Re: [regrep] Reg/Rep & UDDI (was: Describing ebXML Reg/Rep ...)

[David RR Webber - XML ebusiness <Gnosis_@compuserve.com>]

Farrukh,

I find this all very interesting here.  And last week in Santa Fe 
there were some good moves from the users to get clear
focus on needs and solutions with a brown bag lunch session
particularly to look at ebXMl / UDDI / ISO11179 roles
and model and co-existence.

I see that Matt is asking the "BetaMax / VHS" question - 
and that is of course nothing to do with technical merits.
Successful software however is not like a cassette 
cartridge, it morphs, assimulates and evolves - (compare
Java today from what it was 3 years ago).

Knowning and understanding the users is vital.
Clearly users are uncertain right now - while at the
same time users are also wanting answers near-term
to deploying real systems and solutions for their two
sets of needs.

One is shared metadata semantics of ebusiness
systems and the other is web service automation.
Right or wrong - people see that ebXML is the 
strong solution for the former, while UDDI is ahead
for the later.   And ISO11179 carries the lead for 
formalized committee driven alignment of complex
sets of metadata semantics within an organization.

So - where is the path forward?  Again the users from
Santa Fe were  focused on having to deploy and
support the right mix of technology.  Collaboration
and federation struck a strong chord as the right
way forward, allowing people to share between
their investments.

Staff training and expertise - who needs to learn
and own these tasks - and what value proposition
are they providing - are also key questions
ahead of which widgert do they use.

Anyway - in Santa Fe one of the things we did 
was look at the raw laundry list of key functionality.

If we want to have interoperability and collaboration
with registries then there are three vital places
we need to start:

1) Classification scheme mechanism - this must 
     be the same across all.

2) Core API functions subset that all support with
     basic search, content storage and retrieval 
     capabilities in a collaborative federated way.

3) Common association model - where links 
     between content across registries are
     maintained and supported by the participants.

If this looks like the right road map - then we 
need more than just a glib "we support all that,
and all the other guys should just implement
our spec's".

Everyone needs to have common ownership,
and more important a roadmap for the future
maintenance and enhancement of those base
specification capabilities.

So part of the solution here is a strong 
executive story that everyone can buy into,
based on fulfilling the users best interest
and needs.

Identifying those components and laundry
list would certainly be an excellent check
point in time and space for a collaborative
working group IMHO.

Now over on the OASIS messaging TC side 
we evolved the ebMS work, and now I'm 
wondering if the right moment has arrived
to create ebRS - eBusiness Registry Services?

This would assume and fold into it all the existing 
base but with an aligned roadmap for the 
future based on collective user needs and then
delegation via focused sub-committees under
ebRS to really bring home the goods for the
users.

I feel really bad about even comtemplating this
however - because obviously this is a chunk of
work - and right now I'm only focused on one
little piece of it with the OASIS CAM component.

But then again Rome was not built in one
day by people with fear of climbing up the
steeper path.

Thanks, 

DW.
==========================================
Message text written by Farrukh Najmi
>
I beg to disagree Matt. This may be a surprise to some but ebXML 
Registry actually provides a much richer ability to publish and discover 
web services than UDDI. Let me explain why I believe this to be true:
<