Re: Comments on ebXML Core Components forms

["Martin Bryan" <mtbryan@sgml.u-net.com>]

Stuary/Bob
>
> I'll give a 'maybe' to synonyms - 1) en:name and fr:name are not synonyms.

Our model has always allowed Names to have an xml:lang attribute to allow
for different language versions. Its just that the data capture forms only
deal with one language, and we were talking on Wednesday about what would
happen when we came to integrate all the different versions.

> 2) Many (most? all?) synonyms turn out to be 'almost synonyms'.
> It is perhaps better to treat all 'synonyms' as 'almost synonyms' and
define
> them independently as subclasses of the same parent class.
> Of course, I wouldn't have any 'almost synonyms' in the Core Components,
as
> these are the definitions upon which the 'off-by' are founded.
> ***Yes, I agree.  I find synonyms invariably equally misleading.  Infact
the
> point I was making to martin was that do not set in stone the fact that
only
> names and descriptions can be associated with different
languages...others,
> such as synonyms, could be as well

The problem is that different industry sectors and/or business processes may
need to assign locally meaningful variant names (note I did not call them
synonyms - they could be names that identify subclasses of the main name, or
that impose some constraint on the set of permitted representations of the
name e.g. EAN-Identifier or DUNS-Number in place of PartyIdentifier, as is
required by specific industries.)
>
>
> PS:Another thing I forgot to mention to Martin was to ensure that the
> language possibilities don't just follow English, French, German etc but
> also allow for Real English, Real French and the others - EN-US, EN-AU etc
> :-)   From EDIFACT we all know that shares - stock - inventory are not
> necessarily related concepts!

Don't worry, I chose xml:lang because it is a standard format taht allows me
to define things like EN-Cockney, FR-CA, ES-BS (Basque, if my memory is
correct) and even X-Cherokee.

Martin Bryan