Re: Can we get patents?

> That depends; is the SFLC offering to pay for the patent applications? Last
> I
> checked, it was somewhere around $6000 per patent.

Nolo press (www.nolo.com) sells a book on patents. Many people file their own
patent applications successfully. The cost is less that $1000.

David

I've talked to a friend of mine who is a patent lawyer.

1) in Europe if it is in the public domain then it cannot be patented
2) in North America you would have to patent before submitting to the
project.
3) His question was why? With a bsd license you can't stop anyone from
using it and nobody
else can patent it since by placing it in the project you are
establishing prior art.

Dave

David Walker wrote:

>>That depends; is the SFLC offering to pay for the patent applications? Last
>>I
>>checked, it was somewhere around $6000 per patent.
>>
>>
>
>Nolo press (www.nolo.com) sells a book on patents. Many people file their own
>patent applications successfully. The cost is less that $1000.
>
>David
>
>
>
>---------------------------(end of broadcast)---------------------------
>TIP 8: explain analyze is your friend
>
>
>
>

--
Dave Cramer
http://www.postgresintl.com
519 939 0336
ICQ#14675561

On Tue, May 10, 2005 at 16:57:01 -0400,
Dave Cramer <pg(at)fastcrypt(dot)com> wrote:
>
> 3) His question was why? With a bsd license you can't stop anyone from
> using it and nobody
> else can patent it since by placing it in the project you are
> establishing prior art.

Nope. They can still be issued a patent and then you will have to come
up with some big bucks to get it overturned. The Patent Office isn't
going to go looking through the Postgres source when checking for prior
art. Even if you got the method published in a journal, the Patent Office
could still not see that the paper was the same technique as was being
used in the patent.

Another difficulty with establishing prior art to prevent others from
obtaining patents is that different inventors and different patent
agents or attorneys use different terminology to describe the same or
similar inventions.

To use a simple mechanical example:

- Alex develops gadget that includes a "... coupling
comprising two pipes aligned end-to-end with a
washer in between and a clamp that grasps both pipes
... " Alex publishes and/or patents her gadget.

- Subsequently, Beth develops a similar gadget that
includes "... a link comprising a plurality of hollow
cylinders, each adjacent cylinder linked to the next by
a circular gasket, the adjacent cylinders being
functionally attached to each other ..."

Are these inventions the same? Is Beth's invention obvious in light
of Alex's? They certainly appear to be the same, but without knowing
the details of the patents and their prosecution history, there's no
way to know. If a patent examiner working on Beth's case relied on
word searches - or was just working too quickly - it is likely that
Alex's would not show up in the search or not be carefully
considered; and Beth's application might grant as a patent.

Now consider complex software patents. The same software function can
be described in a myriad of ways. It is quite possible for the first
inventor to establish prior art that SHOULD block another application
for the same or a similar invention, but nevertheless the second
application is granted as a patent.

If the second inventor tries to enforce the patent it SHOULD be
declared invalid in court; but no one wants to be accused of patent
infringement, forced to pay a fortune in legal fees, and dragged into
Federal Court just to prove a point. One advantage of the first
inventor actually patenting the invention, rather than just publishing
it, is that then at least the first inventor can threaten to counter
sue, and perhaps reach an quick settlement.

Other than increase the price of applying for a patent (again) and
hiring more and better examiners, I don't know the solution to this
problem.

Richard Tanzer