Last week I attended QCon in London and have been mulling over the experience. Two themes stood out for me, though bear in mind that having recently attended Adobe and Microsoft events, I avoided the sessions which related to those companies and focused on Java and more general issues.
The first was the extent to which QCon accepted that mistakes made in the early years of J2EE have been costly for the industry. In this respect the key speech was from Spring inventor Rod Johnson, whose session was called The Cathedral, The Bazaar and the Commissar, “an opinionated view of the story of Enterprise Java innovation.” Johnson stated that the rush to standardize everything in the Java platform had done almost as much harm as good (though he did say it was a “net positive”). He identified numerous problems in the standardization process. In particular:
- The expert groups are dominated by large vendors whose interests are not aligned with those of developers. Johnson gave the example of JDO. “What happened was, this was intensely political. Database vendors hated JDO. Application server vendors hated JDO. Because in each case it wasn’t good for their business. And the result was that JDO was indeed taken out and shot.” Johnson also noted that J2EE serves to protect the existing licensees:
“It pretty much ensures that there will be no competition, because who at this point would want to implement the whole enchilada of J2EE?”
- The JCP has a history of ignoring prior art, which means on occasion it comes up with specifications that duplicate the functionality of existing products, or which are inferior to what is already available. In consequence good technology that was outside the JCP was lost. He cited O/R mapping, persistence and logging as examples. Here’s Johnson on Toplink:
“Hibernate is the de-facto standard for O/R mapping today. It’s a pretty good product. How many things does it do that TopLink didn’t do in 1998? The answer is not many. There’s probably just as many things that TopLink did in 1998 that Hibernate doesn’t do now. Toplink wasn’t as robust back in those days. But … what we’ve done is somehow managed to kill all innovation for ten years in O/R mapping, while in the meantime Microsoft has both figured out that O/R mapping is interesting and with new technologies like LINQ they’ve actually started to push the envelope. They shouldn’t be relevant in this space because there was such a lead in terms of Java technologies.”
- The JCP standardized specifications that did not work, in particular Entity Java Beans. The industry accepted them because there was an obsession with standards. “There was a complete transformation where non-standard technologies could no longer compete and competition was no longer on merit.”
Johnson asked what seems to me to be a key question: what should be standardized? He said that it is silly to try both to innovate and to standardize at the same time, because the committee will get it wrong. You should standardize in areas that are well known, understood, and proven in the market.
Despite appearances, Johnson is not an enemy of the JCP. He spoke warmly of the current chairman, Patrick Curran, who is trying to reform the organization; and feels that real progress is being made. Curran was also at QCon seeking opinions on the JCP and its future.
Johnson also feels that Java has moved on. “The Java world is no longer a one-party state,” he said.
What I found interesting was that everyone at QCon seemed to agree with what Johnson said. Even Curran did not try to dispute it. In the closing panel, Martin Fowler from Thoughtworks introduced Johnson as “the man who single-handedly destroyed EJB”, and was applauded.
I suspect there is a Microsoft factor in all this. In the early days of Java, there was huge fear that Microsoft would corrupt Java and make it a Windows-only technology. Standards would prevent this happening; hence the standards religion that Johnson claims was not thought through.
There is an interesting underlying issue here, which the extent to which our industry meekly follows the lead of influential gurus and big vendors in the solutions which it implements.
That brings me to the second major theme which I noticed at QCon: the rise of REST. Despite sincere attempts at balance, there was an unmistakable tendency to promote REST at the expense of other stuff including WSDL and SOA. In many ways this makes sense to me, but it would be rash to assume that this year’s fashion is immune from the issues that now beset fashions from previous years. One delegate asked what stuck me an an excellent question in the closing session:
“What if you came along last year and saw the rise of SOA, and perhaps went off and did a whole load of SOA, and came back this year and found, SOA is on its way out, I should have done loads of REST stuff – how are we supposed to recognize which are the good horses to back?”
It is this kind of fear that leads many senior IT folk to be conservative in their technology and vendor choices.