Why Linux is not (yet) ready for the desktop

[Samuel Thurston, III]

Hello,

On Mon, May 25, 2009 at 12:38 PM, Peter Davis <paddax at googlemail.com> wrote:

> Regarding PulseAudio, its a good technology but what Linux needs is a single technology that everyone can sign up to, if thats PulseAudio then great but please could we STOP inventing new one audio sub systems.

OSS was kind of a hack job, that made audio work for most cards on
most systems, and provided an adquate API to get apps on linux using
sound.
ALSA is a great subsystem that replaces OSS, adds support for a number
of cards, adds a lot more control of card features (mixer settings
etc) and in general is a great improvement.

Sound servers like esd, jackd and more recently pulseaudio are another
beast entirely.  They utilize the underlying sound system to provide
advanced features like routing, real-time processing, mixing,
resampling and other things that you just expect your computer to do.

The separation between the two is what permits for continued feature
development.  It is in line with the unix-development philosophy of
using a chain of many small tools to get the job done.  You may say
"quit making new subsystems" but go ahead and uninstall pulseaudio and
go back to using OSS and esd for a while... they're both very mature
and well-supported, and they still suck. Developers come up with new
ones because they're improvements on a less than ideal situation.

Unfortunately sometimes the problem cannot be fixed by just continuing
to rework an inadequate design.  The good news is, there's room to
rebuild indefinitely.