
we were previously. Sure, if we were at a major revision/non-backwards 
compatible release, there are a number of things that I would urge 
changing. The "Future Notes" in the current draft reflect some of those 
concerns.

[ ... ]