Re: SBS and input specs

["Stephen Green" <stephen_green@bristol-city.gov.uk>]

Micah

Greetings

Sorry to have to say, we don't quite have it yet but we have been
considering a suitable format (e.g. designing some kind of subset 
definition language in XML). All we have at present are the XPaths
published to this list. I did produce Schemas in order for Ken to 
generate the XPaths but I'm wary of distributing these as they
have the same namespaces as the full UBL Schemas and that might
cause confusion.

Can you continue without them, just with the text file XPaths say
(this file should be machine readable to some extent)? If you would
help us (especially Ken if he has any time) architect a suitable subset
definition language Schema, great. I had it in mind as a side-product of
the soon-to-be-formed Small Business SC but I see no harm in doing
it in HISC if the need is urgent. Otherwise us doing it in SBSC, perhaps 
as the first work item might be nice :-)

I can't help wishing for an extension to XSD to allow folk to produce
Schemas with some flag to say they are subsets and therefore not to
be confused with the full Schemas even though they might have the 
same namespaces but I guess that is just dreaming and certainly
not what I think you are asking for here (it would be more of use for
secondary validation in an e-trading production system perhaps).

Your example is pretty much what I was thinking we'd need but might
I suggest we add attributes for cardinality and maybe others
for the further metadata Ken has in the existing XPath files. Otherwise
I'd very much agree with this being an appropriate format. Really it
might then be up to Ken to adapt his XPath generator to output in this
format too and then run it for SBS 0.5. Perhaps that is again a reason
to do it under the SBSC umbrella but I don't suppose it matters very much.

Perhaps a 'container' would be a good addition like 
<XPathList><XPath>...</XPath><XPath>...</XPath>...</XPathList>

Thinking of schedule, didn't I gather correctly from earlier minutes that
the HISC plan is to wait for a UBL 1.1 SBS? This would likely be the same
or almost the same as a UBL 1.0 SBS, mind, but the timeframe that introduces
would allow time to introduce the subset DL format. However, I could
appreciate the need to use something in the meantime during this development
phase and the desirability of not having to change it too much later.

What could we call it? SubsetDL (Subset Definition Language)? (confusing
with ...Schema Definition Language which is also ...SDL might be an unwanted
side effect. How about UBL SDL (UBL Subset Definition Language)? There's still
the problem of folk mistakenly thinking it might stand for UBL Schema Definiition
Language. How about UBL SubsetDL?

Anyway, glad this has been brought up.

Many thanks

All the best

Steve




>>> Micah Dubinko <micah@dubinko.info> 22/02/05 02:52:27 >>>
Stephen,

Where can I find machine-readable (and transformable!) specs for SBS? I 
have a package here that includes samples (.xml) and html versions of 
the spreadsheet.

Ideally, I'd like to see something like: (glossing over ns declarations, 
but you get the idea)

<sbs for="Order" version="1.0">
  <xpath>order:Order/cbc:IssueDate</xpath>
  <xpath>order:Order/cbc:Note</xpath>
  ...
</sbs>

Are those xml sample files inclusive? (Do they include every possible 
element in SBS?)

If I could figure out how to get something like this, I could make some 
serious progress.

Thanks,

.micah

-- 
  Available for consulting. XForms, web forms, information overload.
  Micah Dubinko                           mailto:micah@dubinko.info 
  Brain Attic, L.L.C.                        http://brainattic.info 
  Yahoo IM: mdubinko
  Learn XForms today: http://xformsinstitute.com