glib comments

[Zack Rusin]

On Monday 24 February 2003 20:10, Havoc Pennington wrote:
> Well, I don't know anything about multimedia, but hopefully the above
> is useful information about GLib policies/intentions.

GLib in itself is not the problem. I and many others KDE developers 
enjoy it and it's already used in ARTS (see arts/gmcop). The issue 
arise due to the fact that GStreamer object system is based on GObject 
which doesn't naturally fit into Qt object system. People are in 
general afraid of GObject ( by the way you should make Mathieu Lacage 
maintain his GObject docs in glib cvs not on his website :p )
People are afraid that this would give Gnome an unfair edge ("hey, KDE 
is super bloated and confusing because it has two object systems and to 
code in KDE one has to learn Gnome and KDE"). These theories are of 
course fallacies and we need to get to the bottom of those and make 
sure everyone understands that this is simply not true. ( it's enough 
to look at Scotts JuK which uses Tim's GStreamer bindings to see that 
). [ Havoc you can end reading right here if you want to save time ;) ]

I think the problem is that although we bring up different frameworks ( 
by the way Tim http://www.networkmultimedia.org seems to be very well 
designed. I just started reading their docs and I'm fairly impressed ) 
we still haven't specified what we are exactly looking for. 
So at this point I'd like to propose to do the following :
- create a complete and comprehensive list of futures we are looking for 
(low latency? easy of integration with our object system? ... ). What 
are our needs? Anyone wants to start? 
- a person who wants to propose a framework will do analysis of it based 
on the above list. That will immediately tell us which frameworks are 
better suited for our needs.

Comments?

Zack

-- 
Notice: Your mouse has been moved. Windows will now restart so this
change can take effect.