[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: LedgerSMB companion projects (was: Light manufacturing modules....)



Chris Travers wrote:
<SNIP>
This certainly has come up in the past.  I am not sure we have a
complete idea of how this is going to work yet.  I think we want to
get more of our own infrastructure developed and tested before opening
it up to other sub-projects, however.  I suspect that the next step
will probably to move the public email lists onto our own servers.
I do not see the benefits of moving list in this manner, it creates a reliance on certain members of the core team. If this is to continue to be a fully open development project these types of reliances should be at a minimum. Google groups or the current sourceforge lists (I now SF has some issues) should be sufficient.

And we need to get our community CMS better developed.  And we still
don't have a public demo up yet.
Do you just mean more content or a new CMS?

I think companion projects like Light Manufacturing Modules would be very
well served to also run in the (or a) LedgerSMB Trac instance, for the
same reasons of attracting developers, testers, and users to particpate in
the sub project.

Personally, I think that a community architecture is going to require
more than just throwing Trac at the problem.  Especially from the
viewpoint of attracting users, I see a lot of needs we need to look at
(for example, online demos, etc).  I think we need an idea of how
feasable it is to meet all these needs before deciding on a solution.
I would think that as long as a module's functionality has any possibility
of being integrated with, or some code moved to, the LedgerSMB core, then
the code base benefits by keeping that development process close to home.
>
Not quite sure I follow you.  I think it is important for the core
team to be available for input regardless of where the code is hosted,
and I suspect that this is more important than whether it is hosted by
us, by Google, or by Sourceforge.
I agree with some of the comments above, we may be getting too excited about tools rather than defining what we need.

If the APIs etc are defined well enough then external products can be developed discretely. For an example check out a CMS I do a lot of work with http://plone.org/products/ hundreds of products developed independant of core. I would prefer a closer build relationship with developers rather than creating the wheel each time but reliance on the core developers means less core product development?

If it can be managed without impacting the core developers' operations, it
would be very nice to have subprojects under a separate path within the
same repository. Subversion seems to be able to use different authorization
lists at different paths in the repository, and a Trac instance can
apparently be rooted at various points in the repository.

I can see positives and negatives to this setup.  I personally think
we need to figure out first what the core project needs (and optimize
this), then we need to find out what companion projects need.  And we
need to bear in mind what our available infrastructure is.  I also
think that those positives and negatives are less important than the
importance in having a core team that is available to help companion
projects minimize foreward-compatibility issues.
Could we look at introducing teams to work on certain aspects. At the moment I see us relying on the core team for technical development, system maintenance (offerings such as http://code.google.com/ could offload some of this burden), direction and vision, training support.. all the stuff a project like this needs.

Is it a viable option to break out into key areas such as marking, technical, infrastructure, documentation etc. There will be overlaps in resources but this could allow a buinessperson with a marketing background, or an accounts person who likes to write documentation a great opportunity to get involved.

Just my 2c worth.

Just to add, I have not yet migrated over to LSMB but willing to provide assistance where I can. I am not a Perl developer but have fiddled with LaTeX templates and can write good england sometimes. Happy to work with technical people and translate into normal speak.

W