About the audio/video chat solution for Plasma Active 3

[George Kiagiadakis]

On Fri, Feb 10, 2012 at 7:01 PM, David Edmundson
<david at davidedmundson.co.uk> wrote:
> I think we need to do this in two parts.
>
> 1) Make a desktop app. Really really simple, don't get bogged down
> messing about with QML, handling multiple users or any other such
> nonsense.

Well, we kind-of have that already, don't we? The code is a total
mess, though, and I think it's not properly plugged into the backend
after my last rewrite of the backend, but well, it exists. I'd really
prefer someone rewriting it, though.

> In order to do this we need:
>  - Code in Telepathy-Qt-Yell
> and that's it.

We need this to release the call-ui, but we don't need this to start
working on the UI. Using some slightly older version of
telepathy-farstream and alin's patch for tp-qt4-yell, the call-ui
should compile and do some proof-of-concept audio calls. This is
enough for someone to start working on the UI.

> We don't need it merged into TpQt to start on this, tbh I don't think
> we even need it merged to ship we can just compile it statically with
> the call-ui.

I really hope we won't have to do that, but in case we absolutely
must, it can be done.

> Once we have a basic app, lots of people will be around to help add
> polish/features afterwards. From experience it's far better to be
> simple and iterative, than to try and do lots, get stuck and just
> crash and burn.
>
> 2) Making a Plasma app.
>
> Afterwards this should be simple enough, I think we should make the
> desktop client first as it's far easier to iron our all the bugs
> there. We did this for the Text-UI and made a plasmoid once I knew
> what was needed and what needed doing. It was a pretty smooth process.

+1

> As for sponsorship, I've been talking to Martin. It's unlikely that
> any one person will code all this, as there's lots of little tasks -
> if we did get any sponsorship it should be split on a per-task basis.
> Personally I'd be happy to be sponsored as I am truly amazing and good
> at getting stuff done. It would allow me to have more time hacking as
> I could cut down on my hours teaching.
>
> I think the key is getting this TpQtYell work done. I think the spec
> is fixed enough now that it can be started. The question is whether
> Collabora are going to do it, which they'll only do if they have a
> client who wants them, or whether we will have to. I'd only hack on
> TpQt if I was being paid to do it as it's boring, I imagine many
> others feel the same. I can ping their mailing list and find out as
> this is the important part.

I have already claimed that I intend to work on it this month.
Regarding the rest of the upstream work, this is intended to be pushed
into gnome 3.4, which has a freeze later this month, so I'm quite
confident that everything will be in place when we will want to
release our call-ui. But as I said above, the UI work doesn't need to
depend on it.

> Dave
>
> 2012/2/10 Mario Fux <kde-ml at unormal.org>:
>> - What the overhead for such a solution on Plasma Active (PA)?
>> - Did somebody of you already worked with PA?

I haven't worked with PA and don't really know much about it, so I
can't really say. However, Telepathy/Farstream audio/video technology
has been successfully deployed in mobile devices in the past (the most
notable example being the Nokia N900), so I don't think we are going
to have any issue here.

PS: "PA"? :S PA=PulseAudio for me :P

>> ---------- Forwarded message ----------
>> From: Martin Klapetek <martin.klapetek at gmail.com>
>> [...]
>> Tasks that need to be done:
>>
>> Telepathy Call1 spec needs to be implemented in Connection Managers
>>
>> this is almost done by Collabora (xclaesse), details unknown

Yeah, almost done, both in gabble and rakia.

>>
>> Also unfortunately the work can't start before the Call1 spec is implemented upstream, otherwise we have nothing to work with. We also need QML bindings for GStreamer, which should be ready by the end of January.
>>

QML video in QtGStreamer - check (see my blog ;)
Work can start without call1, as I said above.

Best regards,
George