erik doernenburg

As you might know one of my main interests at the moment is software visualisation, and the success of the talk that Gregor and I are giving at conferences at the moment shows that there’s something to it; I guess.

So far, we’ve mostly been using ad-hoc tools, made up from small building blocks such as GraphViz Dot and custom scripts to munge file formats. This can get you quite far, but how much further can you get with real tools?

With this question in mind, I visited Michele Lanza at the Univerity of Lugano when I was in Switzerland last week. I had been playing with his CodeCrawler for a while, especially since it started working with Java codebases courtesy of java2cdif, but I was curious what Michele and his PhD students are working on now.

The individual projects are a clear evolution of the ideas behind CodeCrawler, making use of interaction and filtering to cope with large amounts of data, and some of them add time as another dimension, which is interesting because I had also found identifying trends one of the most powerful uses of visualisation on several projects.

If you are interested in this field, have a look at Michele’s site, all the tools and related papers are linked from there.

I enjoyed last years Symposium, both the one in Las Vegas and the one in Barcelona, quite a bit and will be back at the European one in Barcelona this year. (Had to skip Las Vegas due to timing issues this year.) Confirmed talks are the Visualisation talk Gregor Hohpe and I have given several times now, and a case study about the building of the new guardian.co.uk website. I am extremely pleased to co-present this with Matt Wall from the Guardian Unlimited.