Home  |   About  |   Energy  |   Politics  |   Software  |   Music

26 July 2012

Loose notes from a conference

Earlier this month I went to an Environmental Modelling conference in Leipzig to present an article on my current professional endeavours. On any field, conferences are mostly important for the networking you build, meeting people from literally all continents and finding new projects to work on. This particular conference was no exception on that regard, but apart from that, there were some interesting points in terms of software worthy of note. Below the fold is a loose collection of thoughts I brought back.

Conferences are logistically heavy events and for that reason often require sponsors. Naturally these sponsors try to push their products by showcasing the sort of research you can do with them. This is all acceptable, up to a point, but in this particular conference this point was clearly crossed in certain sessions. Some address were pure marketing, showing how good a certain closed expensive product is. In another flavour there was a new independent product for which the presenter asked the audience for a good licensing price. And another genre still was of those "open source" products that are totally dependent on commercial software, in some cases doing things for which fully open alternatives exist. This last flavour is more akin to science, but still a dicey terrain to be in. Noticeable was the fact that all this sort content came from the US; at some moments this gave the impression that some US universities are nothing but mere marketing agents for US software companies.

After one of these eerie sessions I ended up talking with a few younger students. One of them put it up in a very simple manner: "this looks so old school". Indeed, were this conference a full sample of the research in the US, it would clearly show a multi year lag from Europe. In reality there are many (and good) open source projects in the US and likewise research institutions in Europe gladly mastered by software multinationals. But it is interesting how doing research based on closed source is gaining such connotation. It is more and more apparent that releasing your code and making sure your products are portable and interoperable is not just a matter of scientific scientific necessity, but more than that, essential requirements to the success of your work.

The most noticeable trend to bring from the conference is a clear shift towards service based architectures. Whereas researches have been traditionally used to work with models enclosed in expensive software packages, more and more folk are showing up with projects that bring up these models to the web as services. Suddenly, an incredibly large number of models, ready to run and cost free (both in licensing and processing time) are becoming available. Software projects that catalogue, integrate and chain these loose services are taking shape, transforming modelling into a rather different activity. The symbiosis of Taverna with BioCatalogue is a good example of this process, that will likely propagate to other fields of Science. The Web Processing Service standard, proposed by the Open Geospatial Consortium, is a cornerstone of this transition into distributed processing, that is quickly going beyond its initial scope. It is a totally new reality in Science that is emerging, reinforcing the demise of licensed software.

Finally, this conference provided a proper test bed for an old issue with the battery of my corporate laptop. After some tweaking I finally got the Linux 3.2 kernel using the energy saving features of the sandy bridge processor in full. Sticking to reading and taking notes the battery can last some impressive 8 hours. Turning on the WiFi board and doing some browsing and e-mail reading gets this down to 6 hours. Long enough to last throughout the whole day of conference; for the first time I came back from a conference without a single sheet of paper written. This experiment also allowed me to understand a fundamental difference between my two favourite browsers: Firefox and Chromium. The numbers I provided earlier are only possible with the former; whenever, by mistake, I launched Chromium, the fan jumped to max and soon the laptop was burning my legs. This was another thing to learn from the conference: Chromium is an energy hog.

Apart from all this, I really recommend the city of Leipzig, a place where I wouldn't mind to live.