Subject: Re: Open Source shareware?
From: "Tim O'Reilly" <tim@oreilly.com>
Date: Mon, 16 Sep 2002 09:00:27 -0700

On 9/16/02 4:03 AM, "Simon Cozens" <simon@netthink.co.uk> wrote:


> 
> Just like the availability of closed source software discourages potential
> programmers from writing free software? No, I don't think so.
> 
> I can't think of any project whose documentation has been harmed by the
> existence of an O'Reilly book, and in fact the only books I can think of are
> for projects which have damned good documentation already.
> 
> I believe there's an evolution law of free software. Actually, I believe there
> are quite a few evolution laws of free software, but here's the only one I'm
> going to preach about in this sermon: if it's necessary to produce a free
> version of something, *someone will do it*.

Amen. I was going to reply to this thread, but Simon said it all.  And this
issue demonstrates as clearly as any the fundamental split between the
dogmatic and practical wing of the free software and open source movements.
If the goal is dogmatic (it must be free or I won't use it), then RMS is
right that proprietary commercial documentation is bad, but if the goal is
maximum utility and benefit for users, then he's wrong.  It's pretty clear
to me that the mixed model, in which fun things are free, and boring things
are non-free, is one that maximizes the existence of the things that people
care about.
  
I'll also point out that RMS is extremely inconsistent when it comes to the
reality of competition for his "free" documentation.  When one of the
authors of some of the GNU documentation wanted to do a project to improve
the GNU documentation and have it republished by O'Reilly (still under the
GPL), Richard objected strenuously, on the grounds that the sale of
documentation was one of the principal sources of FSF income.  I even
offered to pay a royalty to the FSF.  But we decided not to go ahead,
because my principles say that you honor the wishes of the creator of a
product, and either accept or reject the terms he or she offers for its use.
And if Richard doesn't want to honor his own expressed license, but instead
asks that we act as though the FSF manuals are to be published only the FSF,
I was willing to honor that instead.  But it has galled me ever since to
have him making pronouncements about how documentation should always be
free.
> 
> If you need a free Unix, someone will write a free Unix, AT&T or no AT&T. If
> you need a free documentation set, someone will write a free documentation
> set, O'Reilly book or no O'Reilly book. Why has one happened but not the
> other? Here's why I think it is:
> 
> For programmers, writing software is fun. It's actually enjoyable to produce a
> free version of something you "can't" have because it's proprietary. Why do
> people write free software? Because it's either necessary to have a
> free version, (whether pragmatically - we need a free version of this software
> because there isn't a commercial one we can actually use - or dogmatically -
> because we believe that everything should be free) or it's fun to do.
> 
> Documentation is different. For most programmers, writing documentation isn't
> fun. So the question is, is it necessary to have a free version? If there's a
> reasonably-priced book available in a shop, it's probably not pragmatically
> necessary. That only leaves dogmatic necessity. And that's pretty weak - the
> only people who are really interested in dogmatic necessity are those who
> spend more time playing politicians than they would writing documentation
> anyway. 


I agree completely.  See http://www.oreilly.com/ask_tim/orabooks_os.html for
something I wrote on the subject three or four years ago.
> 
> It's all very well for people to complain that there's no free documentation
> for project X, but if they cared that much about it, nobody's stopping them
> from solving it.

I'd go even further.  I've made standing offers to several free software
projects to give them subsets of some O'Reilly books to incorporate into
their documentation, but no one (e.g. FSF, xfree86) has ever taken me up on
it.  The only project where we've done this routinely is with Perl, because
the authors of the commercial books and the authors of the free
documentation are often the same people.  An update to the camel book and an
update to the perl docs generally happens in parallel.

Unless of course you count the projects where the commercial book (some of
our Linux books, Samba) *is* the free documentation.

-- 
Tim O'Reilly @ O'Reilly & Associates, Inc.
1005 Gravenstein Highway North, Sebastopol, CA 95472
1-707-829-0515 http://www.oreilly.com, http://tim.oreilly.com