Subject: RE: [openip] Re: "rights" and "freedoms"
From: "Ross N. Williams" <ross@rocksoft.com>
Date: Wed, 20 Oct 1999 00:47:39 +0930

At 11:29 PM +0900 19/10/99, Stephen Turnbull wrote:
>>>> From: Richard Stallman [mailto:rms@gnu.org]
>
>    >>> Subject: [openip] Re: "rights" and "freedoms"
>    >>
>    >>> The US Patent Office is incompetent in all fields, and issues
>    >>> trivial patents habitually.
>
>Easy for us to say....  Let's say doing a good job researching prior
>art takes a work week (5 days).  (Ross said "several days," and patent
>examiners are generally not specialists.)  With 500 software patents
>granted per week, the P.O. needs 100 "prosecuting examiners" just for
>the software patents granted.  How many do they have?

I vaguely remember that they hired a whole army of new ones to do
all the software patents. Actually, they MUST have. A few years ago,
it was only 5000 software patents a year.


>Still, it sounds like something practical that can be done now.  And
>the Patent Office ought to be happy about it, since fewer really dumb
>ones would get through, which makes them look bad.  This should up the
>quality of applications (and reduce quantity), allowing them to do a
>better job of review.

Have you got ANY idea how BORING this process is? It's SOOOOO BORING
busting patents. It's incredibly tedious, boring and frustrating
looking for some damn CONCRETE example of a "while loop with incremental
counter means". You remember some program in the late 1980s that
did something like it and you go looking and can't find it and it's
horrible and I wouldn't wish it on anyone.

So before everyone gets too excited about this idea, please have a
think about what's actually involved. Every TWENTY MINUTES. :-)
I like my Patentbusters Inc idea better. At least the poor souls
who go busting will get paid. Presumably getting paid is the only way
the patent examiners can stand it too.

Note: If software patents is a "profit centre" for the USPTO, they
may not actually be too happy about the process getting tighter.


>And the anti-racketeer brigade can concentrate on the ones that look
>most vulnerable, which should help to bring down the amount of time
>needed to research the prior art significantly.  (How much of that
>"several days" is research and how much preparation of the
>submission?)

1 day trying to understand the convoluted language of the patent.
4 days trying to contact some person you remember did something like
  this ten years before.
4 days trying to track down some ancient instruction manual that the
  person refered you to that's not even available on eBay.
1 day writing the submission.
9 days going and doing it all again when your patent attorney points
  out some trivial, but critical point that you missed.
  "no, it has to be a FLOATING POINT incremental means".

Ross.

Dr Ross N. Williams (ross@rocksoft.com), +61 8 8232-6262 (fax-6264).
Director, Rocksoft Pty Ltd, Adelaide, Australia: http://www.rocksoft.com/ 
Protect your files with Veracity data integrity: http://www.veracity.com/