Re: Proposal: A Registry Browser GUI tool

[Nicholas Kassem <Nick.Kassem@eng.sun.com>]

David,

Clearly you don't see the regrep Client and associated services as just 
another ebXML Business Process (granted that it is a specialized BP).

I don't see any reason why a WAP enabled device can't be a consumer of 
regrep services. But I wouldn't put a Cell Phone at the design center and 
gut all the ebXML TRP functionality to meet the needs of this single 
device, I would proxy the client and expose it as an ISP service. Whether 
TPA assembly is likely to happen over a shoe phone is an entirely different 
matter - personally I think not.

I know you have one specific solution in mind but do you feel comfortable 
that you have the necessary consensus *within ebXML WGs* to say that the 
regrep work is all done and the answer is 42 ? You may have but I haven't 
been privy to any votes within regrep.

What makes RegRep so specialized an application that it doesn't need ebXML 
TRP - can we get passed this obsession with the Client Run-time environment 
- is there a technical basis for the position you are taking ?

Nick



At 04:04 PM 9/14/2000 -0400, David RR Webber wrote:
>Message text written by INTERNET:dick@8760.com
> >
>ebXML message REQUEST: I want your trading partner profile
>ebXML message RESPONSE: OK; here it is
>
>In the above scenario the ebXML header document in the REQUEST message
>identifies the service (a.k.a. procedure) being invoked, in this case it
>would
>be a regrep service (e.g.  TPINFO) and the payload contains the "call
>parameters" (e.g. DUNS Number of trading partner profile being requested).
>The
>RESPONSE contains an ebXML message indicating a SUCCESS status in the ebXML
>header document, the payload  contains the requested information (trading
>partner profile) in XML (tpaML).
>
><<<<<<<<<<<<<<<
>
>Dick,
>
>OK -  your note is the high level interface.  Between this an the actually
>RegRep
>"a miracle happens" to resolve the "I want your trading partner profile"
>into
>bits-n-bytes that understand the RegRep structure.   I'm concerned about
>that
>lower level.   Also - RegRep should be accessible via that lower level
>directly - and that is obviously a performance and access path we want.
>
>Within the bit-n-byte level the access profile should marry to the TRP one,
>so all that jives consistently.
>
>That's what I'm seeing here.
>
>DW.