Re: revenue in a Free Software Industry

[Tom Lord]

Wow.  Gnu-misc-discuss is turning into software economics 101.
Listen up, kids!  (Really.)



       Also, the proprietary closed source model
       allows you to seize control of your user base
       and once they're locked in, you have to freedom
       to charge outrageously high prices and enjoy
       huge profit margins.

       Nowhere is this practice taken to the greatest
       extreme than the electronic design automation
       (EDA) industry where licenses can run into the
       tens or even hundreds of thousands of dollars
       per year per seat.  Once a semiconductor company
       chooses to go with an EDA vendor, and spends
       months designing its methodology around its software
       and validating the process, the company becomes
       very reluctant to change it unless there is
       a very compelling reason.

You've just defeated your own argument.

"Once a semiconductor company [...] spends months [I would say years]
designing its methodology around its software and validating the
process, the company becomes very reluctant to change it unless there
is a very compelling reason."  Therefore, that semiconductor has an
excellent reason to pay its EDA vendor a premium to sustain the
engineering process that produced the design software in the first
place, regardless of licensing.  Do you suppose the average chip
designer is smart enough to realize that?

Now if the EDA vendor gently switches from a proprietary to a Free
Software model, that "makes no never-mind" for their customers.  Those
customers are not really paying for the alleged "right to use" the
design software, but for the development process that created and
maintains that software.

There are, no doubt, degenerate cases.  There may be vendors who write
some software, fire the developers, and become little more than a
legal department that enforces a license.  What can we say about such
organizations other than they have neglected their own future?  For
example, design validation, the example you cited, is a moving target.
Those PhDs you can't figure out how to pay for (and the corporate
infrastructure that organizes them) are precisely the source of value
our hypothetical chip manufacturer wants to be paying for.

Chips are a funny case.  It seems to me, with my limited knowledge,
that chip designers will often want the advantage of hiring those PhDs
directly -- keeping their talents strictly "in house" in order to gain
a competitive advantage as technologies advance.  It's a fine line:
will the designers be better off if the EDA PhD's pool knowledge and
effort in an independent corporation or if they split up and compete?
As a total guess, I'd bet that the bigger designers simultaneously
persue both strategies.

-t