I must admit that it
has been a little frustrating to define a widely accepted articulation between
BPSS and BPML. Political issues aside, they should be very complementary.
When I worked at BPMI
I settled for a binding framework between BPML and BPSS, such that BPML could
be bound to other B2B infrastructure, not just ebXML.
It appears to me that
there is only one B2B infrastructure left on the table today (ebXML) and this
is not going to change. BizTalk framework is dead, RN has pledged to adopt
ebXML, large industries like the Automotive industry have standardize both
demand and supply chain on ebXML….
In the light of this
observation, I am now in favor of bringing some of the semantics of BPSS (not
the protocol part) into BPML. Namely BPML must deal with Business Transaction
as well as it deals with Web Services.
BPML is much more
focused on Web Services rather than B2B, this is unfortunate – they are
missing a huge opportunity. Web Services as they stand today (and in light of
the design of BPSS) will not be able to efficiently support complex B2B
transactions. Everybody knows this and talk about it in private.
In the ideal world,
BPML would bring User, Enterprise systems and B2B interaction in one compelling metamodel.
Web Services are just one side of the equation. BPSS is another one, missing is
how do you model user interactions effectively? I actually created our own PML
which does exactly that. I don’t think it is really complex to do it, it
is rather a mater of vision (or pragmatism).
I run a little web
site that collects information about PMLs like BPML, WfMC, … and try to
describe an articulation between ebXML, PMLs, Web Services and business
languages. Here is a link that shows a possible binding between BPML and BPSS (http://www.ebpml.org/bpss_binding.htm).
To be fair XLang, and
WSFL are not smarter but MS and IBM are so entrenched in selling Web Service
infrastructure that I can excuse them, why do BPMI has to do the same mistake?
I don’t know. My prediction is that BPML will suffer the same fate of
other Process-Oriented languages which all failed to look at the problem
globally and just addressed one aspect of business processes, therefore
limiting tremendously the value and the adoption of the technology.
There is more in the
book that Pim references in case you want to go further.
Feel free to
contribute to www.ebpml.org if you want to.
200 Fifth Avenue
Waltham, MA 02451
From: James Bryce Clark
February 22, 2002 10:47 AM
Subject: Re: [ebxml-dev] BPMI?
To: "James Bryce Clark" <email@example.com>,
Is there any kind of documentation/minutes
available that gives the details of the cross-discussions held and what was the
outcome of the discussions? Would also be great to get pointers to the
work being done in making all the available, related specifications
Herr van der Eijk
is exactly right -- the most careful thought on this that I've seen is JJ's
published chapter, and JJ's own comments in our various discussion
groups. If you want to look at the informal discussions as well,
ebXML policy is to keep them publicly available. I believe they are
presently located in the last 3 months of the ebxml-bp v1.0 archive at http://lists.ebxml.org/archives/ebxml-bp/.
James Bryce Clark
~ VP and General Counsel, McLure-Moynihan Inc.
~ Chair, ABA Business Law Subcommittee on Electronic Commerce
~ 1 818 597 9475 firstname.lastname@example.org
~ This message is neither legal advice nor a binding signature. Ask me