[ver] recommendation

[<stephen_green@seventhproject.co.uk>]

There was a versioning team meeting yesterday
with Arofan, Tony and myself attending.
Arofan intends to send me for fowarding to this
list today a written-up recommendation agreed
on the call. Unfortunately I'm having trouble 
receiving emails at the moment so I'm unsure
when I'm likely to get it - hence this message.

The recommendation will define two options:
one for a 'redeclaration' of all BIEs (as
SSC took as its default) and one for polymorphic
Schemas. I was to add a few suggestions for
possible NDR amendments for the first option
and add a few notes about the second.

In case I don't get the paper out for reveiw until
later today I'd like to give those notes about
option 2 here.

There would need, we agreed, to be one particular
NDR change for the major versions which would be
to make EVERY element global (NOT making an
exception for IDs and Codes). This would allow
the necessary functionality in an importing
minor version or customisation Schema (along
the lines we recommend for option 2).

I believe we should definitely make this change
in the next version of UBL Schemas since even
if we decide not to use the polymorphic design
for minor versions we would be allowing 
customisers to use it (and I believe this was
a requirement for the considerations of the
versioning team and an essential aspect of the
NDR).

This, I believe, would require the next version
to be a major version.

Then there would be time to consider the design
for the version after that but our recommendation
option 2 gives a workable design to allow the
polymorphic functionality for minor versions as
well as providing a reliable design along the
same lines for customised Schemas. This time
could be used testing further details such as
how multiple minor versions can be managed with
regard to imports and whether to create a new
element and type for every BIE or just for those
which are changed, then for refining the NDRules
accordingly (and possibly adding more detail to
the Customisation Methodology paper).

We agreed to this in the meeting yesterday as
our recommendation to the TC.

Many thanks

Stephen Green