Re: The GNUstep code challenge.

[Nicolas Roard]

On 5/14/06, Marko Riedel <markoriedelde@yahoo.de> wrote:

Hi marko,

> Folks, GNUstep definitely needs more visibility.

it wouldn't hurt..

> That's why I'm suggesting we establish
>
>     The GNUstep Code Challenge,
>
> a biannual coding competition with problems that are set up to require the
> study and use of Foundation/Appkit classes. We need a sponsor for this so
> we can offer real prizes (say three prizes in the 250-500 Dollar range),
> otherwise people won't be interested. I would be willing to volunteer as a
> problem designer and evaluate submissions. There are many aspects of
> GNUstep that I don't have a grasp on, but that's what teams and
> collaboration are for.
>
> What do you think? We'd announce it to the media and it would get us some
> coverage that we really need, as well as motivate people to learn
> Objective C and familiarize themselves with the OpenStep framework.
>
> I eagerly await your comments and observations.

Frankly, I'm not sure. First, you'd need quite a few $$$ to be
attractive I'd say, or it would look a bit ridiculous... which is not
what you want if if the goal is use it as some kind of PR device :)

Secondly, apart from the slight media buzz created by it, what will be
the consequences ?.. if it's a competition, it also means a lot of
duplicated efforts..

That's why I'm not personally very convinced, but well if you do
organize that, it's cool -- any kind of initiative towards GNUstep is
:)

But it wouldn't hurt to see what exactly do we want to do for GNUstep
-- I mean, what's the point of GNUstep and how to attract more
contributors ?

There's three broadly defined populations that have an interest in
GNUstep, in my experience:
1) - people that use -base and just that, eg server-side
2) - people that want to recompile/port Cocoa application
3) - people that want a GNUstep desktop

#1 is brainstorm, opengroupware, manuel, david, etc... historically
important but somehow I'm not sure it will grow -- the server-side
area is pretty crowded with a wide range of solutions nowadays, so
what are the advantages of GNUstep here...

A lots of recent effort went toward #2 as well -- improving the
Windows port, the recent effort for nib compatibility, etc.

And It seems to me that #3 has also many people interested in.

Good thing with #3 is that it actually  provide a _target_ for
developers of gui applications : "hey I'm not developing an
application for 3 persons using gnustep from svn but for a complete
desktop!". The other possibility is to work for more integration with
the other desktops (kde, gnome, or even windows..), which answer the
same problem : "I'm using GNUstep because I can develop faster, but it
doesn't prevent anyone to use my program properly integrated in their
own desktop!" (win-win)

So it looks to me that most of the things that would help a lot
GNUstep is to work on -gui in general and progress on the
interoperability aspect as well. Plus themes obviously, which you need
anyway to be properly integrated in an other desktop.

The only problem with that conclusion is that even if -gui is central
to the future of GNUstep, very, very few persons are actually working
on -gui... most other developers seem to work at the application
level... It's perfectly fine and normal, but everybody's waiting for
bug fixes or progress on -gui, while there's actually just a few
persons working on it regularly -- if you take from one year ago,
there is only 136 commits until now according to the changelog... :-/

So well, attracting people, sure.. I effectively truly believe that a
"killer app" would attract lots of people to GNUstep as well...

but we should seriously focus on -gui, because really, it lacks
attention and progress. It's particularly frustrating because -gui
works quite well for many things, but... things are lacking. A few
"big" things, but many tiny ones as well. I would say, applications
developers should have a look into -gui and try to tackles some
problems... the code is pretty clear most of the time and it's not
like there's no documentation to check against...

So I would be more interested if we could focus our energy to make a
proper list of the areas that need work (top of my head, printing,
slow drawing, drag and drop using XDND, nsformatters, etc) and then
try to work on them :)

Short of organizing a gnustep coding party were all developers could
meet for a few days and work together on gnustep (think of the KDE
parties or the *BSD hackatons: they prove to be very successful to
improve the projects) we could at least organize "virtual hackatons"
using irc and one area at a time we'll progress!

Also, GNUstep code is not that difficult to work with -- do not
hesitate to have a look and see if you can improve things, you'll be
surprised :)

So well, more than attracting people to gnustep, we should try to
attract existing gnustep app developers to gnustep-gui :-P

Cheers,

--
Nicolas Roard
"I love deadlines. I like the whooshing sound they make as they fly
by." -- Douglas Adams