Path: utzoo!mnetor!uunet!husc6!hao!ames!elroy!devvax!lwall
From: lwall@devvax.JPL.NASA.GOV (Larry Wall)
Newsgroups: comp.sources.d
Subject: Re: `free' software
Message-ID: <992@devvax.JPL.NASA.GOV>
Date: 12 Jan 88 23:58:51 GMT
References: <14224@think.UUCP> <9983@mimsy.UUCP>
Reply-To: lwall@devvax.JPL.NASA.GOV (Larry Wall)
Organization: Jet Propulsion Laboratory, Pasadena, CA.
Lines: 71
Summary: Larry Wall bares his soul.

In article <9983@mimsy.UUCP> chris@mimsy.UUCP (Chris Torek) writes:
: >In article <2377@dasys1.UUCP> samperi@dasys1.UUCP (Dominick Samperi) writes:
: >]Question: What motivates a programmer to develop a large complex software
: >]system, perhaps requiring many weeks (or even years) of development effort,
: >]which is then posted for the world to use, for free?
: 
: (Why do I do crossword puzzles?)
: 
: In article <14224@think.UUCP> rlk@think.COM (Robert Krawitz) writes:
: >Perhaps a desire to help others?
: 
: Or for my own reasons: the challenge of solving a puzzle; the desire
: to be famous; and probably most important, a neurotic need to prove
: myself :-) .

I suppose I can use myself for an example.

All of the reasons mentioned above play a part, but I feel like they all miss
the point slightly.

At the start of any project, I'm programming primarily to please myself.  (The
two chief virtues in a programmer are laziness and impatience.)  After a while
somebody looks over my shoulder and says, "That's neat.  It'd be neater if it
did such-and-so."  So the thing gets neater.  Pretty soon (a year or two) I
have an rn, a warp, a patch, or a perl.  One of these years I'll have a
metaconfig.

I then say to myself, "I don't want my life's work to die when this computer
is scrapped, so I should let some other people use this.  If I ask my
company to sell this, it'll never see the light of day, and nobody would pay
much for it anyway.  If I sell it myself, I'll be in trouble with my company,
to whom I signed my life away when I was hired.  If I give it away, I can
pretend it was worthless in the first place, so my company won't care.
In any event, it's easier to ask forgiveness than permission."

So a freely distributable program is born.

At this point I'm no longer working for a company that makes me sign my life
away, but by now I'm in the habit.  Besides, I still harbor the deep-down
suspicion that nobody would pay money for what I write, since most of it
just helps you do something better that you could already do some other way.
How much money would you personally pay to upgrade from readnews to rn?
How much money would you pay for the patch program?  As for warp, it's a mere
game.  And anything you can do with perl you can eventually do with an amazing
and totally unreadable conglomeration of awk, sed, sh and C.

It's not so much that people don't value the programs after they have them--they
do value them.  But they're not the sort of thing that would ever catch on
if they had to overcome the marketing barrier.  (I don't yet know if perl will
catch on at all--I'm worried enough about it that I specifically included an
awk-to-perl translator just to help it catch on.)  Maybe it's all just an
inferiority complex.  Or maybe I don't like to be mercenary.

So I guess I'd say that the reason some software comes free is that the
mechanism for selling it is missing, either from the work environment, or
from the heart of the programmer.

What programmers like me need is a benefactor, like the old composers and
artists used to have.  Anybody want to support me while I make beautiful things?
My hope is that some billionaire who reads the net for pleasure(?) will
someday say "I'd like to pay you for all the people who have used rn over
the years..." and drop $1,000,000 or so on me so I could live off the interest
and finish the new rn.   :-)

If you are a benefactor, my address is 9132 Kester Ave, Panorama City, CA 91402,
and my phone number is (818) 893-8182.           only 1/2 :-)

If you aren't a benefactor, that's ok, I really expect my rewards in heaven.

Larry Wall
lwall@jpl-devvax.jpl.nasa.gov