WG: Syntax-Neutral Definition of Business Semantics

[Hermes Hartmut <Hartmut.Hermes@MCH11.siemens.de>]

Dear all 

Maybe this e-mail helps to clarify some misunderstandings.
Kind regards / Mit freundlichen Gruessen
Hartmut Hermes     
Siemens AG EL LP	   D-80286 Muenchen     
Tel: +49 89 9221 4564     Fax: +49 89 9221 3753 
Tel: +49 8233 600 222     Cellular phone: +49170 22 97 606  

If you want to get information on the Basic Semantic Register please visit:
http://forum.afnor.fr/afnor/WORK/AFNOR/GPN2/TC154WG1/index.htm <http://forum.afnor.fr/afnor/WORK/AFNOR/GPN2/TC154WG1/index.htm>  



-----Ursprüngliche Nachricht-----
Von:	Alain Bezos [SMTP:alainbezos@WANADOO.FR] <mailto:[SMTP:alainbezos@WANADOO.FR]> 
Gesendet am:	Montag, 13. März 2000 09:39
An:	ISSS-WS-EC-DATATYPING@LISTSERV.CENORM.BE <mailto:ISSS-WS-EC-DATATYPING@LISTSERV.CENORM.BE> 
Betreff:	Re: Syntax-Neutral Definition of Business Semantics

Martin,

I am afraid you missed my points.
I fully concur with the idea of a clear separation between: 1/ the definition of the semantics of information units in multiple languages, 2/ the definition of information sets, 3/ the definition of messages, and 4/ the format used for the transaction.
I fully suppport the idea that XML (due to its popularity) is one of the formats that can be used.
The definition of the semantics of information units, sets and messages should be independant of the format used for the transaction.  Using XML doesn't help for the definition of semantics of units. (It is mainly a dictionnary exercise). In this specific field, the work done in TC154/WG1 (BSR) should be used. It is the purpose of BSR to define an international, standardized, multi-language dictionnary for Electronic Commerce.
If you feel that something is missing in the BSR, or if you miss information on BSR, please contact TC154/WG1 and provide your input to international standardization. There will be a presentation of BSR in Paris during the EWG on Wednesday 22 March.
In your paper you use a "basic language" to define sets and messages from units, made of "," "+" "/" ... This is a very weak approach to define a complex structuration of information. This was one of the difficulties in EDIFACT, that needs to be solved. (CEFACT recommends to use UML to solve it.) This is where I propose to use a standardized language, EXPRESS.
Then you translate this language into DTDs. As I said TC184/SC4/WG11 is currently working on transaltion from EXPRESS language into DTDs.
The use of a real language for the definition of messages would not prevent the use of XML, as an implementation form. (by the way there is a query language associated to EXPRESS language: SDAI which is aslo an international standard ISO 10303-22). At the same time it would not impose the use of XML either. (EDIFACT syntax or STEP neutral file syntax could also be used depending on the context of use).
I hope this clarifies my position.
Best regards
Alain Bezos

----- Original Message -----
From:	Martin Bryan <mtbryan@sgml.u-net.com <mailto:mtbryan@sgml.u-net.com> >
To:	Alain Bezos <alainbezos@wanadoo.fr <mailto:alainbezos@wanadoo.fr> >
Cc:	<>
Sent:	Friday, March 10, 2000 8:35 PM
Subject:	Re: Syntax-Neutral Definition of Business Semantics


> Alain
>
	> > As far as the structuration of messages is concerned, you propose sort
of a
	> > "modelling language" to represent the combination of information units
into
	> > information sets into information messages.
>
> I prefer to see my approach as a separation of what is currently a mixed
approach to message semantics in both the EDI and XML worlds into clearly separated approaches for defining the semantics of information units in multiple languages separately from the definition of information sets suitable for multiple domains and the definition of information messages for specific business processes. To the best of my knowledge BSR only covers the first two of these aspects, and ignores the question of the controlled use of semantics, whereas EDI has traditionally mixed all three up in.
>
	> >I would strongly recommend to
	> > use an existing standardized modelling language : EXPRESS (ISO 10303-11)
,
	> > which was specifically developped to describe such structures (in the
field
	> > of industrial data, but looking at your document it makes no
difference).
>
> The reason I did not use EXPRESS is that there is no way I know in which
an XML system can query an EXPRESS model, but there is an in-built mechanism for querying XML descriptions. Hence I deliberately chose to model my data using an XML model as the starting point. In doing this I was seeking a way of interchanging queryable semantic definitions that would work across domains.
>
	> > As far as the semantics is concerned, the main difficulty is not in the
	> > "attributes" defined by 11179, but in the definition of what you call
the
	> > meaning. The BSR should be refered and used to define the information
units.
	> > Its structures is currently based on ISO/IEC 11179.
>
> I stated quite clearly in the document that the examples were not based on
"real semantics" such as those in the BSR at present. Unfortunately the BSR tables are not presented in a format that is based on ISO/IEC 11179 semantics, so it is impossible for an "outsider" such as myself to understand the mapping between BSR and 11179.
>
	> > Not being a specialist in SGML/XML, I will not comment the part of the
	> > document dealing with the XML representation of information sets. What I
can
	> > say, is that the use of EXPRESS language to describe information sets
and
	> > messages, would allow to use the existing work on XML representation of
	> > EXPRESS defined data (ISO 10303-28) to automate thi process.
>
> Not in an easily queryable way.
>
	> > To conclude, I feel that the goal of the project is not to define new
	> > "standards", but to provide users with guidelines to consistently
implement
	> > and use existing international standards.
>
> Please show us how you can represent your EXPRESS model of BSR in a manner
that can be interrogated using an XML Path statement attached to a URL. If and when I can understand how to do this it might become feasible to use your approach. At present I do not understand how this could be achieved, let alone enough to document it. I can only write up what I understand. You are, off course, welcome to submit an alternative approach to DAMSAD.
>
> Martin Bryan
>