Re: XML Business Document Library Project Team

[Jon Bosak <Jon.Bosak@Sun.COM>]

I spent all last week thinking about how to respond to the message
of 21 September announcing the formation of the XML Business
Document Library (XBDL) Project in the eBTWG.  My preference would
be to say nothing.  Unfortunately, that message, sent to the
eBWTG, ebXML, and ebXML-dev discussion lists, contains some
misstatements of fact that are too important to let pass without a
reply.

First we have the following:

| The eBTWG Executive does understand that the XBDL Project is in
| concept identical to the recently approved work of the OASIS UBL
| TC. After consultation with the CSG we agreed to proceed with the
| approval of our project. The reason is very simple, this work is
| not related to ebXML infrastructure work, but related to ebXML
| content and context work which is the agreed responsibility of
| UN/CEFACT.

The characterization of UBL as "the agreed responsibility of 
UN/CEFACT" is false.  The stated aims of UBL are as follows:

   1. To avert a crisis in electronic business caused by competing
      XML business-to-business document standards by choosing as a
      starting point an existing XML business document library as
      the basis for creating a new "Universal Business Language"
      that will be a synthesis of existing XML business document
      libraries.

   2. To begin with xCBL 3.0 as the starting point and to develop
      the standard UBL library by mutually agreed-upon changes to
      xCBL 3.0 based on industry experience with other XML
      business libraries and with similar technologies such as
      Electronic Data Interchange.

   3. To develop UBL in light of standards/specifications issued
      by UN/CEFACT, ISO, IEC, ITU, W3C, IETF, OASIS, and such
      other standards bodies and organizations as the UBL TC may
      deem relevant.

   4. To harmonize UBL as far as practical with the ebXML
      specifications approved in Vienna (May 2001), with the work
      of the Joint Core Components initiative (a joint project of
      ANSI ASC X12 and the UN/EDIFACT Working Group), and with the
      work of other appropriate business information bodies.

   5. Ultimately, to promote UBL to the status of an international
      standard for the conduct of XML-based electronic business.

This program of work does not belong at all to the part of ebXML
assigned to UN/CEFACT.  It does, however, fall squarely within the
charter of OASIS, a nonprofit consortium that has been devoted to
the advancement of structured information standards since 1993 --
years before ebXML was even an idea. The development of industrial
standards for tag languages has been part of OASIS from the
beginning, and the OASIS TC process is specifically designed for
the creation of XML languages.  OASIS did not agree to abandon its
raison d'etre when it cooperated with UN/CEFACT in the ebXML
initiative, and it did not agree to abandon it in Vienna.
Assertions that OASIS agreed to cede the development of XML
syntaxes to UN/CEFACT are simply untrue.  This was not the intent
of the representatives of OASIS in Vienna, and is not supported by
the Joint Memorandum of Understanding signed in Vienna.

Another part of the message of 21 September that needs correction
is its statement that "the approved XBDL Project has been part of
the UN/CEFACT work plan as it relates to the approved UN/CEFACT
eBusiness vision, it is not a totally new item."

The fact is that the development of XML syntax for business
messages was formally ruled out of scope for ebXML at the May 2000
meeting in Brussels.  It was as a result of that decision by the
ebXML Steering Committee that I started trying to organize an
OASIS effort to develop XML syntax -- as the CEFACT leadership
knows, since they were among the first people I approached with
the idea and the first to reject it.

It's no secret that the people trying to develop UBL would have
preferred to do the work in UN/CEFACT.  At their meeting in April
2001, the members of the UBL Organizing Committee resolved as
follows:

   That the Organizing Committee is chartered to prepare and
   propose the formal establishment of a new T-level XML Syntax
   Subworking Group of the EDIFACT Working Group, subject to the
   following conditions:

    - xCBL 3.x is the starting point

    - The SWG's Terms of Reference are similar to the OASIS
      technical committee process

    - The work is governed by an intellectual property policy that
      vests ownership of the proposed work in the U.N. and ensures
      royalty-free licensing

    - The membership policy allows the members of the Organizing
      Committee and other experts to join the new SWG as voting
      members

   That if the effort to establish a new SWG in EWG fails, the
   Organizing Committee is chartered to prepare and propose the
   establishment of a CBL Technical Committee in OASIS.

If the EWG XML Syntax group proposed in this resolution had been
allowed to form, it would have given the UBL effort a home in
CEFACT.  But the CEFACT Steering Group (CSG) discouraged the
formation of a subworking group that would have been governed by
the EWG membership on the grounds that the EWG was about to go out
of existence.  Denied the creation of an appropriate subworking
group in the EWG, the UBL Organizing Committee fell back to the
alternative it had resolved upon earlier and proceeded with the
formation of the OASIS UBL Technical Committee.

Everyone should clearly understand that there is a basic
difference of philosophy between the CSG/eBTWG leadership and the
people involved in the UBL effort.

The CSG/eBTWG leadership are recognised advocates of top-down
design methodologies. Anyone who doubts this needs only look at
the committee structure they originally proposed for the eBWG and
the committee structure that is now proposed for the eBTWG.  The
technical program of the eBTWG is organized around concepts
inherited (inappropriately, in the opinion of many) from
object-oriented programming.

The CSG's goals were summed up in a document titled "UN/CEFACT's 
Expectation for ebXML's Deliverables" that was prepared by members of 
the CSG for the ebXML Executive early in the course of the ebXML 
project. The core of the CSG position with regard to the design of XML 
business schemas is stated in that document as follows:

   In addition [to identifying and defining standard "core
   components"] the CC team was to create the transformation rules
   that would automatically create, from a UML class diagram,
   (stored in XML in the repository) the XML information
   parcels. These would become the base business objects for any
   ebXML compliant business and information model in order to
   achieve interoperability at the data level.

   In addition to UN/CEFACT many leading international industry
   groups have followed UN/CEFACT's direction. In addition to GCI,
   SWIFT, RosettaNet, TM-Forum, and EWG (UN/EDIFACT) all have made
   public statements about their use of UML for business process
   and information modelling. All use business objects and
   transformation rules to generate automatically the XML payloads
   from those models.

Unlike the proposed CSG/eBTWG methodology, UBL takes an approach
more in keeping with modern software development. It asks, What
already works?  What do we already know about business? How can we
easily realize what we need? How can we model what people really
do rather than what object-oriented programmers think they do? And
then it sets about implementing the already existing business
knowledge in the fastest and surest way possible -- by taking an
existing set of XML business grammars and evolving it into a
standard, aligning it with the ongoing semantic definition work in
UN/CEFACT, and modifying it as necessary to include the best parts
of other existing XML libraries and incorporate the hard-won
practical business experience gained in 20 years of electronic
data interchange.

Thus, while the UBL project is intended to develop a standard set
of XML business schemas from the bottom up, based on existing
practice, the CSG/eBTWG approach is a top-down effort intended to
develop a standard set of formal business models from which XML
schemas can automatically be generated. In the CSG/eBTWG vision,
XML schemas are just one of an unlimited number of syntaxes that
could automatically be generated to instantiate "business
objects." In the UBL vision, XML schemas occupy center stage, and
their creation is the collaborative work of XML experts and
business experts.

To those of us trying to develop UBL as a practical near-term
solution to the XML business document interchangeability problem,
these appear to be different strategies and goals.

The CSG/eBTWG leadership did not like the UBL idea in May 2000,
and, despite their blessing of the new XBDL Project Team, I am
convinced that they don't like it now.

In advance of the UBL Group meeting 13-14 August in Montreal, I
was told that if the UBL Group did not agree to put its work under
control of the eBTWG, a project team would be started in eBTWG to
compete with it. The XBDL announcement delivers on that promise.

Unlike the UBL TC, however, the XBDL project team will have to
engage in a serious examination, both technical and legal, of the
already existing commercial XML libraries and all the other
logical alternatives before it can begin to define a syntax (UBL
begins by assuming xCBL as the starting point, in no small part
because it is already widely deployed and legally unencumbered).
It's hard to see how this prerequisite analysis can properly be
carried out without materially slowing the XBDL effort.

For those of us interested in developing a practical near-term
solution to the XML business document interchangeability problem,
it's time to get down to work.  If you share this goal, I invite
you to join us at

   http://oasis-open.org/committees/ubl/

If you are not an employee of an OASIS member organization, you
will have to buy an individual OASIS membership to participate.
The annual dues of USD 250 per year compare favorably with the
cost of participation in other organizations that charge meeting
fees, and joining OASIS will entitle you to participate in a
number of other important industry XML initiatives as well.

I sympathize with everyone who wanted to see the UBL work take
place within UN/CEFACT; I wanted that, too, and for several months
this summer tried hard to make it happen.  I am greatly encouraged
that we may eventually achieve this goal by the "EWG Proposal for
the Future Structure and Organization for e-Business
Standardization within UN/CEFACT" that was approved by the EDIFACT
Working Group at their recent meeting in Rotterdam.  Unlike the
structure originally proposed by the CSG and still embodied in the
eBTWG, this new initiative by the EWG membership is far better
suited to the work we're attempting in UBL.

As I have recently said in the UBL Group:

   The EDIFACT Working Group embodies two decades of hard work by
   some of the world's leading experts in electronic data
   interchange.  UBL needs to be tied into that expertise so that
   it can be realized anew in syntax appropriate to the world of
   web services.  It needs to be integrated into the assessment,
   harmonization, and approval processes that the EWG has
   struggled so hard to implement.  It seems to me that the
   proposal approved by the EWG membership offers our best hope
   for accomplishing this integration.

I urge everyone who shares this goal to work with us in UBL to get
the beginnings of a practical solution in place and to support the
members of the EWG in their efforts to organize UN/CEFACT in a way
that would provide an appropriate long-term home for such a
solution.

Jon Bosak
Chair, UBL Group
Designated Chair, OASIS UBL TC