Monthly Archives: August 2006

Why We Need SO Tooling

An interesting discussion is underway at the Yahoo SOA Forum about the relevance of object oriented (OO) tools to service orientation (SO).
This kind of discussion illustrates why we sponsor the SOA Tools Platform Project at Eclipse, working together with IBM, Sybase, and others to create tooling better suited for SOA than tools based on UML and MDA.
One of the participants in the discussion, Stefan Tilkov, and I have discussed this previously in earlier blog posts (ref1 and ref2). And one of the forum participants, Lukas Barton, referenced a thesis he wrote about MDA in his post, including some perspective about the disconnect between OO and SO with respect to tooling.
Some things have changed since the original discussion, but as far as I can tell UML and MDA still have not been adapted away from object orientation toward service orientation. It still looks to me as if the UML and MDA based tools want you to derive your WSDLs, XML Schemas, and SOAP messages from objects rather than the other way around.
Certainly you can use UML to create services – as has been mentioned many times you can really use any technology, including procedure oriented technologies such as COBOL or PL/I or stored procedures, or asynchronous messaging systems. And SO is abstract enough to encompass them all, which is one of its main benefits. You lose some of that when you tie SO too closely to OO.
For IT software, OO concepts can often create more problems than they solve. (Implementing services using objects, procedures, queues, etc. is ok but let’s not design them that way.) It is much better to model, design, and develop using services natively since they map more cleanly to business functions. If you have to design and model your functions first as things (i.e. objects) you are introducing unnecessary complexities, and often the industry ends up trying to resolve unnatural and impossible mappings such as OO-Relational and OO-XML (impossible to do automatically or transparently I mean).
Software tools need to improve the level of abstraction. MDA and UML do not really do this since they expect us to first learn how to reinterpret the world of functions into objects with methods and only then think more abstractly about services. What we need is a good set of contract-first SO tools specifically designed for SOA that natively support the service abstraction, whether the implementation is OO or not.

Advertisements

Wikipedia Discussion while Visiting W3C

Wikimania is this weekend and by chance I happened to meet Joseph Reagle on the way back from lunch yesterday with Philippe LeHegaret and Steve Bratt. Joseph used to work at W3C with Steve and Philippe and they were kind enough to introduce me.
On the way to D.C. Tuesday I had read the very interesting article in the New Yorker about Wikipedia. True to form, I suppose, the existence of this article has already been noted on the Wikipedia entry about The New Yorker.
The article made it sound as if Wikipedia was turning into a forum for endless arguments about who is right. The article noted the proportional increase in background chat about entry content compared to actual entry editing activity. Having worked in the technical community for longer than I like to remember, it is all too easy for me to believe this.
The main theme of the article, however, was about the potential for Wikipedia to replace the Encyclopaedia Britannica (or any other traditional encyclopaedia for that matter). The Britannica only has 120,000 entries while Wikipedia recently surpassed 1,00,000. One of the biggest question marks is how authoritative and objective can Wikipedia be if anyone can edit an entry, as opposed to selected expert or experts.
Joesph’s Wikimania talk is related to this point, and he seems to believe that Wikipedia is for the most part reliable. I don’t think he or Steve thought my characterization of the New Yorker article was particular apt – this would be I guess a little like saying that the whole thing is bad because there have been some problems, and criticizing the thing itself because of the discussion it creates.
The morning downpour had cleared and we were standing in front of the main entrance to the recently opened Stata Center, where the W3C now has its offices. It was pretty cool seeing it, and I wished I’d brought my camera. But of course there are already some photos online that are better than anything I could have captured just walking to and from.
I was there to continue the discussion on the “Web of Services,” and what, if anything, the W3C might be able to do to make it as successful as the “Web of Documents” – e.g. Wikipedia and the Stata Center photos.
It was a good meeting, and hopefully there will be some news very soon. Of course, any and all comments and suggestions are very welcome.

12:01 Blues Band

Jane and I went to see the 12:01 Blues Band at the Acton Jazz Cafe last Saturday night.
It was a great night out. Both cafe and band were very enjoyable.
As I mentioned before, Jane and I lived in Chicago just after college (1978-1982, can it really be that long ago?!?) One of the great things about living there then, and living where we did (in the DePaul neighborhood on the near North side), was that we could go see world class blues musicians pretty much any night of the week.
So it was great to see our neighbors from Swampscott, Mass. singing and playing the blues the other night in a nearby venue. If you get a chance, I would definitely check them out. They can all play and sing too, and they offer a great mix of traditional blues and original tunes.