RE: [regrep] Should IBM being working with us on this?

["David RR Webber \(XML\)" <david@drrw.info>]

Bill,

 

Yes indeed. 

 

Added to this is that IBM and Oracle already have "registry" components in their product suites that provide widely differing roles.  Oracle's is for partner registration and configuration across their A/S product.

 

The phrase "your own mileage may vary" resonates here.  The challenge for us as a group is to engage with folks such as IBM in an open way that allows everyone to clarify use cases and share a better understanding of applicability rather than knee-jerk marketing of hammers and nails.

 

Putting together that SOA reading list is definately a good first step. 

 

From my stance the OASIS BCM specification and its discussion of registry functionality is #1 on the list.

 

Thanks, DW

"The way to be is to do" - Confucius (551-472 B.C.)

-------- Original Message --------
Subject: Re: [regrep] Should IBM being working with us on this?
From: Bill Majurski <bill@nist.gov>
Date: Mon, April 30, 2007 2:50 pm
To: "David RR Webber (XML)" <david@drrw.info>
Cc: "Farrukh S. Najmi" <farrukh@wellfleetsoftware.com>,
regrep@lists.oasis-open.org

David

One reason why IHE/XDS is moving forward is that it is backed by concrete Use Cases.  That allows folks who are not at the bleeding-edge of technology thinking to align with the group-think behind the technology. Given everything I have read about SOA so far I am still trying to figure out what new-thing SOA offers. Sometimes SOA seems like marketing, sometimes like technology. Most of the time it is just confusing.  

The first activity needs to be to assemble a technical reading list so the community can be on the same page when discussing this.  I am, so far, at the "King's New Clothes" stage of understanding.  I doubt that I am alone!

bill

 

On Apr 30, 2007, at 6:24 PM, David RR Webber ((XML)) wrote:

Farrukh,

 

Yes - I think we are agreeing here - it is NOT about the spec'!!!

 

There's already more than enough in the spec' to do this all IMHO.

 

It does come down to people purposing it and fashioning it - and our own positioning here.

 

Ann Manes et al - have consistently dissed ebXML reg/rep as not the right answer without even trying it.

 

When people do try it - as with IHE/XDS group - then we win converts.

 

However here's the rub in all this.  Folks what instant results - TV dinners - not "some assembly required".

 

So we are back to the implementation(s)...

 

IH/XDS is succeeding because people see the value and are prepared to invest significant effort themselves developing their CPs that overlay on top of the registry functionality.

 

The SOA crowd are different.  Their ROI is a lot less tangible - so they need low onramp low hanging fruit stuff.

 

An OASIS 60 day discussion list may help us descern the gaps here and bring us closer to this potential audience.

Cheers, DW

"The way to be is to do" - Confucius (551-472 B.C.)

-------- Original Message --------
Subject: Re: [regrep] Should IBM being working with us on this?
From: "Farrukh S. Najmi" <farrukh@wellfleetsoftware.com>
Date: Mon, April 30, 2007 1:52 pm
To: "David RR Webber (XML)" <david@drrw.info>
Cc: regrep@lists.oasis-open.org

David RR Webber (XML) wrote:
> Farrukh,
>  
> Errrr.  I think this is likely to drive folks away not bring them in
> here!?
>  
>   : -(
>  
> Remember - what your kids told you - "Dad - you may not always be
> right, but you are never wrong...."

Sorry if it appears that way. I did not intend to imply that they are
wrong.
My intent was to set the record straight that there *is* a registry and
repository standard that exists todays.
I am genuinely interested in what gaps exists in regrep specifically
compared to what their repository has to offer today.
Perhaps an IBM colleague on the lists would contribute to this
discussion?

>  
> So folks don't need reminded of that.  Frankly I could agree with IBM
> on this one - its NOT about the spec'.
>  
> Always put yourself in your customers shoes - and ask - why are they
> not using my product already?
>  
> Here are some inklings of what IBM wants:
>  
> 1) Highly extensible GUI with support for AJAX et al
>    - ok - this is work in progress for us with REST and
>    - "Any way, 2 minutes with Eclipse and the JBoss-IDE and I have
> auto-genenerated EJB3 (JPA) entity and SLSB facades from the existing
> omar 3 database schema.  How cool is that!"

I feel that both of above can be done with the current regrep spec.
IMHO, GUI is not a spec issues but a product / implementation issue.

>  
> 2) Easy deployment model
>    - so you are re-writing OMAR to get there...
>    - current installation and setup is a bear - definately not SME
> friendly

IMHO above is not a spec issues but a product / implementation issue.

>  
> 3) Built-in WSDL plug-n-play
>     - OK you have to buy the Sun release for that - not something IBM
> wants to hear...

Not sure I follow. Whatever Sun offers for above is available in
freebXML Registry and is based upon the draft ebXML Registry Profile for
WSDL.
>  
> 4) Simplier classification and management UI
>     - this is really linked to 1) - so content management is much more
> intuitive - something like the TortoiseCVS interface on the client
> side would be scary good...

Please see my point above that UI is not a spec / standards issue.
Recall the discussion was around a SOA Registry Repository standard.

>  
> 5) Linkage to Eclipse IDE so it aligns with the mainstream core toolset

IMHO above is an impl issue not a spec / standard issue.

-- 
Regards,
Farrukh

Web: http://www.wellfleetsoftware.com