Re: FW: Results of "downloading compressed program images" request

[Alec Cawley <alec@cawley.demon.co.uk>]

In article <c=US%a=_%p=Systemix_Softwar%l=MAGGIE-980122145757Z-
36@maggie.systemix.com>, Brian Cuthie <brian@systemix.com> writes
>
>Now, I would argue the opposite (naturally ;-).  The way I see it, there
>are three options:
>
>1) Make everything freely available, the natural consequence being,
>IMHO,  that we all go out of business.

That does not seem to me to follow at all. The "technology" of making a
burger is hardly secret, yet McDonalds don't go out of business. It is
perfectly possible to run a legitimate busines manufacturing things
using publicly available designs and software and succeeding because of
the economies of scale inherent in most quantity production. I have
recently bought some configured Linux boxes at work. I could perfectly
well have bought Linux and "empty" PCs, but I have better things to do
with my time. The guy who sold them to me makes a respectable living
selling Linux plus add-ons, all of which he is happy supply as GNU
licensed.

>2) Patent everything, thus protecting intellectual property rights while
>publicly dicslosing the implementation details of everything.  The
>cosequences here are much more widespread than (1), since eventually it
>will become difficult to build *anything* without tripping over someone
>else's patents.  That software *can* be patented seems like a real
>stretch to me.


Patents expire - as is part of their design. When they do, you can use
them for your design as above. I could argue that the patent system has
a lifetime designed for a slower-moving world than today. Also, many
patents can be bypassed or invalidated - admittedly at a cost. I don't
think this approach is viable; the cost in lawyers fees would cripple a
business far faster than any competitors.

>3) Keep trade secrets secret.  This approach seems to have the fewest
>consequences.  The company that invested in a product's development gets
>to make a living selling it, while everyone else is free to reimplement
>it at will.


The "price" of GNU software is the GNU terns and conditions. If you
don't want to pay the price, no-one is forcing you to buy. But if you
don't buy, you can't complain about you haven't bought. And if the price
is too high, shop elsewhere.

Somebody, a few days ago, pointed out that embedded designs are very
different from PCs and PC plug-in hardware. He/she reckoned that this
made it pointless to insist on GPL/GPLL conditions. I disagree, for the
following reasons.

If nothing novel is revealed by publishing the software under those
condigions, nothing has been lost. I presume, therefore, that something
can be learned by studying the reluctantly-published source. If it was a
revolutionary idea, you would have patented it. So what is revealed is
good design, neat solutions, handy approaches. 

It seems to me that the publication of good hardware designs is as much
a public good as the publication of good software design. If a designer
has a greater library of proven solutions at his fingertips, he can
create better producs. He has to spend less time reinventing the wheel.

By insisting on GPL/GPLL for hardware designs, the FSF is promoting (to
the best of it's limited ability) the approach of the originator loosing
a small potential monopoly profit in order to achieve a much greater
public benefit into the hardware arena. It seems to me exactly in
accordance with the FSF philosopy and aims that it should do so. The
benefits, however large or small they may be, to those who chose to take
the proprietary path, are no concern to the FSF.

In fact, it seems to me that the barriers to copying in hardware are far
greater than in software. I can copy a complete software product with a
few keystrokes. To duplicate the hardware, I have to get a machine
readable circuit diagram, lay it out, get it manufactured, get it
populated, obtain sources for programmables....


-- 
Alec Cawley
Newbury
Berks, UK