Tags and semantics ( was Dotted Names)

[martin.me.roberts@bt.com]

Folks,
	The debate about how the syntax represents the semantics is all very
well, but we end up with atleast two missing points.

	1) The core components willnot hold the complete set of semantics
sent in messages, some of this will come from the application of context
rules.  This means that any message that has had context applied will have
to be 'explained' to both the sender and the receiver in order that they
interpret it in the same manner.

	2) This is the hardest part of ebXML. We are arguing about how two
parties can interpret items such as Firstname.  As simple concept, yet in
asia Firstname means surname.  I am working with a published library (xCBL)
and I have to explain each and every element and attibute to the community
of interest in order that we are all speaking the same language.  XML syntax
seems to make the messages more accessable thatn CSV for example, as a human
can make a stab at understanding them.  It also helps in discussing message
set in that the two human developers at each end can talk directly from
examples of the real messages with out requiring a further translation step.
However,  as it stands XML does NOT of it self solve the semantic problems.
I even have my doubts about the taxonomy approach as I find that language is
not strictly based on fixed taxonomies and flexibilty of meaning is
sometimes required.

	So what do we do?  I feel that the Core Components, (as slim as they
might be) need to have their semantics honed to a precise and unambiguous
manner.  I.e that there should be a definition of what firstname means for
example.  With this we need to then start to expand the semantic definitions
beyond to include typical applications of context.  This could be done in
two ways, 1) the publication of context rules libraries with clear
definitions attached or 2) adding components to the core library in such a
way that context has been applied.  An example of this would be ,

	Core component Party would be defined
	Context component Supplier.Party would be defined as well.  This
would mean that the level of context rules would be at the margins and not
on every message.

	Sadly, I feel that some of the sniping that has gone on in this area
has belittled the difficulties that the discovery teams have encountered and
also in making some decisions in the ebXML process, the technology has been
given the front seat (because it is relatively easy to define a new
messaging protocol for example) and the semantics have been pushed back
because it is very VERY deifficult, but I feel non the less important for
that.

Martin M.E. Roberts
xml designer, BTexaCT
01473 643775
martin.me.roberts@bt.com


-----Original Message-----
From: David RR Webber [mailto:Gnosis_@compuserve.com]
Sent: 18 April 2001 03:26
To: John McClure
Cc: 'ebXML Core' (E-mail); DCN Architecture; LegalXML; Probert, Sue;
CRAWFORD, Mark
Subject: RE: Dotted-name Tags (was RE: Long Tags Codes etc. again)


Message text written by John McClure
>Wow. I'm an old REXX guy, and that language has dotted-names too. Doesn't
Python also use dotted-names? 
<<<<<<<<<<<<<<<

John,

Ooops.  Too many TREX out there!

But now I know just how old you are ; -)

REXX - huh?  Real men use CMS scripts and Xedit macros and 360's,
none of them highflutting fancy systems...

No - the TREX you need can be found at :

 http://www.thaiopensource.com/trex/

and

 http://www.oasis-open.org/committees/trex/index.shtml

 "There is a genuine need for a straightforward, easy-to-learn XML schema
language, and TREX delivers that," said James Clark, who will serve as
chair of the OASIS TREX Technical Committee. 

enjoy, DW.


------------------------------------------------------------------
To unsubscribe from this elist send a message with the single word
"unsubscribe" in the body to: ebxml-core-request@lists.ebxml.org