[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [ubl-dev] SV: Where to next? - It's not the traction, nor supporting two stacks, nor any of this!
Mikkel, That was a not-so-short short note! I'll try to be shorter too - there is really no excuse for selling only one kind of "ice cream". I believe you are kidding yourself if you think this will make adoption easier and go faster. History, life, and prior experience tell us otherwise. You key statement was "I hope that we can all agree that as long as UBL payload can flow from one framework to another in a reliable and secure fashion, then we are all OK!". Re-reading the CBDI report - and that line "while the ebXML stack is working to be interoperable, the same is not true for WS-*" - oh really? We have learned over 20 years that there is no "magic" to B2B. I can give you the exact formula - and indeed that formula was embedded into ebXML - partners, transactions, process, context, transport, authentication, security, reliable messaging, queues and open marshalling/unmarshalling. You need to manage all those pieces. Problem is the W3C is not about B2B - so they built WSDL for point-to-point - mono-structure exchanges. So - horror of horrors - web services are only a subset of what B2B requires. Problem. Enter WS-I to "fix it". What does work here is the approach that Oracle and other vendors of B2B solutions ALL HAVE today. Basically they all have the ability to use the B2B formula above - and because the transport is pluggable - they can simply add in EDI, AS/2, ebXML, RosettaNet, HL7, etc, etc. But that is the rub - when they do that - there has to be compromise - and they make their own mini-standards - that cause incompatiblities and lack of full functional behaviour and interoperability across those interfaces. Then along came ebXML - and everyone realized - YES - we can make this so that it is an open, standard and everyone can support it. When I used BPEL and ebXML B2B together on the Helena project is worked seamlessly because of this very reason - the CPA-based approach already has all the B2B-formula baked in. But unfortunately WS-* is taking us back into those same dark mini-standard caves again...of 100 customizations and extensions for this and that and the other. All broken down into mydrid fragments of XML functionality that each have their own fiefdoms. And conveniently enough the "Big 6" have teams owning and working on building more complexity into each... I hope you implementation is going well - I too have implemented an open B2B and web service solution for NIH eReceipts - and it supports BOTH the WS-* interfaces and ebXML seamlessly out of the box. It's all about open design, open queuing systems and status control, and simple use of XML enabled mechanisms. I just question why - when I can do this with a team of 12 people - AND develop a full eGrants system at the same time - that NES and the whole of the Danish Government - and their consultants - cannot do it as well? I hope that "baker on the corner" is going to be able to do UBL himself - without needing a team of Big-6 consultants to push the buttons and hold his hand. But I fear that is not the case - instead he will be calling the 800-# help hot-line A LOT and getting very frustrated compared to just printing an invoice and mailing or faxing it. And your online forms - well - those will not cover everyones needs easily - the Grants.gov approach has already gone down that path - because the rules have to be too open, and no context can be applied - because XSD has no context mechanisms...and without context central rules cause downstream errors and failures. Again - been there done that - seen it already for the eReceipts project. Offering "treatments and pills" and people to talk to - to take the pain away - is no substitute for good simple design that prevents the pain in the first place. Cheers, DW "The way to be is to do" - Confucius (551-472 B.C.) -------- Original Message -------- Subject: [ubl-dev] SV: Where to next? - It's not the traction, nor supporting two stacks, nor any of this! From: "Mikkel Hippe Brun" <MHB@itst.dk> Date: Mon, February 05, 2007 4:38 am To: "David RR Webber (XML)" <david@drrw.info> Cc: "ubl-dev" <ubl-dev@lists.oasis-open.org>, <fulton.wilcox@coltsnecksolutions.com>, "Roger Bass" <roger@traxian.com>, "Sacha Schlegel" <sacha_oasis@schlegel.li>, "Stephen Green" <stephen.green@bristol.gov.uk> Dear David et al This will be a short post since I have to get back to work on rolling out UBL infrastructure. I think that we share the same goal: 1. That it should be possible to exchange business messages with very low barriers 2. and that it should be possible to do this on an open standards based and freely available platform. We are establishing this freely available platform. I think of this as putting up an ice cream machine with free ice cream. The ice cream comes in two variations, but we have to pick one of them: We have the choice of an ebXML organic flavor and the WS-* synthetic flavor with lots of GMO's. The kids says "Give us the WS-* flavor and we will eat lots of ice cream. We will integrate the ice cream in the heart of our applications and do all our home work." I don't mind serving synthetic ice cream if it can motivate my kids to do their homework. Once we get the ice cream flowing it will affect the ways we do business immensely. Adoption rate is critical because the business case for establishing an infrastructure is incredible. But we have to act fast! Public sector institutions are building infrastructure from the ground up every day based on WS-* standards and we have to stop the insane redundancy of efforts and waste of tax payers money. We can save 300 million USD every year in the public sector by establishing a shared infrastructure (We are only 5.5 million inhabitants in Denmark). It makes a world of a difference how fast we can establish the infrastructure. We do not have the luxury to choice of the organic road because it will affect adoption. We have support from very small companies as well as big enterprises. The toolkit that we are developing will make it very easy to use our infrastructure. I do not share the concern that our approach will only be to the benefit of governments and large enterprises. It's a business requirement that the baker on the corner should be able to plug into our infrastructure with OSS components. We have demonstrated that this can be done. Well - I am getting back to work. I am so thrilled by what's happening to UBL at the moment. NES has invited other EU member countries to a workshop in Brussels on February 15th to get more support. The UBL TC has succeeded in creating a viral technology and I congratulate all of the hard working TC members for their contribution. You have given us the payload, which will drive the establishment and roll out of infrastructure in Denmark and around the world. I hope that we can all agree that as long as UBL payload can flow from one framework to another in a reliable and secure fashion, then we are all OK! Best regards Mikkel >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]