Service Oriented Enterprise

Saturday, September 20, 2003

WSDL of the Day?  

If Don Box can do a Win32 API of the Day... surely I could do a 'WSDL of the Day' !

Hmmm... sounds like a lot of work. Anyone want to help? (If not, it will likely turn into WSDL of the Month :-)

posted by jeff | 10:26 AM

Friday, September 19, 2003


I just stopped by the OASIS Messaging and Coordination page and got a good laugh. It appears that Karl Best and friends have decided to take on a few more specs. So, in addition to BPEL, we have:

Business Transaction Protocol (BTP)
OASIS Asynchronous Service Access Protocol TC
OASIS Web Services Composite Application Framework (WS-CAF) Technical Committee
W3C Web Services Choreography Working Group
Web Service Choreography Interface (WSCI)
Web Service Composite Applications Framework (WS-CAF)
Web Service Context (WS-CTX)
Web Service Coordination Framework (WS-CF)
Web Services Transaction Management (WS-TXM)
Web Services Choreography Description Language (WS-CDL)
Web Services Conversation Language (WSCL)
Web Services Transaction Framework
Web Services Atomic Transaction (WS-AtomicTransaction) [replaces WS-Transaction-V1, Part I]
Web Services Coordination (WS-Coordination) [Version 2]
Web Services Business Activity (WS-BusinessActivity) [to replace WS-Transaction-V1, Part II]
Web Services Transaction (WS-Transaction) [Version 1]
Web Services Coordination (WS-Coordination) [Version 1]
WS Choreography

A Bounty
Ok, enough is enough. Can we put out a bounty to be paid to anyone that manages to kill a working group? Honestly, I will kick in my fair share. For starters, let's whack WS-Choreography - I'll pay $500 USD to the person that disassembles this working group. Surely there are others that will kick in too... My fear is that as quick as we knock them down, the fine folks at Oracle, Sun and Iona will create new ones to replace the old ones. Hence, I propose that we find the professional spec writers at the aforementioned companies new jobs. Got a startup? Offer one of these guys a job! Not because you need them... because you'll have to spend less on marketing to educate the world on why your product doesn't support some bullshit specification that these people made up. In the end... it will all pay off

posted by jeff | 7:45 AM

Tuesday, September 16, 2003

OASIS to look at Methodology  

Name of the TC: OASIS Framework for Web Services Implementation (FWSI)
Technical Committee

Statement of Purpose

The purpose of OASIS FWSI TC is to facilitate implementation of robust
Web Services by defining a practical and extensible methodology
consisting of implementation processes and common functional elements
that practitioners can adopt to create high quality Web Services systems
without re-inventing them for each implementation.

It solves the problem of the slow adoption of Web Services because of
lack of methodologies to implement Web Services, and a lack of
understanding of whether solutions proposed by vendors have the
necessary components to reliably implement an application based on Web

posted by jeff | 7:20 AM

Monday, September 15, 2003

Model Driven Services  

The model for creating business software has leveraged the concept of utilizing a base engine (or server) and extending its functionality with specialized models, templates or other consumable metadata. In J2EE, we use a JSP template engine to consume JSP's, we use an EJB container to consume EJB's, etc. Typically we then chain together engines (JSP engines + EJB engine + DB engine) to fulfill some use case.

Many of the web services that I have created were "home grown", that is to say that they were written from scratch and didn't leverage any engine. The more I looked around, the more I realized that others (including vendors) seemed to be caught in the same boat. Moving web services into the (model + engine) world is obvious.

I found myself asking how come we haven't seen more model driven services? I think one reason is that many developers are too concerned about the SOA triangle (producer, consumer, directory). The SOA triangle is an architectural pattern that will be used over and over again inside of service-based applications. However, it isn't the end. Extending the triangle, or leveraging other patterns is absolutely necessary.

So, here are some interesting questions to ponder....
1. To what extent does one attempt to standardize authoring environment interfaces?
2. Does an authoring environment service provide a default UI customizer that is shippable (JNLP style)?
3. Should the customizers (models, meta data, etc.) be logically held together at the use-case / scenario level?

posted by jeff | 6:21 AM

Sunday, September 14, 2003

IBM, Batty, Autonomic Computing  

I've been blogging how IBM is Batty on a few subjects... most recently on web services, grid (OGSA), model driven architecture (MDA)... and now autonomic computing.

The convergence of the aforementioned technologies could create one interesting model!

posted by jeff | 7:28 PM

From UML to BPEL  

Keith Mantell at IBM wrote an article on using UML and MDA concepts to gen your bpel. See:

He creates a UML profile that has the semantics to represent bpel and then discusses making a map from the model to the bpel code. The base process is represented as a class and the orchestration is modeled as an Activity Diagram.

All in all, it is an interesting concept. One concern that I have is that people don't try too hard to shove a square peg (UML) into a triangular hole (SOA & BPEL). The base UML models were developed some time back and don't always represent the concepts that we need. For instance, I use a Service-Based Sequence Diagram rather than Activity Diagrams to represent orchestrations... just seems easier.

posted by jeff | 7:23 PM