Re: Two reasons why software patents are bad

On Fri, 24 Sep 2004 18:23:00 -0400 (EDT), Stuart D. Gathman
<stuart(_at_)bmsi(_dot_)com> wrote:
> On Fri, 24 Sep 2004, Michael Hammer wrote:
>
> Mostly it is a difference of speed and a matter of legality.
>
> Speed:
>
> There might be 2 or 3 people inventing a hardware idea independently of one
> another.  For example, the telephone (Bell and Gray).  The first one to the
> patent office wins - but now the idea is published for engineers everywhere.
>
> There are hundreds to thousands of developers arriving at the same software
> idea simultaneously, and thousands of new ideas every year.  There is no need
> to publish the ideas because just about everyone has already thought of it or
> talked to someone else who has.  There are way to many such ideas to justify
> the expense of taking out a patent on them - unless you are an unethical corp
> out to exploit people.  You just wait for a good invention that
> everyone uses, then be the first to the patent office.  Now you've got
> them over a barrel and can extort money from them.
>
> Here is an example of a new original idea that came to me last night
> in the course of my work.  I needed to install squid as a reverse web proxy.
> External users need to see squid when they brwose the domain, but squid
> needs to see the real web server.  Recommended solutions are: split DNS
> or override DNS with /etc/hosts for squid.  I came up with a third solution:
> Keep the same IP for both, but use NAT to make that IP go to different
> places depending on the direction of the connection.  Pretty slick, eh?
> I'm sure I'm not the first to think of it, but I thought of it all by
> myself.
>
> Well, the only way for me to recoup $100 for a patent application on
> an idea like that is to promote it until everyone depends on doing reverse
> proxy with NAT, then reveal my submarine patent and start requiring
> licenses.  Since a corp is supposed to make money for its stockholders,
> software patents practically *demand* that a corp use them expoitively.
>
> The only way around this for corps trying to be ethical (IBM) is use their
> software patent portfolio defensively - i.e. to block exploitation
> by unethical corps.  But even that leaves individuals like you and me out in
> the cold with no defense.  Just about every new and original idea you
> or I have ever had while programming - has been had by thousands of
> other programmers - and is probably patented by someone.  All we
> can do is keep a low profile and try to stay under the radar of
> the unethical corps - and work to abolish software patents.
>
> There are, of course, a very few software patents that were actually
> unique - RSA, for instance, was not a widely thought of idea.  But they
> are very rare.  LZW was mentioned, but compression schemes using nearly
> identical principles were invented by many, many people - some of which I know
> personally.  The only thing unique about LZW was that it was the particular 
> way
> you had to do it to create GIF files that web browsers could read.
>
> Legality (IANAL):
>
> Software patents were never specifically authorized by law.
> Software and ideas could not be patented at all until the early '90s, when a
> landmark case decided that a chemical process patent could not be denied 
> merely
> because it used software.  That very reasonable decision was quickly twisted 
> by
> crooked lawyers into a case law basis for patenting software itself, and
> finally the ultimate banality of patenting business processes at the start
> of this century.
>
> Software patents are yet another example of the courts making the law.
> It is ridiculous that Congress would have to enact yet another law
> to specifically prohibit something that was never authorized in the
> first place, but short of impeaching activist judges, that is the only way to
> slow judicial tyranny.  Of course, with powerful lobbyists like M$ pushing
> it, they might just go head and actually authorize software patents.
>
> At least the Europeans are honest enough to realize that they need to
> pass a law first before they can patent software.
>
> --
>              Stuart D. Gathman <stuart(_at_)bmsi(_dot_)com>
>    Business Management Systems Inc.  Phone: 703 591-0911 Fax: 703 591-6154

So what's the difference between your idea as software or your idea
encoded on a silicon chip? By your logic the latter should be
patentable and the former not.

The real problem is that patents are being given for things (software)
which are not really inventions.

BTW, your approach with the snat is something that load balancer
vendors have been doing for years (for example F5 and Nortel/Alteon).

Mike