|
|
|
| Authored by: halfhuman on Thursday, June 14 2012 @ 03:38 AM EDT |
I see your point: you have received a clear answer. However, I do not believe
that M Risch has clearly acknowledged the import of your question and that he
has failed to indicate explicitly in which way he believes he has answered it
(well, without the first he can't very well do the second of these).
I believe there is every danger that this conversation will now be used as a
source of FUD and trolling. It will be said, perhaps not by prof Risch himself,
that on Groklaw this or that remark in support of a mealy-mouthed position on
software patents was made. That the very possibility that software is a process
has had some acceptance here. That PJ respects M Risch's position (as distinct
from respecting him as a person).
I believe your earlier articles indicated very clearly that software is data.
The machine reads the data and processes it. Sometimes this data modifies the
OS, sometimes it is merely regurgitated slightly changed or even unchanged.
Sometimes it is completely ignored. None of these make a new machine, none of
these is even a process. Your analysis of the CPU and its basic cycle clarified
these things completely.
It follows that any patent on software is a patent on some aspect of data. Prof
Risch, if he is as competent in computer science as his claim of reading so much
of it seems to indicate, must be able to understand the previous paragraph. But
then he must clearly indicate whether he thinks it is false. If so, he adheres
to the fictional computer science you described. If not, he must indicate why it
is irrelevant. This he has NOT done.
Until he does, I will regard his approach to PJ and Groklaw as purely
self-serving and perhaps cynical.[ Reply to This | Parent | # ]
|
| |
| Authored by: RMAC9.5 on Saturday, June 16 2012 @ 01:26 PM EDT |
PolR said Also Pr Risch has indicated he is doesn't mind stopping
writing software patents
as machine patents based on the facts of computer
science. He believes this
would not matter in practice because the same patent
could be drawn to a
process. It is just a matter of writing the right magic
words in the claim.
If I understand process patents correctly, they
are limited to the process being patented. Thus it should be possible to
reuse the "patented idea" for a different software process. This is in stark
contrast to software machine patents which give unlimited protection to the
"patented idea" on general purpose computers.
PolR said
That may be true on the long run, but from a short term perspective
there are a
lot of existing patents which are drawn in machine form. If we can
get a ruling
that ban them, all these patents will instantly be invalidated but
they will
remain in the USPTO database and be available as prior art. From
pour
perspective, this will reduce the magnitude of the problem. We are making
some
progress here. I agree. Banning software patents
entirely is probably impossible. Too much water[/money] has already
passed under the bridge[/been spent] for the current big corporation good guys
like Google and IBM to give up their existing [defensive] software patent
portfolios [and no one else has the political muscle to make it
happen].
IMHO, our best bet is to concentrate on limiting software
patents to "process type patents" that allow for some competition, making sure
that the processes being protected are narrowly defined, and insisting that
source code be provided so that "persons ordinarily skilled in the art" can
understand and benefit from what is being patented. [ Reply to This | Parent | # ]
|
|
|
|
|
