[TowerTalk] Patentability (OT)

[Dick Green]

As someone else who has been through the patent submission process, I agree
with most of what Norm says -- it's not as easy to get a patent as you might
think. I especially agree that an examination of prior art is an important
part of the process. Unfortunately, the examiner may not always be as
familiar with the prior art as we would hope. The applicant is required to
divulge all known prior art, but I sometimes wonder how well this is
enforced.

BTW, the process is *very* expensive. It costs about $20,000 for each U.S.
patent (mostly attorney's fees), and as much as $200,000 more to broadly
protect the invention internationally (mostly translation fees and more
attorney's fees.)

I've seen a number of patents on technology that either seemed obvious to me
or were based on commonly-known prior art. However, a closer examination
usually reveals one or more very specific techniques that make the approach
unique. For example, my company is involved in developing technology for
Internet voting and has submitted four patent applications on related
cryptographic protocols and computational processes. When scanning the
Patent Office database, I saw  what appeared to be a patent on the basic
process -- i.e., using a remote computer network (like the Internet, which
wasn't specifically mentioned) to cast and tally votes. There were
absolutely no cryptographic protocols described, although the patent said
they would be required to make the system feasible (little did the author
and examiner know how difficult those protocols are!) I was astonished,
because the general concept of remote voting has been described in published
papers and discussions for nearly 20 years. The hard part has always been
the cryptographic theory. However, closer inspection revealed that all of
the claims relied on using specific authentication techniques to "positively
identify" the voter. The process required using such extreme techniques as
fingerprint scans for this. Although ideas like this had probably been
discussed informally, this patent offered up several specific implementation
techniques for positive authentication that I doubt have ever been fully
described in the specific context of remote voting. My guess is that the
examiner required that for approval. Otherwise, the patent would have been
too general and too much like prior art. It's possible, however, that the
patent could be successfully challenged for being based on relatively
obvious techniques and for relegating several crucial steps, like the
crytographic protocols and fingerprint scans, to "black boxes" that are not
described in any detail. Unfortunately for the inventor, not only are most
of his authentication schemes complex and costly to implement, but
provisions of the Motor Voter bill prohibit using such strong means of
identification because stringent proof of identity tests have been used in
the past to discriminate against persons wishing to register. The inventor
is from Jamaica, so he probably didn't know this.

Anyhow, I'm sure that Tom's patent is based on prior art, but probably has
one or more unique features or approaches that qualify it for patent
protection. It might have been sufficient that the technique is described
for a yagi (not a dipole, as in the K9AY patent) and uses more than two
frequencies.

However, I also feel that the Patent Office has issued some highly
questionable patents in recent years, particularly in the area of
information technology. Many of you have probably read about the patents
issued for "business processes" and fairly obvious software interface
techniques like Amazon's "One-Click". Although I don't think those things
should be patentable, Amazon's CEO did offer what I thought was a reasonable
compromise -- reduce the life of the patent protection to five years (I
think three years would be more reasonable in this day and age.) That way,
someone who is first to market with a good idea gets some level of
protection, but competitors are only excluded temporarily.

73, Dick WC1M




--
FAQ on WWW:               http://www.contesting.com/FAQ/towertalk
Submissions:              towertalk@contesting.com
Administrative requests:  towertalk-REQUEST@contesting.com
Problems:                 owner-towertalk@contesting.com