g-s-t in Ubuntu

I never wanted to be part of a blog debate, especially if it contains ranting. I would have very much preferred to see some action on the spec, which is nothing more than a spec skeleton – it doesn’t mean anything at all at the moment.

I proposed this spec because in the months of GNOME maintenance in Ubuntu I noticed that:

  • normal users really do need the system tools,
  • we faced a lot of bugs,
  • I personally had problems trying to fix them, the reasons for this seemed to be to me:
    • the code has grown a lot over time,
    • Perl (for the backends) and C (for the frontends) seemed to be not to be the easiest solution for the problems g-s-t is trying to address.

I hope you don’t get me wrong. I’m grateful for all the love and work that Carlos Garnacho and everybody else put into g-s-t. It’s just that users faced more problems than we are able to fix currently.

My gut feeling was that some PyGTK scripts

  • re-using the glade files of g-s-t,
  • talking to the existing backends or using tools we use anyway

would be easier to debug and easier to maintain. Maybe I’m driven by sheer ‘start of release cycle actionism’ and wanted to be part of a nice Ubuntu Desktop Team hacking project that could gradually try to improve the situation. ‘Umbrella’ was the name for the ‘do something about the g-s-t situation’ project, Michael Vogt and I chose in a fun telephone call.

At the conference in Paris we’ll talk about the spec and our options. Please don’t make more of the spec as there currently is.