Another Potential Slave... I mean Programmer

Joe Tennies rotund at fatnsoft.com
Mon Feb 4 01:01:23 CST 2002


Okay.  I'll admit I am biased toward GNOME, but my real problem is that I
didn't explain myself much, and I doubt people actually went to look at the
gstreamer site.  I had found it before, but forgot about.

Why I think KDE should adopt gstreamer as well.  It is not a replacement for
ALSA, OSS, ESD, or aRts... it's an add-on to it.  My understanding of the
gstreamer project is that it's Linux/UNIX's response to the Windows CODEC
system.  It is a single interface to output video and audio from various
types of input file types.  They have support for MP3s, DIVX (coming), Ogg
Vorbis, DVD, AVI, MPEG, Video 4 Linux, etc. for input files.  It also
supports OSS, ALSA, esd, aRts, X11, aalib (gotta love it), SDL, etc. for
outputs.

This is obviously not for programs that need low latency.  It is for your
everyday stuff.  It is more than good enough for saying "you got mail," yet
it would have support for lots of files types w/o specifically writing
support for the new file type.  This is also good for multimedia players and
web browsers due to the support for lots of file types (I could see XMMS
switching to this from its plugin system).

Is this the answer for recording studio programs, remakes of Rebirth (WE
NEED ONE OF THESE), or people that need the audio fast, but for the typical
desktop user, it will be a god send.

Only downfall I could see to the KDE crowd is the need to install GLIB.  It
doesn't require the full GTK+, just GLIB (linked lists and other
datastructures).  KDE could recommend using aRts for output and write its
own input plugins for implementations that depended on a library KDE didn't
need installed but KDE did install another one that implemented something
simiilar.

HECK, I don't think anyone would mind KDE writing their own underlying/core
part, required you use aRts for output, didn't require GLIB (though this may
be difficult), but used those input plug-ins.  I understand KDE and GNOME
have very different things in their head, but we need to get some standards
here.  XDND was way overdue (though I hear it wasn't a great implementation,
at least they agreed on how to do something), and the two groups need to
make more decisions like these to help each other strive.

Remember everyone.  I WAS a HUGE KDE fan myself for years.  I had installed
GNOME a couple times and hated it.  I used KDE through 2.1 (I think...
definitely in the 2.X).  Then I discovered I had Blackbox on my machine, and
it flew.  I like the QT library (though I won't use it because I am now a
GNOME guy).  I like the layout of KDE... simple yet sophisticated.  It
tailors to the geek and the 'typical home user.'  I just like GNOME's CORBA
interface they are adding (which is where gstreamer will fit in).  Something
that can be so easily updated and contain potential problems is powerful
(though VERY frustrating at times).  I also love Ximian, who I feel keeps
pushing the things that need to get done to make Linux a full fledged
desktop system -- not to mention their love of all things monkey =)

I salute KDE.  They did the large scale desktop environment for the masses
first (CDE is NOT for the masses anymore).  They are probably still doing it
best (I hear lots of complaints about some of GNOME's design).  The problem
I see is that there's no one really pushing KDE to keep its lead (and the
slow speed it had under Mandrake).

-----Original Message-----
From: Christopher Morgan [mailto:chmorgan at speakeasy.net]
Sent: Saturday, February 02, 2002 11:46 PM
To: Joe Tennies
Subject: Re: Another Potential Slave... I mean Programmer


It would be interesting to see a latency comparison between gstreamer, arts
and
jack.  If gstreamer doesn't get very close to what jack can do then it isn't
the
solution for the future.  We need low latency or tons of audio apps will
continue to find it through direct alsa or jack interfaces.  Ideally we
should
pick the best solution and much of it is based on the latency issue,
features
should come after and should be designed such that they don't increase
latency
by any more than they have to.  It sounds a little like you want to go with
a
gnome solution because you use gnome.  I think more work needs to be done to
come up with a linux(open source or whatever as well) wide solution to the
problem.

Chris




On 2 Feb 2002, Joe Tennies wrote:

> On Sat, 2002-02-02 at 05:42, Hetz Ben Hamo wrote:
> > > I'm not trying to start a Gnome/KDE flame war, but as far as I know,
esd is
> > > unmaintained and the Gnome developpers are thinking about switching to
> > > aRtsd. Wouldn't it be more useful to write an aRts driver ?
> >
> > Actually if I'm not mistaken - GNOME people are moving to gstreamer, if
I'm
> > not mistaken. I'm not sure since I work with KDE here...
> >
> > Hetz
> >
> Actually, I just looked at gstreamer.  It lies above aRts, esd, alsa, and
oss.  It is a platform independent (well, not ported to everything yet, but
working on it) system.  I think KDE SHOULD be heading this way and am
impressed that GNOME is.  I think that gstreamer would be the way to go for
WINE.
>
>
>
>





More information about the wine-devel mailing list