[GSoC2012] Introduction / Impress smartphone remote control

[Andrzej J. R. Hunt]

On 26/04/12 10:04, Tor Lillqvist wrote:
>
> In your proposal you said from "middle of April", so let's hope this
> doesn't slip too much.
>
> Have you already managed to get a build of LibreOffice? Please hang
> around on the developer IRC channel, #libreoffice-dev on freenode, and
> don't be afraid to ask even what seems like silly questions publicly.
Yep, I've got a build working fine (on Linux that is -- I'm hoping to 
set up a build environment under windows once I get back home in June). 
I'll probably start coding tomorrow or saturday (I want to get a small 
patch finished off tomorrow before I start on the project).

> Hmm, I am not sure about this, I think we would much prefer to have
> the code as C++ code inside LO itself, and thus licensed as LGPL/MPL
> (the combination which we use for all newly written code).
>
Yep, I was thinking about this after I'd sent my email; and based on 
what other have said, I've decided it is best to use C++ and LGPL/MPL 
Licensing to allow integration into LO -- I'll then abstract away the OS 
specific bluetooth stuff as far as possible (since I don't have a mac 
nearby to test on, I'd need someone to do that testing -- but that's one 
of the last things to be done). Initially the plan is to use IP/Wifi for 
the communication, with bluetooth just being an additional option added 
once I've got a basic controller over IP working. Since I was planning 
to use RPC (specifically: Apache Thrift), I'd simply have a serial 
connection over either TCP or bluetooth -- the bluetooth specific code 
is limited to finding devices and setting up a serial connection for the 
RPC to communicate over.

(The library I was considering using was PyBluez -- I can still look at 
this to see how bluetooth is used on Windows and Mac when I get to that 
stage.)

For authentification I thought it would be easiest to have a passphrase 
which is used to encrypt the serial stream between LO and the app -- 
this is mainly to prevent people playing pranks rather than to provide 
high security.

On 26/04/12 10:04, Noel Grandin wrote:
> Hopefully we should be trying to integrate with existing standards and 
> solutions.
> On Windows that would be things like LIRC and the Bluetooth-HID
> http://winlirc.sourceforge.net/developer.html
> http://en.wikipedia.org/wiki/Bluetooth_profile#Human_Interface_Device_Profile_.28HID.29 
>
I was planning on using RPC (see above) -- this is necessary since I 
also want to be able to preview slides etc., with a HID device this 
wouldn't be possible. (There is already an app for android implementing 
HID, which can send keystrokes to the computer, i.e. for a simple 
back/forward control that could quite easily be modified : 
https://code.google.com/p/androhid/  -- but I want to be able to do more.)


Cheers,

Andrzej