[ubl-cmsc] Context Methodology Issues

[Matthew Gertner <matthew.gertner@schemantix.com>]

Welcome to the context methodology list. I'd like to kick off discussion by
establishing a list of issues that need resolution in order to finalize the
specification. I'd like to stick to gathering issues for a few days before
starting discussion on these. So unless someone objects, I would suggest
that we reserve this thread for suggestions of new issues, rather than
discussion on the issues themselves. I'll split all of the retained issues
into separate threads to kickoff discussion towards the end of next week.

1) Do we consider the "Document Assembly and Context Rules" document from
Vienna (v1.04) to be the starting point for our work, which we will be
refining? Or so we want to start from a clean slate and pull in input from
the document as needed? The question here is whether people consider the
Vienna work to be deeply broken in some way or whether it simply needs to be
finalized.

2) Should XSD derivation be used when context rules are applied: always,
sometimes or never? In others, what is the derivation relationship, if any,
between contextual components and the base components that yielded them?

3) Are we assuming an additive methodology based on minimal components, a
subtractive methodology based on maximal components or a combination of
both?

4) Should the transformation be represented as an adhoc XML document (like
the Vienna approach) or using some standard transformation language (such as
XSLT)?

5) How do we determine how hierarchical context driver values are matched?
For example, when does "France" match "Europe" (since Europe contains
France) and when does it match only "France" exactly? The Vienna document
uses the "Apply" attribute for this purpose.

6) Do we still need assembly rules? There has been discussion that the
construction of larger BIEs from core components using standard schema
languages eliminates the need for a separate assembly language.

7) Syntax of the context rules in general. Assuming that we decide to use a
canonical syntax language (probably XSD), should we align the names of the
context tags so they match the tags in the schema language? Can we handle
all of the various constructs (global types, local types, local elements,
etc.) that are retained by the Naming and Design subcommittee? Are any of
the constructs redundant (e.g. what is the different between "CreateElement"
and an "Add" tag containing and "Element" tag)?

This is all the occurred to me, but I'm sure that discussion will lead to
many other points. If anyone has anything to add, please post to the list
under this thread.

Cheers,
Matt