gtk

[Michael Van Canneyt]

On Thu, 29 Apr 1999, Sergio A. Kessler wrote:

> Michael Van Canneyt <michael.vancanneyt at wisa.be> el día Thu, 29 Apr 1999 
> 09:37:16 +0200 (W. Europe Daylight Time), escribió:
> 
> >> btw, Michael, I'm the only one that think that the FCL is the
> >> "rigth thing to do (tm)", and then (or parallely) building the
> >> IDE (call it Megido, Lazarus, etc) from the FCL ??
> >
> >Lazarus builds on top of the FCL.
> >Megido will take a completely different approach.
> 
> yes, I know this, but why I do not understand is why the FCL
> is not populated with GUI components from the Lazarus team
> (I understand the case for Megido :)

see below

> 
> >> What is the future of the FCL if everyone and his brother
> >> are doing components separately from the FCL ?
> >> 
> >> The FCL has a roadmap ?
> >Yes, it has.
> >
> >I see the FCL as a subset of VCL which is Non-GUI, non Toolkit
> >specific. So All non-gui components can go in the FCL.
> 
> but ... it is a bad idea to put GUI components using the GTK
> in the linux directory, and put GUI components using the win
> API in the win32 directory and so on ??
> 
> thus making a real VCL clone
> 
> Even you can divide the linux directory in GTK, QT, Motif, etc
> the people interested will fill up this directories.
> 
> It is a bad idea ?
> Why no GUI components ?
> What is the intrinsic difference betwen a GUI component for
> linux and one unit of the RTL for linux ?

Because GUI stuff depends not only on a OS, but also on toolkits
(Native Xlib, Qt, GTK, X-Forms, Fox etc) which is highly a matter
of personal taste. The non-gui stuff can be implemented relatively
OS-independent. 


> Also I think this is good to prevent code forking, rigth now,
> there is Megido, Lazarus and Nico that want to go alone,
> tomorow you will have dozen teams doing the same, and yes,
> competition is good, but this is beginning to sound like 
> nonsense, just because the FCL, *IMO* has left this hole.
> 
> I'm missing something ??

Yes, if we do as you suggest
1) the FCL will be much harder to maintain.
   (Agreed, this is largely an organizatorial problem,
    which can be solved )
   
2) We then must take 1 project (e.g. Lazarus) and say
   'this is the one'
   and I am not sure this is 'The right thing (c)' :)

But I will speak with some other developers, because you seem quite
convinced that we should do this, and maybe the other core developers
agree more with you than with me...

I will propose them your stuff from the mailing list:

  /fcl/win32
      /linux/gtk
            /Qt
            /x-forms
            /Xlib

I'll post any results to lazarus.com

Michael.