Patents - An Alternative View (Groklaw) [LWN.net]

Patents - An Alternative View (Groklaw)

Posted Oct 11, 2004 22:40 UTC (Mon) by iabervon (subscriber, #722)
In reply to: Patents - An Alternative View (Groklaw) by job
Parent article: Patents - An Alternative View (Groklaw)

Actually implementing mpeg4 decoding? Invention.

Right, so if you build a blueprint for this invention, which is what pseudocode for it is, you can submit this to the patent office for patent protection, because that's how you protect inventions. If you used a copyright on the pseudocode, this wouldn't cover implementations in actual languages, which would be independant works, as the similarities between the pseudocode version and the runnable version are determined by conformance to a standard (c.f. the limitations on copyright which protect Linux against claims that ELF implementations are derived works of the ELF specification).

The standard for all patent descriptions is that they be sufficient for a person knowledgable in the field to be able to produce a working version from the description. This is essentially the pseudocode for making a toilet; if you know how to do the usual things (like a computer), you could build the invention without needing any insight into the particular problem.

We have patents not primarily because it is difficult to keep secrets (although that is a factor) but because it is beneficial to society for people not to keep secrets, as others can learn from the invention.

(Log in to post comments)

Patents - An Alternative View (Groklaw)

Posted Oct 13, 2004 13:51 UTC (Wed) by mly (guest, #2171) [Link]

> If you used a copyright on the pseudocode, this wouldn't
> cover implementations in actual languages, which would
> be independant works

Are you sure about that? Can you point me to any particular
court case?

Concerning books, it's certainly clear that translation of
a text to another language doesn't create an independent
work, but a derivate, even if grammer rules etc are very
different in the languages involved.

Just try to write a book (or a computer program for that
matter) in any language where a young Magician called Harry
Potter appears in any kind of context, and you will see that
copyright law has a fairly long reach. Even if you change
the names of all involved characters, you'll soon run like
a fox to avoid Rawlings pack of lawyers if you are reusing
a substantial amount of her ideas in a text you claim as yours...

I don't think that describing something like a particular plot,
or concepts like "the butler was the killer" is enough to
protect all possible implementations with copyright, but this
is a matter of the judgement of the court in each actual case.
How far a pseudocode description will reach isn't clear cut as
far as I understand.

Patents - An Alternative View (Groklaw)

Posted Oct 13, 2004 22:29 UTC (Wed) by iabervon (subscriber, #722) [Link]

Most algorithms are described by copyrighted pseudocode in textbooks, and this doesn't seem to affect use of the algorithms. As far as I know, nobody has tried suing someone over an implementation of an algorithm explicated with copyrighted pseudocode, and I'd guess that someone would have tried, and I'd have heard, if it were at all plausible.

I think it would be like publishing a plot summary for a book, and then suing anyone who actually wrote a book that followed the outline. Pseudocode really glosses over a lot of details which are a substantial portion of the effort, and it's a far freer translation into an actual language from pseudocode than any translation of human language text into other human languages as a derived work is.

What tends to get people writing books or drawing cartoons is copyright on characters, not copyright on the actual text. This allows an author who has sufficiently developed a character to prohibit the same character appearing
in text of another author's entire invention. I doubt that code really has anything approximating a developed character, so this isn't a concern. (Furthermore, there tend to be trademarks involved as well, which are even less relevant to code).

Patents - An Alternative View (Groklaw)

Posted Oct 21, 2004 16:54 UTC (Thu) by job (guest, #670) [Link]

Almost. You are wrong in assuming that copyright is only a protection of
verbatim copying. It would be ridicolous if no one thought it to be the
same work anymore just because someone changed the syntax or punctuation
slightly.

Ask a book publisher what they would do if you did an inofficial
translation and sold on. You can't even tell the same story as another
author with your own words! But you are free to re-use ideas and themes
as you see fit. The same is true for computer programs, at least in those
parts of the world that hasn't allowed idea patents yet.

If you apply your description of a patent to the software field, it would
be a good enough description of something so a skilled programmer could
implement it. That's everything! Computers are Turing complete things so
they can by definition do anything. It is not the same in the physical
world where there are lots of practical problems to solve in order to
make something work! (Before you say "but that's true in software as
well!", remember that *when* you've done that much, you've effectively
written a copyrightable work!)