by homeworld.cygnus.com with SMTP; 24 Sep 1998 08:54:38 -0000
	by shell17.ba.best.com (8.9.0/8.9.0/best.sh) id BAA01354
	for netbsd-advocacy@netbsd.org; Thu, 24 Sep 1998 01:54:00 -0700 (PDT)
From: Todd Whitesel <toddpw@best.com>
Message-Id: <199809240854.BAA01354@shell17.ba.best.com>
Subject: empirical method for generation of netbsd book.
To: netbsd-advocacy@netbsd.org
Date: Thu, 24 Sep 1998 01:54:00 -0700 (PDT)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

I have long held a theory about generating good technical documentation,
which unfortunately seems to be too uncomfortable or elaborate for people
to actually use unless their manager is asking them to do it.

It is simply to gather lots of empirical data about what newbie users
run into while installing and configuring their systems -- what kinds
of questions they ask, what common problems they encounter, and so on.

In short, you simply hold a well-stocked installation party and videotape it
extensively. Make rough transcripts of the various installation efforts and
use the topics they cover as the raw material for an installation handbook.

Then, repeat until diminishing returns starts to take over. Assuming your
newbie sample and target hardware sample are both representative enough,
the final data set is guaranteed to be of high quality.

However, I think it makes a lot of sense to wait until 1.4-BETA is mostly
converged before we marshal the resources to actually do this.

Todd Whitesel
toddpw @ best.com