🚀 go-pugleaf

In article <7sLLeB3w164w@mantis.co.uk> mathew@mantis.co.uk (en font terrible) writes:
>mib@geech.gnu.ai.mit.edu (Michael I Bushnell) writes:
>> The purpose of the FSF is not to write software that everyone can use,
>> but to promote the creation of a world in which we avoid the
>> tremendous costs associated with the current methods of selling and
>> marketing commercial software.
>
>What tremendous costs?

I am a graduate student.  At the moment, I cannot even afford to buy
a computer.  Even if I could afford a computer, I couldn't afford the
extra $800 to $1000 to run Unix on it.  And then, most likely, the
essential things like the C compiler and libraries would have been
unbundled.  Even if I restrict my attention to MS-DOS, the "serious"
type of development software I need (compilers, libraries, utilities)
costs hundreds and hundreds of dollars before I have enough together
to constitute even a pathetic approximation of a usable system.  Go
to the kind of workstation class software that is considered the minimal
acceptable standard in computing science research, and I would be
looking at an investment of thousands of dollars, all in software that
would be obsolete within three years.

Perhaps those people who make a lot of money in the software industry,
or those whose employers shelter them from these expenses, do not see
this as a problem.  But consider this:  How do you make the transition
from being a "have not" to being a "have"?  In order to be a "have"
you would have to make a lot of money in the software industry, or
work for an employer who makes a lot of money.  In order to do either
of these, you must first develop your own programming skills.  It is
impossible to develop your programming skills in a vacuum--you need
access to Real Hardware and Real Software.  The limited perspective
on computing afforded by the widely available microcomputer operating
systems (MS-DOS, the mac OS, etc) produces at best brain-damaged
programmers.  The record of most university CS programs is dismal.
Thus, we have a chicken-and-egg problem:  You can't become a good
hacker unless you've got software, and you can't get software unless
you're a good hacker and hence employable.

Adequate hardware (to my mind: hardware that can provide multiple
large, protected address spaces) is for the first time in history
becoming available as a commodity basis.  For less than $2000 I can
get a 386 based system that has the necessary hardware to support
the kind of software I am interested in.  It is soon plausible that
almost anyone will be able to afford an adequate hardware platform
for serious computing.  The software, however, is another matter.
And without the software, the hardware is meaningless.

Now, I've been very fortunate in the opportunies I've had to develop
myself.  My family was able to send me to a decent college, which had
adequate hardware and software resources, particularly SOURCE CODE,
for me to take advantage of.  Working for the GNU project and later for
Bell Labs helped me improve my programming abilities in ways no CS
program or typical commercial software development shop could.

I would like to see every programmer have opportunities just as good
as mine.  Now, the benefits I've had came from three sources:

	1. Access to interesting, good quality source code.
	2. Access to good books.
	3. Access to talented people.

The GNU project is obviously directly addressing #1 and #2.  The quality
of GNU code is far from the best I can imagine, but friends of mine in
the software industry tell me it blows away a lot of commercial software.
At the very least, the GNU project will raise the least common denominator
of code quality that programmers are exposed to before they start inflicting
their creations on the world.  With more people being exposed to better
quality code, the number of truly talented people being developed will
hopefully increase, and hence people will have a higher probability of
being exposed to someone really talented.  I think the long term effect
could be self-reinforcing.

By a similar argument, the costs of commercial software development
and marketing transcend the merely monetary:

	1.  The high cost of software and nonavailability of source code
	cripples the development of programmers' abilities.
	2.  The crippled programmers write lousy code that needs major
	surgery to keep running, thus driving up the costs in (1).
	3.  The relative rarity of talented programmers further drives
	up the costs in (1).

Here we have a vicious spiral that could ultimately have the effect
of creating a self-perpetuating technological elite of "haves" vs.
"have nots".  I am against the creation of such an elite, because
its existence would be contrary to the principles of democracy.
Whether the system now is leading in that direction is obviously
a matter of speculation, but note that I am certainly not alone in
worrying about such a trend--many scientists and sociologists are
concerned about just this possibility.

Please note that I am not against selling software.  What I am against
is preposterous cost and the non-availability of source code.  I do
not think the commercial software industry is in any danger of
being destroyed by the free software movement.  I do hope, however,
that the competition will force the software industry to become
reasonable.  And finally, I expect the free software movement will
be a blessing to those of us who are simply not interested in
using typical commercial software--we will be able to build whatever
we want.

	Mike Haertel
	haertel@euclid.uoregon.edu