🚀 go-pugleaf

In article <4893@network.ucsd.edu> mbk@jacobi.ucsd.edu writes:

>A ham radio reciever could conceivably infringe on maybe 5 circuit design
>patents, at most perhaps.  There are only a few different things that
>a radio reciever can possibly do, and in each submodule, most of the circuit
>topologies will be mutually exclusive.

When I was  a young electronics engineer and radio receivers were still
constructed from vacuum tubes, a good amateur radio receiver could
easily contain two dozen miniature (7 or 9 pin) tubes.  A simple way to
"bias" one of the electrodes with respect to another was to insert a
small resistor between the cathode and ground.  Current flowing through
this resistor raised the DC voltage of the cathode a few volts positive.

At the time, RCA held a patent on this "cathode bias" technique.  They
were entitled to a royalty on every USE of cathode bias.  In principle,
20 usages could bring 20 royalties.

>Contrast to computer software:  the possible spaces of programs is far
>larger than that of all radio recievers (but not necessarily all electronic
>circuits), and more importantly, is composed of hundreds or even thousands
>of potentially patentable parts or methods.  I mean, when you make a ham
>radio, do you have to research all possible patents on the possible ways
>to manufacture transistors of a certain type, the patents on the processes
>needed to extract the raw materials used in your solder?  If you make
>ham radio recievers, you look up patents on "ham radio recievers", naturally.

And this is just one piece of the topology of a radio receiver.  Every
time a ham "home-brewed" a receiver using this technique then sold the
thing because he/she tired of it or wanted something new, he/she
infringed upon the  patent.  Such a patent was not restricted to "ham
radio receivers" but was covered by a much larger  universe -- that of
"vacuum tube circuits."

>Now, when you write a software program that has many parts to it, what
>subject do you look up for patents?  There are conceivably thousands---because
>you create almost all of the program by yourself.

This is a good question.  Would the second inventor  of the tube biasing
circuit above have thought to call  it  "cathode bias?"  How would  one
research this before  the days of  computerized indexes?!

>  This is the type
>of burden that would be placed upon programmers were software patents of
>commonly used procedures to become widespread.   It is because software
>is frequently created and recreated from nothing, with essentially
>only intellectual powers that makes it inherently different from other
>endeavors that are presently patentable.  (In analogy with the ham radio,
>I see nothing wrong with charging royalties for the use of commericial
>libraries, just as you have to buy the transistors in the circuit)

>This is the same argument as to why mathematics is not patentable.  Any
>mathemetician can think up some new mathematical structure and prove things
>about it.  Now, suppose that theorems and methods were patentable:  now Mr.
>Math would have to pay anytime he used say a diagonal counting method
>(originally discovered in the 19th century) in a proof, or Cauchy's integral
>formula, or a Fourier Transform.  (All of these are in my opinion utterly
>more important than any software patent).  And all engineering work that
>depended on those theorems would have to pay off the inventors. 


>If software is patentable, then why isn't mathematics?  There is no
>real distinction.

Isn't it true that mathematics is (was) thought to have always existed,
much as the physical universe described by physics has "always existed"
(or evolved)?  I believe that mathematics was forbidden from  patenting
because it was thought that mathematics is "discovered" and not
invented and that it was created by a Higher Being.

>The last reason I oppose software patents is sociological.  Much of the
>standard practice of engineering and science and mathematics is based upon
>the published and unpatented work of academic and government scientists.
>In comparison, much of the standard practice of programming and engineering
>(I don't mean "computer science"---which is often math) comes from a vigorous
>commercial sector, which is more likely to attempt to file patents.  
>Scientific & Engineering companies of course do file patents, but in
>general for much more focused and specific applications than for software.
>So the overall cost to society does not overwhelm the benefit to the firm.

You can get yourself into a box here.  In many, many cases, something
done in software can also be done in hardware.  (This may be the basis
for the RSA patent.)  So while I may have trouble getting a patent for
an algorithm in software, I'll build a chip that does the same, novel,
thing and get a patent for the "engine."  Then, I'll use the patent to
prevent anoyne else form doing the same thing, even in software.

-- 
_Brinton Cooper		BRL - Where "Research" is our Middle Name.