[gst-devel] tags and metadata

Thomas Vander Stichele thomas at apestaart.org
Thu Jan 22 09:42:07 CET 2004


> It's about semantics all the time. If you don't define the semantics for 
> something, it's of no use, because everybody can use it for anything and when 
> you get around requesting it, it means just that: anything.
> That's the point in writing a library in the first place: You abstract away 
> functionality by making it adhere to your semantics. This requires semantics 
> first.

The semantics should be wider though than what the pretty crackful id3v2
spec has.

> This "I'll implement what I think is useful, we'll think about what it means 
> later" has brought us stuff like the last metadata implementation, locked 
> states and our current threading state.
> 
> Write first, think later is ok for apps. Libs should do think first, write 
> later.

Sure, I agree - in an ideal world.  In the GNOME world, however, if you
provide functionality in 2.4, you need to provide it in 2.6.  Whether or
not you agree with how that functionality is used is a different
matter.  Feel free to argue those points through the relevant channels.
But saying "We don't want to do it this way anymore, deal with it" is
not acceptable.  So I'm not sure if you understood what you signed up
for when you proposed gst 0.8 for gnome 2.6

It's not too late for you to ask that GNOME 2.6 not use GStreamer if
that is what you want.  It's not what I want though.

Thomas



Dave/Dina : future TV today ! - http://www.davedina.org/
<-*- thomas (dot) apestaart (dot) org -*->
You know well
Just never tell
If someone's got a secret hell
<-*- thomas (at) apestaart (dot) org -*->
URGent, best radio on the net - 24/7 ! - http://urgent.fm/






More information about the gstreamer-devel mailing list