JavaPolis 2006

Actually, the talks might have been much better than they were last year, but I would not know, because we hardly went to any. On the other hand, there were certainly a lot more people than last year; around 2800 attendees made it the biggest JavaPolis yet, pushing the limit of the venue’s space. That is what made the conference so good for us, because it was a great opportunity to swap dozens of blue Lunatech cards for an international collection from diverse and fascinating people. Hugh MacLeod explained this pretty well after going to some other conference:

‘I don’t go to these shows so I can sit in an auditorium and listen to folk speaking for hours on end. I’m lucky if I average two hours per day. So what if the schedule changed this time, that’s not why we shell out the money to attend. I go to these events to meet and hang out with people like Sifry, Mayfield and Weinberger, over a cup of coffee or a beer behind the scenes. I got there to commune with my professional tribe. I go there because I like and believe in the people organizing the event. I go there because I like and believe in the other people attending. The stuff in the auditorium is just the hub, as far as I’m concerned. The real action is in the spokes. The real action is in the corridor conversations.’

JavaPolis was just like that: the most interesting people to talk to were presenters, people working in the same area as us geographically or technically, and the organisers. In particular, talking to Robin Mulkers (JavaPolis steering committee) really got me thinking about the real-world community aspects of Java software development, which includes conferences like JavaPolis, user groups, recruitment, and generally going out for beers with other Java geeks. Needless to say, Antwerp is even better equipped than Rotterdam for drinking beer.

The most interesting presentation I went to was A Software Framework for Human Interactions, where Keith Harrison-Broninski explained human interaction management, and presented his humanedj software, which supports this. I liked the presentation because the idea was so new for me, and also because Keith’s presentation skills are excellent. This was a welcome contrast to one of the presentations that I left after ten seconds because the speaker was so terrible, even though his software is great. The humanedj software, however, was harder to judge from what we saw in the presentation; I plan to try it, but probably not soon, because I do not think it is applicable to Lunatech’s environment.

The most interesting software we saw presented was Alfresco, an enterprise content management stack that aims to provide an integrated open-source alternative to Microsoft SharePoint. It is always good to see open-source software move from application to platform/stack in another software area, and we are certainly curious to know whether this software is ready to replace the functional but clunky commercial enterprise search software that we have used on external customer projects.

I also am hoping to find time to check out Alfresco’s web content management capabilities, because it would be useful to find a Java-based web content stack that we like, so we can develop a capability to build simple content-based web sites quickly. We are currently too expensive for this, because we have only done that kind of thing a few times, and we focus more on custom software development, so there are other people who can do it faster and more cheaply. The catch for our would-be customers, is that the average web site builder implements the web site in PHP. This is fine for the initial web site, but if there is follow-on customisation work, the code is likely to get worse and worse until a year later, the developers become too bogged-down to continue making progress, and the whole thing ends up getting thrown out.

The most interesting vendor in the exhibition hall was Atlassian, whose excellent JIRA and Confluence we use. JIRA is perhaps the most impressive of the two products, because it has so little competition, which is why it was more interesting to talk about Confluence, and how it compares to other wikis and other software that you can use for documentation. For example, I had previously talked to Ludovic Dubost about his XWiki project. XWiki is a capable open-source wiki, with interesting features for structured data, and easy scripting in-line in wiki pages. It is not as pretty as Confluence, which matters for daily use, but Ludovic’s vision of wiki as CMS/database platform is certainly intriguing, and its structured data support would let you solve simple problems, like expense forms, in the wiki. Atlassian do not have any immediate plans to allow spreadsheets to be pages in Confluence, but there is certainly interest in at least simple structured data functionality: Mike Cannon-Brookes told me that the Confluence Metadata Plugin is ‘huge’.

All of these presenters and companies missed one thing, though; the most fun place to be was the JBoss stand, where they were serving Jupiler. Some people preferred the Leffe Blonde that BEA were serving, but it is too sweet and strong to drink a large amount of, especially if you start at 4 pm - you are better off with the lighter and more refreshing Jupiler. Adobe, meanwhile, proved that Java geeks are more interested in free beer than talking to pretty girls, because their big stand was deserted. Of course, that could be because Flex is not cool enough yet. (Hint to Adobe: try open-sourcing it.)

So despite all of the talks, software and technical discussions about things like closures in Java, we ended up spending a lot of time socialising with various members of the JBoss team, which basically involved beer and conversation with a bunch of cool people.

All in all, a great conference. I plan to catch up on some of the talks I missed by watching them on Parleys. For beers at the Kulminator (Vleminckveld 32), which I also missed, I shall have to go back to Antwerp.