UBL and CCTS compliance

["Michael Dill" <dill2@gefeg.com>]

Dear Stephen,

1. I do understand, that the CCTS document itself is not everywhere easy to read. But the UBL NDR, even in old versions, has a number of explanations, which are taken directly from the ISO 15000-5 CCTS and which give IMHO an answer to some of your questions. There is not a single word in them, that the BCC and ACC go into a schema. I do not feel, that this should be considered to be changed. Just BIE and documents go into the XSD.

The difference between the cc and BIE is more a conceptional one. The cc shall contribute to the interoperability and are not business context related. UBL did NOT take into account the CCTS specific way of inheritance. This effects both the way ABIE/ASBIE/BBIEs base on ACC/ASCC/BCCs and the customization on model level, but NOT the XSD derivation.

This means for a spreadsheet oriented organization like UBL should extend and improve the spreadsheets, but not the schemas. The below mentioned example PISCES shows that UBL like schemas can be derived from a data model, which uses ACCs and ABIEs.

In case, you want to see, what the UBL data models - from a modelling view - look like, then I recommend to download the free UBL Reader, which GEFEG provides http://www.gefeg.com/downloader/(d2r425egrvgo2s55zf3nbg45)/InsertForm/InsertForm.aspx
In case, you want to see, how ACC look like, then you can download the TBG17 Reader from the same website.

2. If a customization will take place on schema level (only?), then the UBL IMO risks to lose the chances for interoperability. That means, maybe it is a good idea, if UBL members could be at least discuss, whether and how CCTS views the extension and restriction of ABIE, which are based on other ABIEs or ACCs. 

3. During my participation in UBL I've listened several times to the argument, that the use of standards like UBL make software cheaper or so. Please let me say quite frankly, that the non-ISO 15000-5 CCTS compliance of UBL caused us quite a lot of additional programming work, i.e. costs. Other software vendors will have the same problem. And the non-compliance is not only caused by missing the ACC level.

4. BTW: Currently at least the following approaches divide between the master model and derived schemas. Most non legacy XML approaches support ISO 11179.

ISO 20022 Finance, maintained by SWIFT, UML with ISO 11179
CICA of X12, with CCT
UBL based on UBL 'CCTS'
EAN.UCC Methodology based on / linked with CCTS (Library uses CCTS), using UML + Conventions
CEFACT based on CEFACT CCTS (Aero, Global Transport, environment)
ISO TC 204 Transport, supported and encouraged by Homeland 
Global Automotive, currently UML, soon CEFACT CCTS
RosettaNet: says there is a data model behind the PIPs, but it seems, this is not available for users. Maybe Jon Bosak - as representative (?) on their board - knows more.
UNECE UNeDocs http://www.unece.org/cefact/forum_grps/tbg/tbg2_edocs/tbg2_edocs.htm, using UML and now as CEFACT UNeDocs changing to CEFACT CCTS
World Customs Organization, using UML, using CCT and TDED; hopefully migrating to CCTS
PISCES, based on CEFACT CCTS and using UBL schema rules

regards
Michael


-----Ursprüngliche Nachricht-----
Von: Stephen Green [mailto:stephen_green@seventhproject.co.uk]
Gesendet: Freitag, 20. Mai 2005 13:21
An: ubl@lists.oasis-open.org
Betreff: Re: [ubl] Groups - UBL V2.0 Model Architecture (UBL V2.0 Model
Architecture.doc) uploaded


Congratulations to Thomas on this excellent proposal
and potential solution to some pressing problems
(not least of which being how to design Core
Components).