ulc-devs team mailing list archive

[Frederik Elwert <frederik.elwert@xxxxxx>]

Hello everybody,

since our promising start, not really much has happened, at least in
terms of visible results and code. Since it would be sad to loose the
initial momentum entirely, I just wanted to document what has been going
on lately, and outline some possible future directions.

Since we decided for Python, Patrick and I created a basic lens as a
quickly project. The second step was to access Folks to get the actual
contact data. Since obviously no one has ever tried writing a Folks
client in Python before, we faced some hurdles. Thanks to the quick
reaction of the telepathy team, libfolks 0.6.6 contains some major fixes
regarding GObject Introspection that allow us to access it from Python.

I tried to package that new release for Oneiric, with some success, but
it turned out that the dependencies were quite heavy, and I experienced
some issues of the new libfolks with Oneiric’s empathy. Instead of
rebuilding everything telepathy related, Patrick and I switched to
Precise for now, since it already ships the new libfolks.

However, when trying to access the actual contact data, we faced even
more issues, which may be anywhere in libfolks or libgee (which, in
turn, libfolks uses for its basic data structures).

While pioneering in this area of PyGI support might be an honourable
task which will serve generations of GNOME/Python developers to come, it
makes development incredibly slow. For every tiny step we rely on the
library developers to fix issues with their GObject Introspection
bindings. Since these issues are largely due to our decision for Python,
and I know that some of you were in favour of Vala, it seems unfair to
hinder the development in this way.

So I just wanted to bring up the possibility of re-thinking our
decision. I for my part have no experience with vala, and I know that
his also holds true for Patrick. But if some of you would volunteer in
writing the lens in Vala (possibly using Pablo’s initial work as a
basis), feel free to pick up the task. I could still contribute to
design discussions and the like, but probably not to the code.

Lately, I also stumbled upon another attempt to write a contact lens:
http://iloveubuntu.net/unity-contacts-lens-brings-ubuntu-one-and-google-contacts-dash
I could not find a public code repository, but it seems to choose a
different path than we did. The lens contains scopes for Ubuntu One and
Google contacts, thus it aggregates different sources as scopes, instead
of using Folks as a unified interface to different data sources.
However, they managed to showcase something working, and even get
mentioned on OMG Ubuntu. So it might also be worth at least evaluation
possible collaboration with these guys.

So, this is the situation as I see it. Please just let me know what you
think, and which direction you would like to take for the future.

Regards
Frederik