Re: [ebxml-bp] IBM to Support BPEL-Based Web Services

["ZBarch" <zbarch@rcn.com>]

David,



It seems to me Kenji meant under "automation" the ability of BPSS to be a

so called "configuration" file for Business System Interface (BSI) - i.e.
automation

of runtime execution of business transactions. If BPSS will be able to
define

such configuration file,  it will, in turn,  allow building of  the generic
BSI application,

which will be simply configured for each particular instance of eBusiness
implementation.

The Business process flow definition doesn't look like the strongest side of
BPSS in its present status.



What IBM has done is very simple - based upon BPEL they've built some EJBs,
included them into

Websphere plus Web Service support (kill two rabbits with one bullet). It's
a strong market pitch, taking into

account the current mind set of CIOs and CEOs.



ZB


p.s. By the way, I could build a prototype of this generic BSI in 4-6 months
having a sponsor with spare $200K -250K.

----- Original Message ----- 
From: "David RR Webber" <david@drrw.info>
To: "Kenji Nagahashi" <nagahashi@fla.fujitsu.com>
Cc: "ebXML BP" <ebxml-bp@lists.oasis-open.org>
Sent: Friday, April 16, 2004 10:54 PM
Subject: Re: [ebxml-bp] IBM to Support BPEL-Based Web Services


> Kenji,
>
> I would disagree here.
>
> I have built a BPEL editor as well as a BPSS one.
>
> BPEL - as John Yunker so rightly noted - represents
> climbing Mount Fuji - just to make a cup of tea
> every morning.
>
> Joe Chiusano has come the closest I've seen to
> defining a simple approach to BPEL - but that
> suffers from being too simplistic.
>
> Don't forget before you can write any BPEL - you
> first have to learn and write WSDL.  It could not
> be more complicated if you tried!
>
> BPEL really is in danger of being a mini-4GL -
> and then there are three ways of doing most
> things - so which one do you pick and why?
>
> Now of course Microsoft and IBM have tried
> to hide all this underneath GUI interfaces that
> write it all for you.  The snag then is that the
> BPEL it creates will only work inside their
> toolset (naturally!) and nowhere else.
>
> Maybe you can reserve judgement on just
> what kind of lowlevel trickery is needed in
> BPSS - my sense is not much.
>
> My experience is that for building
> eBusiness process definitions BPSS has
> an excellent feel to it - and the right level
> of detail and control mechanisms to make
> it attainable by business designers (this was
> John Yunkers another point).
>
> That is why I am focused on the editing
> tool as the way to bring realization to the
> marketplace.  Once business designers
> realize they can do this themselves quickly
> and easily and are not beholden to a
> classic situation where BPEL programmers
> have to dictate everything to everyone else,
> as technology, not business - then the
> better off everyone will be.
>
> I believe there is every much chance that
> the marketplace will reject BPEL as being
> too complex and too arcane.  I say this
> because I've seen alternative process control
> scripting built by another market leader in
> this space - and their approach was compelling
> and simple.   Therefore my conclusion is
> that's only a matter of time before we have a
> Python style event horizon and someone like
> a James Clark is talking up their simpler easier
> method....
>
> DW.
>
> ----- Original Message ----- 
> From: "Kenji Nagahashi" <nagahashi@fla.fujitsu.com>
> Cc: "ebXML BP" <ebxml-bp@lists.oasis-open.org>
> Sent: Friday, April 16, 2004 10:00 PM
> Subject: Re: [ebxml-bp] IBM to Support BPEL-Based Web Services
>
>
> > Hi,
> >
> > David, I think you're misisng Duane's point... He's talking about
> > prioritization for greater market.
> > I see there are views on BPSS: One view puts higher priority on
> > "documentation", and the other puts higher priority on "automation".
> > For automation purpose, we need to define very formal and detailed
> > computational model for the language, pretty much like BPEL is doing. We
> > need lots of implementation experience for this goal...
> >
> > And my personal view is that "automation" almost always gets greater
> > market share.
> > See how BPEL is doing.  BPEL attracted many people for the ability to
> > automate the process and some BPEL engine has been built before any
> > design tools. If language can provide automation with good level of
> > abstraction, people just use it even if no good GUI is available for it.
> > How people used C compiler before Visual C?. How people used HTML before
> > any web authoring tools?
> >
> > I'm not saying "automation" view is right and "documentation" view is
> > not. "documentation" is an important usage of BPSS. And it is also
> > important to think about what BPSS can do for "automation". I'm just
> > wondering if many in ebBP are interested in this aspect.
> >
> > Duane, thanks for your kind words for our Ottawa presentation... It was
> > a great opportunity for both of us!
> >
> > Regards
> > Kenji
> >
> > David RR Webber wrote:
> >
> > > Duane,
> > >
> > > Its a Catch22 - and also resources.
> > >
> > > However - I can build the designer tools for V2.0 in a few days.
> > >
> > > Once I have those - then that creates more of a compelling
> > > reason for the announce on the open source BPSS engine.
> > >
> > > DW
> > >
> > > ----- Original Message ----- 
> > > From: "Duane Nickull" <dnickull@adobe.com>
> > > To: "David RR Webber" <david@drrw.info>
> > > Cc: "Matthew MacKenzie" <mattm@adobe.com>; "ebXML BP"
> > > <ebxml-bp@lists.oasis-open.org>
> > > Sent: Friday, April 16, 2004 4:01 PM
> > > Subject: Re: [ebxml-bp] IBM to Support BPEL-Based Web Services
> > >
> > >
> > >
> > >>
> > >>David RR Webber wrote:
> > >>
> > >>
> > >>>My top priority coming few weeks is to get BPSS V2.0
> > >>>editing tools done - so people can build their
> > >>>process models easily.
> > >>>
> > >>>Next up we need someone to announce an open source
> > >>>BPSS engine... room for a joint collaboration on this
> > >>>me thinks.
> > >>
> > >>This is highly illogical.  Who would have a need to have BPSS design
> > >>tools or BPSS before you have an engine to run them.  I would argue
for
> > >>execution engine first, then the designer toolsets or at least at the
> > >>same time.
> > >>
> > >>Fujitsu demonstrated some pretty cool stuff in this space in Ottawa
last
> > >>month.
> > >>
> > >>Duane
> > >>
> > >>
> > >>
> > >>-- 
> > >>Senior Standards Strategist
> > >>Adobe Systems, Inc.
> > >>http://www.adobe.com
> > >>
> > >>
> > >>
> > >>
> > >
> > >
> > >
> >
> >
>