Google Summer of Code

Luke Schierer lschiere at
Sun Mar 8 11:42:39 EDT 2009

On Sat, Mar 07, 2009 at 03:38:33PM -0500, John Bailey wrote:
> Gregor Dick wrote:
> > John Bailey wrote:
> >> I'd really like to get the native win32 UI project going this year if we have a
> >> good applicant.  I'm hoping that starting the project will spark interest in the
> >> Windows-inclined developers out there.
> > 
> > Hello,
> > 
> > I'm also considering throwing my hat into the ring for the Win32 UI
> > task. My name is Gregor Dick and I'm a final-year undergraduate
> > mathematics student at the University of Edinburgh. I would propose to
> > implement the UI using the Win32 API directly, Petzold-style, without
> > using any particular framework.
> > 
> > As far as I'm aware there hasn't been any concrete discussion regarding
> > particular requirements for a Win32 front-end, for instance in which
> > respects, if any, it should attempt to be self-consciously Pidgin-ish.
> > If there's any established opinion on such matters, I'd be grateful if
> > someone could please point it out so that I can work it into my proposal
> > and prototype when the period for applications begins.
> I was hoping that the design would be overall similar to Pidgin, but with
> behavior that makes sense for the Windows users.  For example, the buddy list
> design works well, but the mini dialogs that we have may not make sense in a
> Windows application (I've never seen such a thing on Windows).  The Pidgin
> preferences window design may not make as much sense as something that's more in
> line with other applications on Windows.  There are other examples, but I can't
> think of them at the moment.
> John

Default key-strokes, font handling, URI handling and so on should all
conform to the expectations that windows users have.  Little (but
important to various subsets of users) things like the alt+decimal
latin1 insert should work. 

Basically, over the years we've heard repeatedly that in various small,
but (when added together) significant ways, Pidgin feels out of place,
foriegn, to windows users.  The native UI should deviate from Pidgin
whereever it makes sense to make windows users feel comfortable with it.

That would be how I would start with the project anyway.  But like John
said, I hope that this will take on a life of its own, and sure, if it
comes to differ from Pidgin as much as Adium does, if it is pleasing a
user base, great!  The goal here is not (primarily) to recreate Pidgin
in native widgets, but to start what can grow from an SOC project into
a project of its own, that people can be excited about in its own right.

The goal of libpurple is to foster choice in IM front ends without the
cost of multiple projects each figuring out the various protocols in a
mostly independent way.  The hope is that we can collaborate on
libpurple, and agree that sometimes we differ on frontends.  That's why
Adium & Finch are exciting, and why I wouldn't expect a windows UI
client or a QT ui to stop at trying to replicate Pidgin. 


More information about the Devel mailing list