[ebxml-dev] Managing the complexity of ebXML is the key of successbesides marketing

[Wolfgang März <w@maerz.do.uunet.de>]

Dear all,

first of all I confess that I am a convinced supporter of ebXML in my
industry (e-business for energy markets).

What I often hear about ebXML is (1) confusion about ebXML versus Web
Services, (2) the doubted need for it in an installed EDIFACT
infrastructure, (3) too complex, too many options, (4) not much vendors, not
mature, (5) no support by the two big ones MS, IBM.

Regarding (1) and (2) it is important for the ebXML marketing to compare
ebXML versus Web Services and EDIFACT on "management level", that means with
less than two pages.

Point (3) is a problem also for global any-to-any compatibility. What is
needed are ebXML profiles to be able to scale the technology from low to
high requirements. But that is not enough. There are also too many
"competing" options. The best example for this is the ebMS. The security and
reliability options fill a one page table. The only light are digital
signatures with only one option (XML Digital Signature /MIME). If it comes
to other options of security you have to deal e.g. with XML Encryption
/MIME, S/MIME v3, OpenPGP/MIME, TLS or SSL v3, IPsec (for VPN). I would like
a profile with simply something like this: reliability no/yes, digital
signature no/yes, encryption no/yes based on only "one" technology. That was
the success of the Internet.

Also not a good example is the content work of UN/CEFACT which is slow,
sometimes not readable and too complex. The ebXML Specs neeed 18 month to
complete. The work of UN/CEFACT seems to be a never ending story. If you
take UMM for example you have many versions, now 10 and 12 is expected soon.
No end in sight. If you model a system using UMM you have only the paper
sheet on which UMM is written (if you understand the stuff). But if you use
tools as e.g. Rational Rose UML and Rational RUP you have no means to
customize your tools in an automated fashion with the UMM meta model and the
UMM stereotypes.

Point (4) I guess will be hopefully no problem in the near future.

Point (5) really hurts. The standing of ebXML would be better if we would
find a standard body home for it. This is the case of the content work done
by UN/CEFACT but not with the Technical Specifications managed by OASIS. For
this the home could be perhaps ISO or W3C.

Regards
Wolfgang Maerz

MCC - Maerz Communication Consulting
Eichenmarkweg 16
D-44267 Dortmund
Germany