From owner-freebsd-doc  Sun Jan 23 17:21:39 2000
Delivered-To: freebsd-doc@freebsd.org
Received: from taurus.cs.albany.edu (taurus.cs.albany.edu [169.226.2.109])
	by hub.freebsd.org (Postfix) with ESMTP
	id D122814C4F; Sun, 23 Jan 2000 17:21:34 -0800 (PST)
	(envelope-from maniatty@cs.albany.edu)
Received: from richard.cs.albany.edu (richard.cs.albany.edu [169.226.2.48])
	by taurus.cs.albany.edu (8.9.3+Sun/8.9.1) with ESMTP id UAA22758;
	Sun, 23 Jan 2000 20:21:30 -0500 (EST)
From: "William A. Maniatty" <maniatty@cs.albany.edu>
Received: (from maniatty@localhost) by richard.cs.albany.edu (SMI-8.6/CLI2) id UAA01945; Sun, 23 Jan 2000 20:23:01 -0500
Date: Sun, 23 Jan 2000 20:23:01 -0500
Message-Id: <200001240123.UAA01945@richard.cs.albany.edu>
To: maniattb@cs.rpi.edu, msmith@freebsd.org
Subject: Re: Learning the FreeBSD Kernel
Cc: FreeBSD-doc@freebsd.org, asmodai@wxs.nl,
	freebsd-hackers@freebsd.org, grog@lemis.com, maniatty@cs.albany.edu
Sender: owner-freebsd-doc@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org

Mike Smith Writes (in response to me):
>
>> 
>> Actually the lack a reference or tutorial document is kind of damning
>> in a way, since it gives hardware vendors a powerful excuse not to
>> support FreeBSD in the way that say Lin*x is supported. 
>
>We have a better solution; they just give us the documentation and we
>write the drivers for them.  The end result is a better driver, produced
>in less time, and with a motivated maintainer.



This might not be right, many vendors don't want their competitors to know
their interface specification (to prevent cheap clones).  FreeBSD does
not have that restriction, so you may have it backwards.


>I've also written skeleton drivers for hardware vendors; typically they
>just have to fill in the very lowest level of the driver for their device,
>the rest can be generated in a couple of hours based on only the very
>simplest description.

So where is that very simplest description written, maintained and
published?

>Writing documentation is a resource-sucking nuisance; supporting outdated
>documentation even more so.  The BSD driver model is sufficiently simple
>that it's fair to say that if you can't work it out from the code, you
>probably shouldn't be writing a driver in the first place.  That's not to 
>say that a document describing the process wouldn't be good, merely that 
>such a document isn't actually going to help _us_ very much at all.

I do not happen to agree.  A better solution is to lower the barrier to
entry so that people (including hardware vendors) can supply drivers.
Right now the situation is that both sides say "Documentation is not
in our best interest, so you cannot have any.", which is bad.
Actually such a document COULD INCREASE DEVICE SUPPORT which is a MAJOR
FACTOR in people's decision to use or avoid an OS.  Perhaps less time
could be spent whining about lack of users and how users don't have clue
and more time could be spent making the good technology accessible.

Regards:

Bill Maniatty


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-doc" in the body of the message