MHonArc v2.5.2 -->
wsia message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [Elist Home]
Subject: Re: AW: [wsia][wsia-requirements][E922]
I agree that this is a requirement for the reader of WSIA rather than
having any technical impact on the work of the committee and has value in
that it explicitly states that we are considering the developer that needs
to integrate with a legacy app. When we get around to discussing exemplary
applications, the idea of explicitly wrapping some legacy app may be a good
reinforcement of that message.
Sean Fitts
<sean@crossweave. To: Dan Gisolfi/Somers/IBM@IBMUS, wsia@lists.oasis-open.org
com> cc:
Subject: Re: AW: [wsia][wsia-requirements][E922]
05/09/2002 11:44
AM
Dan, I have to respectfully disagree with the statement that no one has
articulated the reason for this requirement. The reason, as articulated
by Stefan and Greg (among others), is that given that most WSIA
implementations will likely be based on existing applications and given
the importance of making sure that WSIA will be widely adopted, it is
important to make a specific statement that the WSIA group intends to
make this possible.
This may have little technical significance, but I think it is very
important
as a statement of intent. There have been too many standards efforts
which have ignored the current market realities and I don't any of us want
WSIA to join that list.
Now, you may disagree with this reason, but it has been articulated.
Let's debate the reasons why you disagree.
Personally, I don't think this goes far enough and I would like to see us
outline specific scenarios that show a company taking its existing
application assets and leveraging them as WSIA services. If this isn't
possible (and fairly easy), then it will hard for WSIA to be widely
deployed.
That said, I support the proposal to use the phrase "MUST NOT preclude"
(in place of MUST enable) as a middle ground.
Sean
At 07:34 AM 5/9/2002 -0400, Dan Gisolfi wrote:
>So far (no wherei sthi sdebate) has someone articulated the reason for
>this requirement. Where did it come from? My position is that we drop it.
>Web Services technologies (namely SOAP and WSDL) will address integration
>of legacy applications. WSIA will adress the description of how those
>legacy (enterprise) services will be interated with..
>
>Dan Gisolfi
>
>
>Rich Thompson/Watson/IBM@IBMUS on 05/08/2002 08:33:45 AM
>
>To: wsia@lists.oasis-open.org
>cc:
>Subject: Re: AW: [wsia][wsia-requirements][E922]
>
>
>
>
>Do we really expect the specification to do anything to enable this
>wrapping of legacy applications? I think the intent to to not preclude
>developers from interacting with any back end system they want to.
>
>
>
> "Beck, Stefan"
> <stefan.beck@sap. To:
> wsia@lists.oasis-open.org
> com> cc:
> Subject: AW:
> [wsia][wsia-requirements][E922]
> 05/08/2002 03:28
> AM
>
>
>
>
>
>Whats about:
>
>The specification MUST enable Producers to provide existing legacy
>applications and infrastructure as WSIA compliant Web Service.
>
>Stefan
>
>
>-----Urspr�ngliche Nachricht-----
>Von: Timothy N. Jones [mailto:tim@crossweave.com]
>Gesendet: Montag, 6. Mai 2002 20:00
>An: wsia@lists.oasis-open.org
>Betreff: RE: [wsia][wsia-requirements][E922]
>
>
>
>Is there a reason this shouldn't be a "must", i.e.:
>
> The specification MUST not preclude Producers from providing the
>capability to support legacy applications and infrastructure.
>
>As long as the protocol between Consumer and Producer is WSIA, it
shouldn't
>matter what else the producer is doing on the backend.
>
>Tim
>
> > Dan, I can see your perspective, but consider the consequences if we
>produce
> > a specification that prevents us from integrating with legacy
>applications.
> > Although we are in the domain of web services, the world will not
become
> > fully WSIA aware for several years, and many of the implementations
will
>be
> > producers exposing existing applications.
> > Without the ability to integrate the adoption rate will be low, which
>will
> > lead us down the path to obscurity.
> > I support Eilon's reworded statement, though I'm not sure that
> > 'infrastructure' adds anything to the requirement.
> > Regards
> > Greg
> >