Re: Discussing issues

On Tue, Apr 15, 2003 at 01:33:18AM +0100, Alan Hourihane wrote:
> > One thing I'm interested in is pulling together a set of components
> > and saying "this is the base GUI platform for X11/UNIX." I'm also
> > interested in growing the scope of that platform over time.
>  
> You can't define that. It's always going to be a moving target.

Yes it's moving, unless it's stagnant. However, there is value
in having only *one* moving target. ;-)
 
> > Creating a shared platform means not only including things, but also
> > excluding/deprecating things. We can't have 6 font subsystems, that
> > defeats the purpose - it would keep applications from adopting any of
> > them.
>  
> You mention excluding/deprecating, but what about superceeding ?
> You've already got bitmap/Speedo/Type1/FreeType etc font systems.
> Applications are written with the tools that are best suited for those
> who are writing the application - they adopt what fits, not what claims
> to be a standard.

Part of what fits is what's part of the platform.

Here is why you need a single font subsystem:

 - if I have two applications, I should have the same list of 
   fonts in each one.
 - if I install a font, using a single installation method, 
   it should show up in all applications.

There are dozens of subsystems with similar properties. The
application is better if it uses the same system as other apps.  And
the sum of the all the applications is better than its parts if all
the apps are consistent and interoperable.

App authors are also interested in whether users are likely to have a
given subsystem installed.

You can speed up progress significantly if people have a place where
they can see a subsystem "blessed." There are many third party widgets 
out there; but we can get those widgets more widely-used by 10000% by
moving them into GTK+.

If there wasn't value to a canonical implementation tree, then nobody
would ever submit patches to any project.

Havoc