On Wednesday I went to the JISC dev8d conference. I wish I could have gone for both days but time doesn’t permit at the moment. In all, I had a trhough provoking day and managed to catch talks on the Mobile Web (which I wasn’t expecting) and Linked Data. Whilst I didn’t attend the programming workshops on languages such as Clojure or Erlang (at the moment I don’t have a need to use either), I was looking for matters that might be useful for my impending move to Janet. (This is one of the reasons why I haven’t posted recently – I was either preparing or convinced I hadn’t got the job.)
I bumped into Eamonn Neylon and we went along to the Mobile Web session with Mike Jones from Bristol and the Molly project. They outlined the two main approaches (either via the various app markets or having a front end which caters for the different phones) and issues such as being sandboxed from the hardware layer at the moment. It would seems from them that you need to do both ideally, though development time doesn’t always allow. The session was slightly hijacked by the Python 2 versus 3 question and if Molly would ever use Python 3 but we gradually got back on track. The main barrier to entry would be the lack of standardisation so you need to target the platforms as well as the hardware issues.
I stayed for the Linked Data session which Chris Gutteridge took sort of control from the array of speakers. The main focus became notions of openness (as defined in the Open Knowledge Definition) and how it is perceived by the academic community bringing back up the ideas of attribution on the web. (An issue which is partly cultural.) The issue of clear licencing came up again as well but there does seems to be some clarification needed on the different models.
I did go to some of the lightning talks before lunch but they didn’t leave much of an impression this time (though Chris Gutteridge did plug his Q &D RDF Browser which I’m thinking of using). After an excellent lunch, I wandered into basecamp where I plugged in my laptop and worked on Open Correspondence whilst waiting for the session on the Linked Data API. I did spend a couple of hours work on it to fix some bugs and little things for the next version to go live (though discovered another one in places with some missing but it is not huge, just needs a couple of hours). Rufus Pollock and Jo Walsh popped by so we managed to catch up and do some hacking. Rufus suggested using Flask which I think I’ll use for some smaller projects in the future (and for some reason Backbone was mentioned but not sure how).
I went along to Chris Dollin‘s talk on his eLDA library and how the Linked Data API works. It seems like an eminently sensible solution to removing the complexity of semantic technologies from the user and to make it easier to use. It is certainly something which will be useful to get my head around completely.
The day, as suggested by Mahendra Mahey, was definitely more useful when doing something and just cracking on with it. We need more days like this as it provides a collegiate atmosphere to try new things and take a look at different technologies which might not appear on the normal radar. The friendly atmosphere was great as well. I’ll book both days off if it comes around next year.