31 July, 2013

oSC13, Strategy and Factory

The openSUSE conference had some chats about strategy. Ralf talked about suggestions from SUSE in his keynote and there were more suggestions discussed in person between various people. I'd like to summarize some of the things I heard. In this blog I will talk about the directions of Factory, what Ralf mentioned as SUSE's ideas from his keynote; later this week I will blog about the openSUSE releases.

This is a combination of stuff I heard (not just at oSC but also earlier - even from the first strategy discussion, 3 years ago) and ideas I have. I just attempt to put it into text so it is easier to shoot at, comment upon, think about.

Factory

Let's start with talking about Factory.

Who is it for

We need to think about who factory is for, how these people are supposed to be using it etcetera. Right now, we have a shared understanding - but it probably isn't always as shared as it seems and it can be good to write things down and talk about them (see my keynote at oSC). Some of these thoughts will come to the mailing lists, I am sure, over the coming months.

Where should it go


On the technical side, there are some things I've heard about and seen several times and which will most likely see some form of implementation:
  • more automated testing - the openSUSE team did a openQA sprint but many agree that there is more room for improvement.
  • getting more people involved - As Ralf said, SUSE will commit to more SLE engineers on openSUSE; and the community & the openSUSE team should work more on both growing our dev pool and training people. In a GSOC discussion about this at oSC, this point was made as well: it would be good to use GSOC also for bringing existing contributors to a higher level.
  • Bringing Tumbleweed, Devel projects and Factory in alignment - As Ralf noted in his keynote, not all work on Devel projects and Tumbleweed benefits Factory and openSUSE ultimately - it would be great to unite forces and bring them all in one process!
  • segmenting Factory in rings - which Coolo has been talking about for a while now. It would create a Ring 0 with the bootstrap- and compiler infrastructure, ring 1 with the base system and other rings on top of that. He has been experimenting with this, hinting at some results in this mail. I leave it up to a later post (perhaps by/with Coolo) to be more concrete about this whole thing once the core Factory team has talked to more people and expanded upon it...

That last point is by far the least concrete - it is an idea which is discussed and toyed with. How much and what exactly will happen - only the Geeko knows ;-)

Some discussion has started on the mailing list, by the way, with the idea of segmenting Factory in even more 'components' than just the rings. I had a chat with Coolo before this proposal came out and I suggested the same thing. He noted that it would lead to a huge mess of dependency issues, including many circular ones (he gave libpoppler as example). This is already brought into the discussion there, let's see where this goes!