Debian Science - umbrella for scientific packages or dustbin for scientific code?

Speaker: Andreas Tille

Track: Debian Science

Type: Talk (45 mins)

Video:

Room: Rex Live Stream

Time: Aug 06 (Sun), 12:00

Duration: 0:45

The Debian Science mailing list was created in 2005 and in 2008 the Blends framework was implemented to turn some random discussion into a structurised team with dedicated coders and packages. Later other existing packaging teams maintaining scientific software were merged and since about 2013 the Debian Science Blend is covering most of the scientific packages inside Debian which do not have any specific Blend team (like DebiChem, Debian Astro, Debian Gis and Debian Med). It tries to serve as entry point for scientists with an interest to package some software for Debian or seeking help with some scientific software. The original idea of the quite general Debian Science was that it could serve as an umbrella as long as there is not yet sufficient manpower inside Debian to be able to run an own Blend for a certain topic. This has worked out for Debian Astro which is a pure offspring of Debian Science - so the concept has at least one example that it can work.

However, there are also things regularly criticised: Debian Science does a bad job to pick up packages where Uploaders vanished from the team without notice and several packages are in a bad state. Admittedly this is the case to some extend. It is caused by the fact that scientists tend to move to different projects or institutions and do not find the time to sort out their code heritage. While this is effect is neither new nor unknown inside Debian this kind team orphaned packages sometimes shades a bad light onto the team. While there are some team members who are doing a great job in picking up such orphans forming kind of Debian Science internal QA team not all cases are covered. At least the good news is that it is very easy to take over an orphaned package by team uploads or even adding an ID as Uploader.

The audience of the talk are all people interested in scientific software. I try to present ideas how we could strengthen the “umbrella side” even more over the “dustbin side” by concentrating more dedicated people for certain topics. Ideas how to reach out further into scientific groups to gather more team members are welcome.

URLS