From alsa-devel-owner@alsa.jcu.cz  Fri Jan  8 17:55:04 1999
Received: from renoir.op.net (root@renoir.op.net [209.152.193.4])
	by marvin.jcu.cz (8.9.1a/8.9.1) with ESMTP id RAA05628
	for <alsa-devel@alsa.jcu.cz>; Fri, 8 Jan 1999 17:53:54 +0100
Received: from someip.ppp.op.net (d-bm3-09.ppp.op.net [209.152.194.73]) by renoir.op.net (o1/$Revision: 1.18 $) with ESMTP id KAA19892; Fri, 8 Jan 1999 10:25:08 -0500 (EST)
Message-Id: <199901081525.KAA19892@renoir.op.net>
To: alsa-devel@alsa.jcu.cz
Cc: Stephen Thornton <Stephen.John.Thornton@ericsson.no>
Subject: Re: Hi - need some help? 
In-reply-to: Your message of "Fri, 08 Jan 1999 12:51:55 +0100."
             <9901081151.AA25556@tele2> 
Date: Fri, 08 Jan 1999 10:24:46 -0500
From: Paul Barton-Davis <pbd@op.net>
Reply-To: alsa-devel@alsa.jcu.cz
Sender: alsa-devel-owner@alsa.jcu.cz
Precedence: list

>my dream for years has been to have an operating system (yes an entire
>operating system) designed specifically for music applications. The

I suggest you look into BeOS. Unlike Linux, it was designed with "real
time", "multi-media" in mind right from the start.

>first step though is to get a good code base in the public domain of
>the layers that sit on top of the OS - to this end Linux is the best
>development system I can think of.

I don't believe that a preemptive multitasking OS is the ideal base
for music applications. RTLinux, which is not really a preemptive
multitasking OS in the normal sense of that phrase, might be suitable,
but regular Linux can offer you no guarantee that what you want
executing on a processor at any given moment will actually be doing so.
The reason the Mac and Windows applications work well enough to show
up in studios is that they *own* the hardware while they run. Linux
won't let you do that. Instead it will offer you the illusion that it
does, and one day, when you've gotten the best player of <insert
instrument here> into your recording setup, and they're playing their
heart out, someting stupid will happen, and you'll lose a couple of
milliseconds of their work thanks to the scheduler kicking in.

>My current idea is to use a
>realtime system running under Linux that can handle timing and such
>properly. 

If you read the linux kernel mailing list, you'll realize that Linux
doesn't do realtime. 

--p

