Re: [Fwd: Re: really attracting developers]

[Helge Hess]

On Aug 25, 2006, at 14:55, Andrew Satori wrote:
> What they aren't is Cocoa complete, and in many areas, it's a near  
> rewrite to port advanced applications from Cocoa to GnuStep.   
> That's not saying it can't be done, or that it's a bad thing, but  
> one of GnuStep's major selling points is it's likeness to OpenStep  
> & Cocoa, and there is a whole lot more Cocoa development going on  
> today than there is OpenStep.

The reason that Cocoa apps are not ported is not that it can't be  
done, its because it doesn't make a lot of sense for the authors.  
Remember that the far majority of the Cocoa apps are not OpenSource/ 
Free Software ones but proprietary ones. That is, people want to make  
money by selling them (there are exceptions, like Adium, but those  
are sparse).

Now even Linux _desktop_ as a whole (including KDE and GNOME) has a  
much smaller market share than MacOS desktop has. So its not very  
interesting from a start. And then, GNUstep-like applications have an  
even smaller market share. If you would want to explore the Linux  
market, you must choose KDE or GNOME to reach any kind of significant  
audience.

So ports are basically restricted to niché enterprise applications  
providing custom applications. Now, how many Cocoa enterprise  
applications are there? A few, but in itself they are extremely niché  
too.

Summary: having Cocoa compatibility (which is getting harder every  
day as MacOS advances, just think ObjC 2.0) for Linux is _not_ a  
selling point. The majority of Cocoa developers simply don't want to  
deploy their desktop applications to Linux/BSD.


IMHO there are two spaces which can be explored if you want to  
advance the GNUstep community:
a) reasonably easy and convenient Windows porting for Cocoa developers
b) server stuff

Obviously a) is interesting for the sales-driven Cocoa developers  
since the Windows desktop market is much larger than the MacOS one. I  
think a lot of people would be _very_ pleased to be able to provide  
Windows versions of their apps with a reasonable effort (it doesn't  
and probably can't be a simple recompilation).

Well, and b), Cocoa people usually want to stick with ObjC, which  
they use for their desktop apps. And they are most likely using Linux  
on their servers. Which makes ObjC server development for Linux  
interesting. Maybe DO Cocoa <-> DO GNUstep interoperability is a key  
thing here.
Though the server development area is extremely crowded too.


Just a few thoughts ;-)

Greets,
  Helge
--
Helge Hess
http://docs.opengroupware.org/Members/helge/