5th July 2013, Lancaster, UK
By Barry Rowlingson SSI Fellow and Research Fellow, Faculty of Health and Medicine, Lancaster University
Getting thirty people from a mixed set of departments and faculties was the highlight.
After a fairly low-key advertising campaign I was pleased to see thirty researchers attend this inaugural meeting of the Research Computing Users Group at Lancaster. I gave a short presentation on the ideas I had for the group, and mentioned the work of the SSI and Software Carpentry. The following discussion included some great ideas including running a mini-conference for research computing presentations later in the year.
After that I gave the floor to Jamshed Anwar, a new professor of computational chemistry at Lancaster, who outlined his plan for a cross-faculty computational modelling group, which would act as an umbrella across people in research from Physics, Chemistry, Environmental Science, Statistics and elsewhere. Software development and sharing of good practice would be one of the outcomes of this group, as well as serendipitous cross-disciplinary research opportunities.
Finally I gave a short, almost lighting talk on reproducible research with R using the knitr system. This for some of the attendees was their first introduction to writing data-driven reports and the light-bulbs could be seen going on above their heads. I also mentioned the use of other programming languages within knitr as I knew there were some python programmers present. The IPython Notebook system also had a place in the discussion.
This talk was meant as an exemplary style of presentation for further group meetings, which I'd like to hold once per term with three or four presenters.
During lunch the conversation included people very interested in the reproducible research process, some calls for Software Carpentry at Lancaster, and more discussion about the research computing mini-conference.
I prepared a short questionnaire for the event to get an idea of people's skill sets and ideas. This only had 19 responses, but still showed some people admitting to advanced skills in programming, revision control and
graphics, and others knowing nothing about them. Clearly there is a lot to do.
Some of the suggestions on the questionnaire included sharing best practice, building community across disciplines, pooling skills, learning 'nifty' tricks, and so on - all good things that the SSI supports.
Overall the feeling I got, and the feeling I get from a lot of people, is that of needing to know what is out there. Software tools and techniques develop so fast that its easy to miss something useful, and unless you have a rich community with interests that give you a diverse set of channels into the ecosystem of research development then that will continue to happen. Setting up a local research computing user group should provide and promote that culture.