"rights" and "freedoms"

[Ben_Tilly at trepp.com]

L. Peter Deutsch wrote:
> > More than that.  Something like 90% of programmers are working on software
> > for internal company use only.  How is free software going to take away
> > these jobs?
>
> By reducing the amount of effort required to work around the problems of the
> black-box environments in which those programmers are having to solve their
> problems today, by reducing the number of those problems in the first place,
> and eventually by producing better environments.  It won't take away all of
> them, but it will take away some of them.  Also, as FS/OSS catches on, some
> of that software that would be useful to others will become openly
> available, reducing the "reinventing the wheel" that goes on now and further
> reducing the need for programmers.
>
I cannot disagree more strongly.  Free software decreases the marginal cost
of additional software.  I think that everyone is agreed on that.  However once
that marginal cost is reduced, the overall demand goes up.

Compare the cost of computing power today and 20 years ago.  There is no
question that 20 years ago computing power was tremendously more
expensive than it is today.  Yet companies spend more today on computing
power.  The availability of very cheap computing power means that today we
think nothing of wasting large amounts of computing power on calculating
which pixels should be part of a small pointing device or a decorative border.

> To the extent that FS/OSS makes good software available more widely and more
> cheaply, it will put programmers out of work or at least cut their incomes.
> The claims that lowering the cost of software will result in more software
> being produced are just that -- claims.  There is no evidence to back them
> up, only general economic hand-waving.  If it results in more units of the

Here is evidence, not drawn from hardware but from software.

The widespread availability of RAD environments for "business productivity"
software has allowed unknowledgable people to be applied to producing
customized software.  The cost of producing mediocre business applications
has gone down tremendously over the last 10 years.  Yet salaries of people
making this software has gone up.  I have heard estimates that the average
VB programmer in NYC makes $70,000/year.  And there are a lot more of
them today than there were 10 years ago.

OK, so VB is not free.  But it is widely available and has lowered the marginal
cost of producing simple applications.  Yet its effect has been to employ
more people at higher wages.

> same piece of software being bought (at a lower price), that doesn't produce
> more software.  It might increase the productivity of the *buyers*, but that
> would require evidence that use of software increases business productivity,
> and from what little I know about this topic, the evidence is mixed.
>
Increases in business productivity are not needed.  IT would be if the world
was rational, but it is not.  It depends upon the tendancy of managers to
want to do new things with software.  Some of the new things are good
ideas, most are not.  Businesses are willing to pay for them though...

> It still floors me that the FS movement effectively values the activities of
> distributors (who are entitled to be paid by the unit for providing FS) more
> highly than those of programmers (who aren't).  But maybe that is an
> accurate reflection of the value perceived by the user.

The FS movement should eventually wind up with Cheapbytes and the
like being major distributers.  They are producing and selling a
commodity and know it.  Companies like Red Hat temporarily enjoy a
very nice niche, but in the long run I would expect their money to have
to come from providing support and services.  I don't think that we yet
know what this model will look like.

Sincerely,
Ben Tilly