Re: [ubl-dev] ebXML approval retrospective

["Fraser Goffin" <goffinf@googlemail.com>]

As I said in my earlier email, I broadly support ebXML for many of the
reasons you and others site. I have absolutely no doubt that there are
and will continue to be large scale and highly successful
implementations that use ebXML, ... but.. the same can be said for the
other options that you mention.

I work for one of the largest financial service organisations in the
UK and we operate a number of distribution channels which cover both
B2B as well as public self service (direct) models (both are highly
secure so this isn't really a differentiator). Our main B2B operation
is via a secure 'web services' portal operated by the UK General
Insurance industry standards body (as are the data standards which are
used to specify business processes). So in general, we not 'punting'
content to the rest of the world hoping to make a sale (if nothing
else legal and regulatory requirements would make this somewhat
difficult).

From the technical perspective, most large organisations need to deal
with highly hetrogeneous environments and of course this is one of the
characteristics that matches with the goals of the platform and
technology independance of service based implementations (web services
or other). Hetro environments also tend to mean lots of different
vendors, so naturally it is important from an operational support
persepective that there is commonality in both the support for
'mainstream' specifications/standards and the (consistent) quality of
those implementations. Like many these days, the prevalent model for
technology solutions is for COTS and off-shore/out-sourcing, so the
case for building and maintaining custom code to paper over the cracks
is pretty hard to make in most cases.

Back to ebXML. This is a suite of specifications, some simple (and
thus more likely to have some degree of support in vendor packages -
or even 'roll your own - thats what we did for ebXML MS), some not
(often with the reverse effect). Of course all have a valid place in
any capability model but they are not without competition.

Vendors will do what the market-place dictates and, although I see
ebXML growing, it is not the loudest voice that can be heard even for
implementations that, as you suggest, would perhaps benefit the most.
ebXML [can] raise the bar somewhat, so it is unsurprising that many
remain unconvinced whether it is worth making that kind of
committment.

So yes, I hear that there is support in Oracle iHub (although our
recent experience in talking with our own Oracle technical account
management is that they are still somewhat playing catch-up - with the
whole services thing not just ebXML - as you will no doubt be aware,
they have made some recent aquisitions to help speed up this process,
but integration takes time !). We are also a very large IBM shop and
there are some possibilities there also. As you might imagine we deal
with a lot of different vendors across probably most platforms that
you could think of. But when you talk to them about ebXML, many no
very little, some know about ebXML MS but not much else, and others
[most] usually prefer to peddle WS-*.

Anyway as you point out, it isn't necessarily (or doesn't have to be)
a contest between ebXML and WS-*, they are complementary and can be
used for different purposes. I'm not sure everyone is aware of, and
ready for, investment in both though. Making technology choices is
IMO, more often than not about 'timing' rather than technical
capability.

Fraser.

On 14/05/06, David RR Webber (XML) <david@drrw.info> wrote:
> Fraser,
>
> I would argue that for trusted broad reliable SOAP-based exchanges ebXML
> offers many advantages over the competition.
>
> I stress trusted and broad - because obviously there's any number of
> lightweight REST type stuff going on - take aws.amazon.com and ebay's
> equivalent for examples - but the model there is one of punting content
> to as much of the world as possible.  Ditto for a lot of SOAP stuff (aws
> and ebay both offer WSDL as well as REST) - people put a server up there
> and expect a ton of people to just hit it regardless of who and where.
>
> When you need secure, descreet, trusted exchanges where the content is
> only exchanged between you and a specific partner - then clearly the
> B2B model outshines the web model.  And particularly in ease of
> credentially using a CPA approach.  Most credentialing I've seen for
> SOAP style stuff is either too superficial or completely unwheldy and
> akin to the old EDI trauma just to get setup and running.
>
> Also - I'd take issue with the assertion on the depth of ebXML
> implementations - as I noted before Oracles ebXML in its iHub is
> definately industrial strength - and then vendors like Cyclone Commerce
> and Xenos - implementing 8,000 dealership interfaces for GM, and health
> insurance for the whole of Norway respectively - show just how robust
> commercial offerings are.  Similarly the Hermes open source solutions
> are also rock solid and production proven open platforms tools with an
> experienced community support base out there. For example electric grid
> distribution marketplace for whole of Austria in realtime.
>
> What I see the win is knowing the strengths of each component and when
> to use it.  This is exactly the same lesson as with realtime and batch
> EDI - now we have WS* and ebXML - same solution mix.
>
> DW
>
>  -------- Original Message --------
> Subject: Re: [ubl-dev] ebXML approval retrospective
> From: "Fraser Goffin" <goffinf@googlemail.com>
> Date: Sat, May 13, 2006 1:22 pm
> To: fulton.wilcox@coltsnecksolutions.com
> Cc: ubl-dev@lists.oasis-open.org, xml-dev@lists.xml.org
>
> > In the same context, SOA/SOAP model is moving ahead none too rapidly either,
> > with most successes being "intramural."
>
> I disagree. The pace of adoption for SOA/SOAP integration for B2B
> varies. In my own case it is the *primary* model not, as many suggest
> one which only follows successful internal integration exercises
> (indeed many are less enthuiastic about using SOA/SOAP internally
> where all parts of the solution are available to implementers).
>
> As far as the challenges of versioning and extensibility are
> concerned, I agree whole-heartedly. Working with a standards body not
> dis-similar to UBL (for UK General Insurance Standards), there are
> identical problems and, not many easy / obvious solutions. This is why
> I (amongst many) having been staying close to the various approaches
> that UBL are considering (genericode, nvdl, etc.).
>
> Some time ago, I took my organisation down the ebXML route for one
> particular project (at least the the MS specification) but as others
> may have found the support in terms of off the shelf implementations
> was not there and there was (is) enormous pressure to use WS-*. I have
> long been a supporter of ebXML, but I remain to be convinced that it
> will eventually gain sufficient market traction. Yes some (perhaps
> even many) of the mainstream players have support for some of the
> specifications, but under challenge many of these prove to be somewhat
> incomplete and sometimes a bit superficial. I am pleased though that
> this effort continues. Many of the ideas that have been adopted by
> other specs have their roots in ebXML so in that regard, it *is* a
> success story.
>
> Fraser.
>
> On 12/05/06, Fulton Wilcox <fulton.wilcox@coltsnecksolutions.com> wrote:
> > Dave,
> >
> > The dismal evaluation I presume relates to Klaus's definition of success -
> > to quote from Klaus's blog, that the ebXML standards would enable "...anyone
> > out of the blue to engage with anyone else anywhere else."
> >
> > The glass is about 1% full in meeting that expectation with
> > machine-to-machine dialogs.
> >
> > In the same context, SOA/SOAP model is moving ahead none too rapidly either,
> > with most successes being "intramural."
> >
> > The currently ongoing debates concerning UBL schema content constraints are
> > symptomatic of the fundamental forces of complexity and "particularity" that
> > must be accommodated before that glass gets very full.
> >
> >
> >                                                Fulton Wilcox
> >                                                Colts Neck Solutions LLC
> >
> > -----Original Message-----
> > From: David RR Webber (XML) [mailto:david@drrw.info]
> > Sent: Thursday, May 11, 2006 5:32 PM
> > To: Bill Chessman
> > Cc: ubl-dev@lists.oasis-open.org; EDI-L@yahoogroups.com; Klaus-Dieter Naujok
> > Subject: RE: [ubl-dev] ebXML approval retrospective
> >
> > Bill,
> >
> > I've just taken a few moments here to read Klaus's piece.
> >
> > Wow - even by Klaus's standards this is a dismal and self-deprecating
> > attack.
> >
> > I'm here to tell everyone that the world of ebXML is nowhere near as
> > forgotten and
> > doomed as Klaus would have you think - and that Klaus really is down on
> > himself
> > when he should be reading more of the items on
> > http://ebxmlforum.net/5-years/
> > and taking some credit for positive outcomes!
> >
> > Afterall noone ever said this was going to be easy!!
> >
> > I'd take issue with many of Klaus's assertions and his outlook.
> >
> > 1) Turns out the big players DO have current ebXML support in their
> > products, Oracle, IBM ,Sun, BEA, SAP, Fujitsu, Sysbase, etc will all
> > sell you their solutions with ebXML support in them.  And Oracles' is
> > brand new - their iHub support for ebXML was released at end of 2005
> > (and its very good BTW as I've seen it close up).
> >
> > 2) He forgets that ebXML B2B remains the ONLY certified XML
> > interoperable solution out there - with more than a dozen vendors
> > completing the interoperability suite.
> >
> > 3) There are industries that have standardized on ebXML - PIDX
> > (agro-chemical), HL7 healthcare, electrical power - and we see the
> > auto-industry also garnering the benefits of using ebMS too and growing
> > its use.
> >
> > 4) There are large scale deployments - the whole of Norway is using
> > ebXML for healthcare insurance, and the NHS in UK is using it for
> > supplychain support.  So OK - in Klaus's self-deprecating stance -
> > these are just tiny little countries on the fringe of Europe that noone
> > pays much mind to.
> >
> > 5) Governments really do appreciate the value of open public standards
> > and open public implementations.  While ebXML may not have gone in the
> > direction Klaus wanted for it - nevertheless it is healthy and primed
> > to be an important part of where the internet and eBusiness is going -
> > driven by the open source and open services revolution.
> >
> > 6) ebXML registry for secure document storage (aka IHE/XDS) is about to
> > be THE solution - and ditto for ebXML Regsitry and semantic content
> > retrieval.  There is some serious momentum and sophisticated community
> > building behind this - and yes - this is exactly the sort of thing we
> > wanted registry to drive.
> >
> > 7) The new OASIS BPSS work is ground breaking - and OASIS just approved
> > the BCM work - that is founded
> >    from the original ebXML / CEFACT UMM concepts.  The ebXML solution
> > stack is changing the way people
> >    approach engineering their solutions - and yes - this does take time
> > to seep into the IT conciousness.
> >
> > Klaus - come on man - GIVE ME A BREAK!  Just because noone is having a
> > huge party and lotsa Hollywood style bruhah does not mean you can sit
> > at your computer terminal and wail into your weak thin American beer,
> > or cheap Californian wine and depress us all!  Just because ebXML did
> > not do everything you thought it should you cannot ignore what your
> > child has achieved!  I feel like the prodigal son here - and Dad just
> > does not want to entertain anything good could have come of his
> > offspring once he left home and went off on his own.
> >
> > This is the new wave internet community world and we know better.  Grass
> > roots is more important than Redmondian brash PR.
> >
> > Get on board the new ebXML and enjoy!  We have an awful lot to be proud
> > of and much to thank you for - so quick raining on your own parade -
> > I'm not going to join you in wanting to suck on lemons when I can eat
> > cake and pizza and drink great English beer!!
> >
> > The best of ebXML is yet to come and we can be very proud of what we
> > have already accomplished and the foundations we have created...
> >
> > Remember when everyone trashed Apple as a crippled and broken spear and
> > Klaus was the only guy we knew who had an Apple Mac?!?
> >
> > Cheers, DW
> >
> >  -------- Original Message --------
> > Subject: [ubl-dev] ebXML approval retrospective
> > From: "Bill Chessman" <bill.chessman@inovis.com>
> > Date: Thu, May 11, 2006 1:01 pm
> > To: <ubl-dev@lists.oasis-open.org>, <EDI-L@yahoogroups.com>
> >
> > All,
> >
> >
> >
> > Anybody else notice that today, May 11, 2006, is the 5th anniversary of
> > the approval of the original ebXML project?  I see that the chairman of
> > the original project (Klaus-Dieter Naujok) has put up some retrospective
> > commentaries on his blog page at
> > http://www.klauskorner.com/MyBlog/MyBlog.html.  For those nostalgic
> > folks that participated, there's even a video from that closing approval
> > meeting.
> >
> >
> >
> > Still wondering where the story goes from here...
> >
> >
> >
> > Best regards,
> >
> > Bill Chessman
> >
> > Inovis(tm)
> >
> >
> > ---------------------------------------------------------------------
> > This publicly archived list supports open discussion on implementing the UBL
> > OASIS Standard. To minimize spam in the
> > archives, you must subscribe before posting.
> >
> > [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> > Alternately, using email: list-[un]subscribe@lists.oasis-open.org
> > List archives: http://lists.oasis-open.org/archives/ubl-dev/
> > Committee homepage: http://www.oasis-open.org/committees/ubl/
> > List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
> > Join OASIS: http://www.oasis-open.org/join/
> >
> >
> > ---------------------------------------------------------------------
> > This publicly archived list supports open discussion on implementing the UBL OASIS Standard. To minimize spam in the
> > archives, you must subscribe before posting.
> >
> > [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> > Alternately, using email: list-[un]subscribe@lists.oasis-open.org
> > List archives: http://lists.oasis-open.org/archives/ubl-dev/
> > Committee homepage: http://www.oasis-open.org/committees/ubl/
> > List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
> > Join OASIS: http://www.oasis-open.org/join/
> >
> >
>
> ---------------------------------------------------------------------
> This publicly archived list supports open discussion on implementing the
> UBL OASIS Standard. To minimize spam in the
> archives, you must subscribe before posting.
>
> [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> Alternately, using email: list-[un]subscribe@lists.oasis-open.org
> List archives: http://lists.oasis-open.org/archives/ubl-dev/
> Committee homepage: http://www.oasis-open.org/committees/ubl/
> List Guidelines: http://www.oasis-open.org/maillists/guidelines.php
> Join OASIS: http://www.oasis-open.org/join/
>
>