[kde-edu]: KEduca resurrection (possible GSoC project)

[Matt Williams]

On 28 March 2010 13:10, Mauricio Piacentini <piacentini at kde.org> wrote:
> After that Matt Williams did some work on KEduca, but as his plate is also
> more than full (between PhD, lots of projects and being the KDE Games module
> coordinator) the effort got stalled again.

I think one of the problems I had (apart from real life catching up)
was being too ambitious with it. I tried to simultaneously port
(rewrite) it for KDE4, make a fancy new QGraphicsView UI and also
change the file format to be compatible with the IMS QTI standard [1].
This all proved to be too much and it was taking far too long to get
anywhere before anything even basic worked.

While supporting a new file type might seem like a big thing I decided
it was worth doing since it would provide interoperability with other
software such as Moodle etc. However, I had trouble narrowing down
decent use cases for the software and so I think I was trying to make
it do too much. I think KEduca really needs some well-defined goals,
use cases and expected work-flows as well as reasons why it is
preferred over other software such as Moodle. We need to know where it
fits in the software landscape. Without this, it is impossible to be
able to plan which features should be added.

> KEduca ships by default in the KDE 3.5.9 based Linux Educacional distro,
> delivered to thousands of schools by the Brazilian government.
> (http://piacentini.livejournal.com/7871.html). So its user base is actually
> increasing, which is a problem as these people will get orphaned soon (next
> year) when Linux Educacional finally makes the switch to KDE 4.x.

For this reason, a raw port alone must be achieved somehow.

> I realized that we will possibly loose KEduca for good if we do not find an
> alternative way of supporting the port. This is why I considered applying
> for a GSoC slot as a student to do it. Several people tried to work on it on
> their spare time. but the results are not good after several years. Working
> in KEduca is not "sexy". The 3.x code base is, well, not easy to navigate
> around. Matt (with the advice of Pino) decided to rewrite it back in 2007,
> but this is an effort that takes a lot of free time, which is something most
> of us do not have, considering our current projects. And, more importantly,
> I think we need someone that could not only do the port as a quick summer
> job, but really take over the application and become its maintainer for at
> least two release cycles. This is involves an attempt to maintain
> compatiblity (write a converter) for older tests, and follow through with
> everything necessary to make sure the application gets included in KDE 4.6.
> So it is something that would start as a GSoC project (and with the initial
> monetary help that comes with it), but needs to go beyond that, with at
> least a 18-24 month commitment in maintaining it.

Give the experience I had, I think that finding someone to do an
almost direct port to KDE4 without adding any new features would be
the best way. It's a simple enough codebase that a full direct port is
well within the scope of GSoC with perhaps even potential for much
cleanup and code simplification.

Once the port is done, then work could start on adding fancy new features.

> At the same time, I am currently working on my Master thesis at PUC-SP,
> related to educational open source software, its social and cognitive
> impact, which limits the time I have for open source development related to,
> for example, KDE Games. But it is possible that this work on KEduca could be
> featured a part of my thesis, so by doing this work as part of GSoC I could
> possibly tie several loose ends:
>
> - the need to port KEduca
> - get Google to support the initial part of this effort (pay for it), as
> voluntary work on this project has failed a couple of times
> - reuse research work done as part of my thesis in the project
>
> All in all it seems like a good thing to do, but for this I need to know
> what you people think about it, and (very important!), if one of you is
> willing to do the mentoring (Matt, Anne, Albert, Pino, Aleix, ..?) I also
> want to make sure the GSoC money is used for something that will be
> long-lived: there is a mix of applications between those that attract
> newcomers (which sadly go away as soon as the project is complete), those
> that manage to attract new and committed contributors (great!), and some
> that are used to support the work of existing community members, like this
> one.

I am willing to help be a mentor for this but I am not sure I have the
time to do it by myself. If someone would be willing to co-mentor with
me then I will be able to help.

> So, opinions on this? Should I post the idea and submit a proposal for it?
> Or is it a waste of a good project slot :) ?

I think it's certainly a valid project and given the real impact and
use it gets, we should get some interest. We just need to work out how
to advertise its sexiness!

-- 
Matt Williams
http://milliams.com

[1] http://www.imsglobal.org/question/