David,
I think you're taking my comment farther than I intended. I merely meant that we have to ensure there is a mapping of the required metadata to the RIM, not that we have to define an API to extract the information out of a UML object. How the metadata gets into the registry can take a number of courses that vendors can develop. I think we need to provide the verification that the registry "can" capture the elements, not the "how".
-Paul
-----Original Message-----
From: David RR Webber - XML ebusiness [mailto:]
Sent: Thursday, June 05, 2003 8:25 PM
To: MACIAS, Paul
Cc: Chiusano Joseph; CRAWFORD, Mark;
; Diego Ballvé
Subject: RE: [regrep-cc-review] How Should UML Be Treated? (Was: Re:
[regrep-cc-review] Kickoff!)
Message text written by "MACIAS, Paul"
>If that is the proper question, then my gut reaction is to suggest that we
ensure that the registry is capable of capturing the Table 6-1 information
and leave it to the implementer/users as to whether they also include the
data in the UML object of the repository. I just think the RIM provides
more control over what goes into the registry than what goes in the
repository.
<<<<
Paul,
Precisely my point - if UML tool vendors / users want to store their UML
binaries as extrinsic objects - that's fine. We need to make sure that
the right metadata and semantics is captured to facilitate CCTS in an
ebXML Registry - regardless of whether the user has a UML tool or
not.
Republica have shown how to build such an interface using just
a Java component that is "CCTS aware".
Therefore creating the XML serialization is the right way to
go to include Table 6-1 metadata content - and then, once
you are there - then you can build out from that - as the
CRI work illustrates - to include support for the "Big 3 C's".
Thanks, DW.
