[NTLUG:Discuss] Free vs $$$ software - was: RMS's Speach

[Christopher Browne]

On Sat, 03 Feb 2001 12:51:16 CST, the world broke into rejoicing as
Greg Edwards <greg at nas-inet.com>  said:
> Being a professional Software Engineer as well as a proponent of
> Linux/Open Source this subject has made for some mixed feelings with
> me.  I think there is plenty of room for both software that is free
> and software that costs users real dollars.  Actually in my opinion,
> if it weren't for non free software we wouldn't have systems
> anywhere near as sophisticated or stable as we have today.

It is not manifestly evident that "nonfree software" has contributed
_vastly_ to the stability of sophisticated systems.

VMS would probably be the characteristic example of a stable OS that
is distinctly proprietary; it is getting decreasingly important,
commercially, and was _always_ primarily tied to sales of Digital
hardware.

In contrast, without nonfree software, we probably wouldn't have
suffered from Melissa, ILoveYou, and many of the "stack-smashing"
attacks to which systems like NT have been prone.

> I also think that the tools necessary to provide communications
> between systems should be included.  Software such as web servers,
> network servers, mail servers, DNS, ftp, browsers, etc., are
> standard equipment type items.

The people that sell Netscape Commerce Server or Zeus would disagree,
as would the folks that sell WebSphere, and, um, what's BEA's web
server?  Novell would disagree with you on the network server side
too.

> Basically anything that is part of the infrastructure should be
> standard equipment and provided through OEM distribution channels.
> Anything that is part of making the use of a computer specific to
> the owners choice is optional and therefore (I think) open to add on
> costs.

Your _thought_ here is not without merit, but the idea suffers from
much the same problem as attempts to define what "bad stuff" ought to
be censored.  

We may be able to agree, albeit fairly loosely, on some of the gross
features of this, but try to define "infrastructure" or "owner's
choice" (or "obscenity") in a way that can be objectively administered
by some form of bureaucracy, and it gets problematic.

What's "infrastructure" and what's "user's choice" (or "obscenity"
versus "art" versus "medical/psychological analysis") is very much in
the eye of the beholder.

> The point that Stallman and others that tout purely free software
> miss is that without a financial incentive software would be built
> by hobbiests only.  I'm not trying to knock the contributions of the
> non professional programmers and/or hackers to the world of
> computing and the open source movement, but that contribution is
> only a drop in a very large bucket.

I don't think this is a point that Stallman misses, at all.  I think
he has put quite a lot of thought into such issues over the years.

It's not that he misses this point; he made it reasonably clear in his
lecture that he doesn't very much _care_ if people make livings
building software.  That's not just someone in an ivory tower not
needing to care; when he got the FSF started, he quit a presumably
reasonably well-paying job at the MIT AI lab, and apparently lives on
quite limited means.  [Presumably other people largely pay his travel
expenses?]

> The bottom line is that if nobody paid for software then nobody
> would pay people like me to do R&D or software development.  If
> nobody would pay people like me then there'd be no technology
> revolution.  If there were no technology revolution then RMS
> wouldn't have anything to talk about:)

But a big chunk of the question is of _how_ the software development
gets paid for.

The bulk of it results from companies paying people to build internal
custom applications, "bespoke software" being the formal term.
"Bespoke software" is really only _useful_ in the context of the
enterprise for which it was designed; the precise licensing of it is
of limited importance as the resulting software isn't normally
terribly useful elsewhere.

The use of the GPL is fairly compatible with construction of bespoke
software; such software never gets redistributed, so the GPL never
gets a chance to have effects people would consider "scary."

If all software had to be free, there _might_ be less R&D work done
than is done now.  Certainly not zero; any time the expected return on
investment was positive, projects would be expected to get funded.

On the other hand, it is arguable that there might even be _MORE_ R&D
work done than is done now.  After all, the present circumstances
involve companies sending Microsoft and Oracle and other big licensors
lots of money to pay license fees.

Consider that lots of people have embedded systems for which it would
be sort of nice to use something like QNX.  QNX licensing fees are
fairly steep, and hence people may decide that it is too expensive to
explore its use.  

If it were free software, with _no_ licensing fee, then license fees
disappear as an economic barrier.

If you don't need to pay for development tools, or DBMS licensing, or
web server software, or OS software, or [fill in other
potentially-punitive licensing fee], lots of projects become more
economically viable.

I think RMS is entirely too pessimistic in assuming that there would
be fewer opportunities; he didn't read Marshall and Eric McLuhan's
_Laws of Media_ which indicate that new artefacts not only in the
"obsolescence" of some things, which everyone seems to recognize, but
rather has _four_ effects:

1. Intensification.  Something gets intensified.  In particular, the
elimination of the barrier of licensing fees allows some development
activities to _increase_.

2. Obsolesces.  Obviously those enterprises that forcibly depended on
licensing fees for income become somewhat obsolete.  That hurts Oracle
_very_ badly, Microsoft _quite_ badly, and other software houses more
or less so.

But note that the GPL does not claim to be an exclusive license;
observe that Troll Tech, makers of Qt, have released that library
under the GPL, allowing its free use for free software, but _also_
license it under another license so that if you wish to use it to
build non-free software, they can receive licensing fees.  So 'tis not
obvious that licensing fees necessarily go away.

3. Retrieves.  A new development may make useful again things that had
been obsolete.  I'm not sure what to associate with this other than a
rewording of what gets intensified...

4. Reverses Into.  When taken to an extreme, an artefact tends to
reverse some of its effects.  Classic case would be that having the
ability to generate custom type fonts allows building more
coherent/elegant looking documents; when Wired Magazines presses that
to an extreme, coherence starts to suffer...  I'm not sure yet what
free software reverses into when taken to an extreme...

--
(reverse (concatenate 'string "gro.gultn@" "enworbbc"))
http://www.ntlug.org/~cbbrowne/sgml.html
`When you  say "I wrote a  program that crashed  Windows", people just
stare at you  blankly and say "Hey, I got those  with the system, *for
free*".'  -- Linus Torvalds