Re: Known MacOS programmer about WO/EOF

[Wim Oudshoorn]

Although rather pointless I guess, but let me add my point
of view as well to this thread.  I could have replied to
any other post, so you best read this as a general reply.
 
Guenther Noack <gnoack@guenthernoack.de> writes:

> Hi!
>
> On Wed, 30 Mar 2005 14:23:00 +0100
> Nicolas Roard <nicolas@roard.com> wrote:
>> Sad but true :-)
>> Mostly, they relay on old 3rd persons accounts about GNUstep and don't
>> try it themselves; and many don't see the point "as they already have 
>> Cocoa". The remaining won't help the project "until it will work"
>> (that's it, until everything will be smooth and all, they won't even 
>> try)

Indeed.  I have not spoken to many Cocoa developers myself, but 
if I imagine myself to be a Cocoa developer I would think:

* I like developing in Cocoa
* Cross platform is only a secondary. 
* If I care about cross platform applications I would mainly care 
  for Microsoft Windows
* If I really cared about cross platform development, I would have
  picked a cross platform development environment, like JAVA.
* GNUstep is interesting if with minimal work I would be able
  to port/recompile my existing application for Windows.
* The last time I looked at GNUstep it does not seem to progress much
  and I haven't heard any success stories porting applications.
  (Note porting from GNUstep --> Cocoa does not count)

Conclusion:
* GNUstep is not ready, and seems to be in the same "not ready" state
  for the last year.

> Wise words. Please note that Nicolas used quotation marks around the
> "until it will work" part of the sentence. ;-) Does it really "have to
> work" to attract developers? Don't the core libraries work very well
> by now? I think the bigger problem is to "make it look like it works".
> ;-)

Yes, I think the core libraries work very well.  But, as Helge Hess
pointed out, not perfect[1].  Also the image is, IMHO not helped by
the fact that it is not clear what GNUstep is.  Is it:

* Base library
* Base + AppKit library
* Desktop environment based on OpenStep (Oh and which one?)
* Libraries + applications
* What about GSWeb? GDL2?

And because all these above are in a very different state of maturity
both are true statements:

* GNUstep is mature and stable
* GNUstep is vaporware and very unstable.

> I personally think that when Camaelon 2.0 is released, GNUstep will
> suddenly get a small boost of public awareness. It's all about the
> look of the interface. 

I personally think this will help almost nothing.  The question is, who 
do you want to attract?

* Cocoa developers?
  If they are really motivated they might try it.  Their application
  will not work, with or without themes.  So, they will conclude
  that the GNUstep people are "hobying" again instead of delivering
  a solid Cocoa implementation.

* Other Linux developers?
  They have had themeability for years, so why would they bother?

> (Would OSX have gotten such a big attention
> if it wouldn't have had this fancy Aqua-look? ;->) Most people just
> can't motivate themselves to try a desktop (*duck*), whose widgets
> look this "old".

Well, end users might want to have a nice look, but they probably
also want lots of nice applications.

BTW.  It is a shame that the Marketing Session at Fosdem
did not materialize.  Because I think we could have 
some meaningfull discussion about how we can change the
image of GNUstep.  

Finally let me add some random notes on the subject:

* I think there should be a CLEAR distinction between
  the parts of GNUstep:
  
  * make/base
  * gui
  * gsweb/gdl
  * developer tools 
  * etc.

* The maturity of the different components on different platforms 
  should be clear.

* There was a section with "commercial" projects that uses 
  different parts of the list above.  So that it can show for 
  example that

  * make/base is used by company X, Y, Z, ...
    because they had to migrate from a ObjectiveC YellowBox application
    to something new.  All testify it was quite easy and base has been
    very stable.

  This will show that make/base is stable enough to build enterprise 
  applications.

  Although this probably only attracts old OpenStep developers, 
  I think it will be good for the image of GNUstep.

  Also, reading this list I think there are a few companies around
  who use GNUstep base for just that.  (I work at one, but 
  I don't think we are the only one.)

* Have a concrete statement of direction for the different parts
  of GNUstep.

  Like:  
  * base will be fully OpenStep compliant.
  * base will follow Mac OS X features except: A, B, C, ... for the following
    reason...
    
  * appkit ...

  Basically this will not be different from the current statement of direction,
  but make it a little bit more explicit.


The above will of course not do anything immediately to add new 
developers/users.
However, If I am developing on Mac OS X and my next application should 
potentially
be cross platform to Windows/Linux.  I want to be able to make a quick
assesment if GNUstep is going to help me.  After reading about GNUstep
I should have a lot of confidence that it indeed is going to work, or
show me very clearly where the gaps are, what to avoid and how much
effort it rougly will be.  If we are honest and clear about that 
it will hopefully avoid disappointments and slowly build a better image.  


Wim Oudshoorn.

P.S.:  Sorry for making a long thread even longer.  As usual
       I will not spent time on this.
      
       Also from today I will be on holiday for about a week,
       so I will not reply soon.


[1] Helge,  I am curious what problem you had building your library
    on windows.