Notes about the three TAG TC documents

[Stephen Green <stephen.green@documentengineeringservices.com>]

In advance of the TAG TC meeting later today I have made
some notes about the three documents - guidelines, model
and markup so I thought it might be helpful to send these
out beforehand, then to go through these points on the call.

1) Test Assertions Guidelines

a) these are fairly stable and mature

b) there is no need for a conformance clause (other than
    basic statement that the document is not normative)

c) Table added on p28 - this has a minor problem where
    there is the second part of 'Test Assertion Document'
    carried over to the second page which is confusing


2) Test Assertions Part 1 - Test Assertions Model

a) not every single attribute in the model has a definition
    in the model specification - semantics covered only in
    in the general description of the class for some attributes
    - is this OK?

b) as minuted, would we be able to ask for some editorial
    technical writer review from Dmitry in Sun?

c) could we remove the word 'class' from the formal
    definitions so that in general
    'class : xyz (...'   becomes just 'xyz (...'    ?

d) is the conformance clause OK?

3) Test Assertions Part 2 - Test Assertions Markup

a) see 2 a) and 2 b) above - same applies here

b) is the namespace OK? I reverted recently to
    '.../090930' to help an early adoption which uses this
    namespace and was not affected by any of the
    later changes to the schema. Bear in mind this is
    all still in draft mode and namespace versioning
    rules do not need to apply as yet.
    We need a namespace strategy for processing the
    markup spec and schema. I'd prefer if we kept what
    we have if appropriate, unless public review leads
    to breaking changes requiring a new namespace. If
    we do need to change it further how would it change?

c) I've a feeling we may still have an issue with the use
    of QNames in content for enumerating the prescription
    level codes. Dennis and I agreed a technique using the
    schema to constrain enumeration extensions as only
    allowing QNames with a prefix. However I have worries
    this may cause problems with anyone using XSLT on
    instances including such extensions.

d) is the conformance clause OK?


Best regards

Steve
---
Stephen D Green