Eclipse forum Europe 2009
May 23, 2009This year’s eclipse forum at Mainz was in cooperation with JAX and SOA. So that sounds great right? If you would purchase the main conference ticket you were allowed to enter all three of them!!! The catch here is that only the Eclipse forum was (almost) entirely in English. JAX was mainly in German and SOA, well I do not know, I was not that interested in it. I was quite surprised because nowhere in the website was mentioned that it would be in German: “3-in-1 Conference Combo: Register for the Eclipse Forum Europe and enjoy the JAX and SOACON conference as well!”. I started getting suspicious only because on the timetable most titles and descriptions were in German… Unfortunately the language barrier limited my options so I had to stick with the Eclipse talks which proved to be quite interesting.
Day 1
The first day started with a talk about the Eclipse Modeling Project (EMP), a top eclipse project that facilitates model driven software engineering. It contains lots of interesting tools for code generation, validation, model querying, database mapping, concurrent access to models etc. Also the EMFT, a project in its incubation stage that aims to extend/complement EMF (the main component of EMP) provides additional functionality such as a query tool more UI oriented, support for the .NET platform, a tool model’s comparisons and a plug-in that enhances eclipse for working with code generation.
Later that day I attended two very fun talks from Ted Neward: Busy developers guide to Scala and the same for ECMAscript. Both talks seemed suitable for developers of none or some knowledge about the particular languages. Scala was definitely one of the hot topics of this conference.
That same afternoon the keynote was ‘Architecting your way through recession: an open source survival kit’. The talk was about recession and open source. As long as I was there, it was mainly about common knowledge stuff with a touch of marketing (the speaker was from one of the conference’s gold sponsors Liferay).
The next talk I visited was called ‘Fresh Ideas for UI – Interaction design in Eclipse’. I went there with a colleague of mine that is really interested in new ideas about UI design, but none of us found them really fresh. We definitely saw some very pretty applications but were mainly comprised of old ideas put nicely together. The presenter was a psychologist that seemed very experienced on user-computer interaction. Even though he gave some good generic tips about UI design, his lack of technical knowledge didn’t help people like my colleague that were interested in eclipse specific information.
During the evening break they were offering beers. When I attended the following talk I found out why.
It was a great keynote by Neal Ford with the strange name ‘Ancient Philosophers & Blowhard Jamborees’. After the talk two phrases got stuck in my mind, the first of which is accidental complexity. It is the tendency of introducing extra complexity on projects and then later having to deal with it. And some of the reasons that he stated sound a bit too familiar: Manager Boards that comprise of people with none or little technical knowledge, meetings that end with the conclusion to reschedule a meeting and software/hardware that has to be used because the company just purchased it at some point, even though it introduces additional complexity. And what he thinks this will lead to if we do not become more aware and active about it? The transfer of most software development to Asia and in particular Chindia. And there is exactly where my second favorite phrase fits shift happens.
Now you understand about the beers…
Day 2
The second day started with a morning talk about ‘what s new in BIRT 2.3‘. Some of the new stuff is the introduction of a JavaScript debugger and improvements on charts and reports. Also two new nice aspects are that each report can now return multiple results sets and that preferences can defer among BIRT projects and not only among workspaces.
Next talk that I was interested in was ‘BIRT within Java Enterprise’. To my disappointment the speaker announced that the presentation was in German because he only found out that it should have been in English the day before… I rushed to a talk about Xtext, a framework to develop external DSLs. The main idea behind it is that one can define a language grammar and then Xtext is responsible generating the parser, the editor etc.
The key note was unfortunately in German. You can find more details about it at my x-colleagues review who happens to be a German speaker. So for me it was time for a break. But what can you do during a break when there is no coffee (they served coffee few times a day, which was disappearing quite fast) and more importantly no Internet? (the coverage was so bad that you could rarely connect…) Oh well at least I could queue early for the lunch which by the way was good and in general the whole conference was very well catered.
Another talk that I attended on that day was ‘Domain-Specific Languages in Groovy’. Also very interesting, DSLs seem to be quite in at the moment. After yet another keynote in German there was an interesting talk by Wayne Beaton where he presented a simple application that runs on a desktop using Eclipse Rich Client Platform (RCP), on a server using Rich Ajax Platform (RAP), and as an embedded application using embedded Rich Client Platform eRCP (see EBERT). Quite tricky actually, since not all platforms have all libraries available, so especially for eRCP the options are quite limited.
Day 3
Thursday started with a talk about user friendly Eclipse applications. They gave some useful tips about what the user wants and they showed that it is pretty straightforward to create cheat sheets, help, info bubbles etc. Following to that, Wayne Beaton presented Eclipse support on top down development with some nice examples. The bottom line would be ‘treat your tests as first class citizens‘.
This afternoon did not differ much from the previous one: a keynote in German, no Internet and me together with few more none German speakers queuing early for lunch.
After lunch I attended a talk by Martin Lipper about general tips on OSGI. A point he made there was that in order to reduce coupling it is preferable to import bundles rather than require them. Also another good point was about extension points; one should not misuse them. For instance if we are not sure who wants to use some particular functionality we should better expose it as a service. Next talk was about the use of OSGI for dynamic application. A demo and discussion can be found at Kai’s blog
Finally the day and conference closed for me with a talk about Equinox p2, the latest provisioning mechanism. Really cool stuff!
The slides from Monday and Friday workshops can be found at:
The other slides will not be online but will be posted to attendee’s house individually (which I haven’t received yet). Hmm…
In case you are interested in more information about one of the talks please let me know. I have some notes and I was there with few more colleagues/friends that attended different ones.