GSoC 09 Project : Write a Native User Interface for Microsoft Windows Based on libpurple

[John Bailey]

Deepank Gupta wrote:
> Specifically, I would like to ask the following questions here, before
> drafting a proposal:

I look forward to seeing your proposal in our queue.

> Q.1 : Do we want to make certain changes to the UI for the Windows
> version or do we want to keep the UI consistent over all Operating Systems?

I think overall we should have similar UI's, but with changes that integrate
better with Windows.  For example:
  * Pidgin doesn't support the alt+xxxx character input that Windows users have
come to expect; this is something that a native Windows UI should do.
  * Pidgin uses mini dialogs in the buddy list windows for authorization
requests, connection errors, etc.  These may not make sense in a Windows
application.
  * Default keyboard shortcuts in Pidgin don't always match what Windows users
expect.
  * Windows users expect completely different font behavior and sizing from what
GTK+ provide.

Basically the idea here is to use Pidgin as a reference point to say "this is
the set of features we should eventually support" and "this is a basic design
that works, but let's improve it in ways specific to our platform."

> Q.2 : Since, the project is large and doing the complete testing and
> deployment of the project might be somewhat tough. What milestones and
> how much amount of work do you see a GSoC Project can be about?

This is a difficult question for me to answer; I have no idea what it takes to
develop a native Windows grapical application.  Realistically, the Windows UI
project isn't just an SoC project.  The idea is intended to get an interested
developer started on it to launch the project and attract additional developers
to continue the project after the SoC finishes.

Basically we're not looking for mercenaries that are going to collect a paycheck
from Google and run.  We're looking for people who are going to stick with and
grow the community.

> Q.3 : There have been two contrasting views put up by the community. One
> being that the UI development should be done in MFC since the .Net Based
> applications are slower. Other view is that .Net Based applications are
> easier to write. Personally, I feel myself capable of doing development
> in either of the two. Though, since we are targetting at a Desktop based
> application and not at a embedded device application, I feel myself
> inclined to write code in .Net since that is going to be the platform
> going into the future.
> Though, it is for the community to decide upon.

I'm opposed to .NET simply because the applications feel slow to me even on
brand-new hardware (although thankfully not as slow as Java).  That said, I know
it's the so-called future on Windows, so I guess I'll have to live with that
reality.  I leave it to the applicant to pick which approach he/she wishes to
use.  We'll rate the applications based on their own merit, not on the toolkit
chosen.

John

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: OpenPGP digital signature
URL: <http://pidgin.im/pipermail/devel/attachments/20090328/dcc28ac7/attachment.sig>