[ubl-comment] Some comments on Interoperability

["Miller, Robert (GXS)" <Robert.Miller@gxs.ge.com>]

Title: Some comments on Interoperability

Hi All,

Not sure if these comments are in your court on not, but they are important.

One of the requirements set forth at the start of the ebXML effort was interoperability with non-ebXML-based business message protocols.  Think for example, the RosettaNet Pips.

I've written about the semantic equivalence of element instances such as:
  <Amount code='TL' currency='US'>
  <TotalUSDollarAmount>

I have further suggested that these two representations could be identified as equivalent if both referenced the same UID.  Yet I don't really think the two should have the same UID, since the latter definition is actually a subset of the former.  Carried to the extreme, there might be but one UID, that of 'string'! 

Given an existing business schema (EBS), interoperability would be aided if the elements comprising the EBS could be related through metadata developed post-EBS development and referenced trough the schema via UID.  But that's the very issue I raise when I speak of semantic equivalence in my response to the Methodology paper.

I understand of course that 'exact' equivalence' is a rare luxury.  More common is 'near-equivalence'.
I also realize that heirarchical structure gets in the way of equivalence. That's one reason I like to start with a relational view.

Bottom line is a need to be able to specify how to locate metadata from a UID; how to discern relationships from the metadata, and so how to be able to navigate among UIDs through a directory.  Of course, unless the metadata we would need has been captured in the directory, we won't have realized much.  Then again, our job isn't to capture metadata, it is simply to provide a structure others might use to capture metadata.  To date, we haven't provided much of that structure. 

We need to think into the future far enough that what we have provided doesn't get in our way in the future.  That means my words to the effect that we use the same UID on an 'equivalent' element are at least suspect, and more likely just dead wrong. It may also mean that UID doesn't quite mean what we say it means. We also need to provide some guidance on how to extend what we have provided to better address the needs of the business community to communicate business messages effectively.

Cheers,
       Bob Miller

Cheers,
        Bob