So it's been a while since the TSG meeting when the RWG was proposed [1];
not a lot of discussion has arisen naturally (not surprising, there's not a lot
for the community to build around yet).

So maybe we can review what we think will be needed.

I think two main issues were raised: [2]
* Is a WG needed?
* Isn't this covered in the MeeGo 'Core' project structure?

I think there is general agreement that Imad was correct in saying "it is larger
than a working group" - so I'd like to recap the proposal here and consider the
points above for each of the areas. The main interest areas for the RWG are,
naturally, the  repositories and arising from that, build and documentation. So
lets review the scope and relate that to the MeeGo Distribution.


Repositories
============

The proposal is that the RWG take responsibility for managing the applications
repository and the community supported packages repository; more specifically
the  associated QA, policies and processes.

The first thing to notice is that there are 2 kinds of repository areas.
a) Individual packages
b) Community non-core packages : Surrounds

The individual packages have a development/release cycle which is independent of
other packages and  independent of MeeGo itself. Indeed they may support
multiple releases of MeeGo over time (eg the GPE suite).

This part of the RWG is comparable to a community app-store and has roots in
Maemo's Extras and the Moblin garage.

As we've seen in Maemo Extras we can provide a lot of community support to this
area; but that support needs to adapt to the dynamic needs of a growing
community. I suggest that a decision making group is needed to enforce, accept,
review and agree changes to policy covering:
* QA
* package acceptance/removal
* security
* delivery
* infrastructure requirements

So is this area managed by the core project?
How can it be? MeeGo core is a distro, not an app-store or sourceforge. It has
no policies or processes for handling non-affiliated applications. There isn't
even community access to the core build system.
This is clearly an area where independent application developers are growing out
of the community and providing an ecosystem around MeeGo.

However it is equally clear that the success of this ecosystem is crucial to the
successful growth of a development community around MeeGo that clearly does not
exist yet. Both corporate and individual members of the MeeGo family should be
looking to allocate resource to this area. Yes a group is needed;
Finally, IMHO this is a technical working group, and I would be surprised to see
it within the CWG.

The second major area is the community non-core packages; this area provides a
more coherent set of inter-related or depended-upon libraries and applications.
This entire collection would follow the same release cycle (which may well be
the same as MeeGo core). The Surrounds area would also probably support
historical releases of Meego for situations where vendors didn't provide updates
for devices.

As originally proposed, it may make sense for the Surrounds area to establish
relationships (close, arms-length, formal or informal) to maintainers in other
community distros to share the burden of maintaining some packages.

So first "Isn't this covered in the MeeGo 'Core' project structure?".
No; this is, pretty much by definition, a wider and community managed set of
applications. MeeGo cannot possibly accept every library, application or
subsystem that is requested by random community members; nor do we want to
create a multitude of independent repositories or one-shot packages.

However, it does make sense to have an area where code is evaluated and may
eventually be sponsored into the core distribution. Packages which become (by
some metric) popular and which demonstrate a suitable level of maintainer-ship
whilst in the Surrounds may be promoted into the core repository.
Additionally, packages which, for whatever reason, are deprecated from the core
may continue to be supported in the Surrounds.
Similar bi-directional migration paths may apply to people as well as code!

Secondly, is a WG neede? Clearly there is a relationship between core, surrounds
and individual packages; managing them in a consistent manner seems rational.
The purpose of a working group is to provide that consistency and focus.


Build
=====
Having outlined what repositories are covered we need to look at how community
code gets into those repositories; the build service.

The build service is a crucial part of both packages and surrounds; the
selection of the OBS allows effective integration between build, QA and
publication (and potentially DVCS too). This is another key reason for the
end-2-end scoping for the RWG in the proposal.

Whilst a case can be made to limit access to the core MeeGo build system to
ensure finite compute resources are available to the core developers, it makes
no sense for the community to have independent build systems for both packages
and surrounds; a single service would cover both; this is another reason why
applications and surrounds are both within the RWG.

I would once again like to clarify: the RWG is not meant be the team managing
the day-to-day operations of any autobuilder-like infrastructure; they would
instead be providing community objectives and goals to that team.

With this in mind the RWG would be the place where the community would work with
the build service implementation team to realise the polices and processes
outlined above. Whilst it is possible to segregate the repository work described
above from the build activity it would appear to make more sense to collaborate
closely; certainly a close relationship between the Maemo community and the
autobuilder teams has been very successful.


Document
========

The community already puts a lot of effort into the documentation; the RWG would
provide a focus for areas within its scope.

Much of the RWG's work would result in additional documentation; indeed it is
quite likely that issues arising would not be considered closed until they were
documented appropriately.




I hope this provides some foundation for discussion and enables the TSG to
consider the proposal further.

David/lbt

[1]http://wiki.meego.com/Proposal_for_a_Repository_working_group
[2]http://trac.tspre.org/meetbot/meego-meeting/2010/meego-meeting.2010-03-31-19.58.log.html#l-100


-- 
"Don't worry, you'll be fine; I saw it work in a cartoon once..."





_______________________________________________
MeeGo-dev mailing list
[email protected]
http://lists.meego.com/listinfo/meego-dev

Reply via email to