From kouhia@nic.funet.fi Sun, 13 Feb 2000 17:12:33 +0200
Date: Sun, 13 Feb 2000 17:12:33 +0200
From: Juhana Sadeharju kouhia@nic.funet.fi
Subject: [Patents] Patent problem

[ I have no idea what would be a correct place to discuss about these
but here goes... ]

Hello. I have a problem in GNU audio software development.

Background:

When playing audiofiles in real-time, we should have a fast start, i.e.,
audio should be heard as soon as possible from a key press.
This is achieved by keeping starts of the audiofiles in memory
while rest of the audiofiles are read from the disk.

I have as early as 1990-91 "invented" that audiofiles could be accessed
via an audiofile virtual memory system which handles also the prefetching,
buffer pools, ring buffers etc. both for random access and sequential
access. This also included the idea of keeping the beginning of audiofiles
in memory for fast start. (We even wrote a real-time synth application
at 1995 when we got first Linux, but due poor programming skills we had
to leave the virtual memory away from that version.)


Problem:

There are patents
  5811706  May 1997
and
  6008446  Jun 1998  (continuation of the above)
which states: "The software routine stores a first part of each waveform
signal in a sample pool of host memory and provides remaining portitions
of selected musical sounds from the hard drive to a stream cell array
without an audio perceivable delay."

Consequence is that one of free software authors refused to write
a software with direct prefetching. (He is looking for an alternative
from Linux's virtual memory system.)

 -*-

So, I would like to know have I lost my right to use my own "invent"
which I have choosed not to patent at the time when I "invented" it?

What if I write a free software with my original ideas and they
sue me? How do I prove that I have "invented" anything before they?
Naturally, I talked about these "invents" with my friends but never
prepared for the situation that somebody could patent the ideas later.
(And boy, do my friends remember my talkings this later?)

 -*-

Even now I don't know how to prepare for this kind of situations!
Richard Stallman once said that
 >The main thing you can do to stop someone else from getting a patent
 >on this technique is to publish it.  If you need advice on what
 >legally constitutes publication, I can ask a lawyer to give you advice.

I'm pretty sure no publication will take half-baked descriptions about
these "invents". I'm not even trying to get anything published -- only
what I need is that these "invents" are public/free.

 -*-

I'm quite disappointed that you rave against software patents while
you really should make sure we don't _loose our rights_ for "invents"
because patent business later steps on our foots.

The patent business is strongly biased because no law enforces me to
do anything for "invents" which are not patented. The patent business
sure takes advantages of this fact! (No, neither they did ask me if those
ideas are already in public domain.)

Please give me instructions how to prevent this kind of situation in future.
I have a few ideas which might end up later as patents. Please be more
specific than R. Stallman --- something concrete is needed.

 -*-

Ok. Writing a free audio software is what I just have to do. I will
implement my original ideas as I have planned them. I just don't like
to give up my rights.

Sincerely,

Juhana




