Re: Review of XSLT 1.0 against SpecGL

At 10:14 AM 11/19/02 -0500, David Marston/Cambridge/IBM wrote:
>[...]
>LH>Do you have any comments, issues, or feedback about SpecGL 20021108,
> >as a consequence of your review?
>
>By far the most obvious thing was that issues of navigation and
>discovery were under-represented in the checkpoints. The discretionary
>items are only detectable by a thorough reading of the entire document,
>and the "Conformance" chapter says nothing about them.

I would like to make sure that this perspective gets captured in some 
issue, that QAWG will discuss and resolve for the next version.  Maybe 
Lynne (SpecGL issues master) and you can try to articulate it more 
specifically, in a way that we can discuss in email and on telecon?  (I'll 
put it in issues list if you do so.]


>LH>One thing caught my attention:
> >"Checkpoint 13.2. Distinguish normative and informative text....
>DM>Informative material is distinguished. Most examples are not marked
>DM>and can be presumed from the context to be normative."
>LH>Is that a typo, "normative"?
>
>Surprisingly, no. If you look at the examples in context, the typical
>presentation is: given this, it must do that. In other words, they are
>more like test assertions.

Just to make sure I'm not missing the point, can you point us to one 
specific such case (chapter/verse) in XSLT 1999?  Flipping around quickly, 
I see (e.g., in ch.8 "Repetition") stuff like:

"Given an XML document with the following structure [...snip a bunch of 
XML...], then a list of employees sorted by name could be generated 
using:  [...snip XSLT code...]"

In the cases you're thinking of, are the MUST's expressed anywhere else, 
outside of the example-like context?  For example, is there a prose 
description elsewhere which mandates the result of the example, given the 
particular XSLT code?

-Lofton.

Received on Tuesday, 19 November 2002 13:30:45 UTC