1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw) [LWN.net]

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 26, 2011 17:53 UTC (Tue) by bjacob (guest, #58566) [Link]

Great article! I fully agree with this, and that has been a common lunch discussion topic among mathematicians for some time. Software patents could stifle pure mathematics research, if they were consistently enforced.

This article is another reason why I'll miss Groklaw, and another example of how Groklaw is about much more than just the SCO case.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 26, 2011 20:36 UTC (Tue) by ccchips (subscriber, #3222) [Link]

Considering the USA's corrupt nature these days, you'd better watch out. Our government (be it the judges, cops, or lawbreakers,) might just come up with a way to patent math.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 26, 2011 21:04 UTC (Tue) by Unladen (guest, #72953) [Link] (8 responses)

Software patents transfer control from engineers to the business decision makers that run companies.
Engineers should join unions to collectively bargain for the right to retain title to their inventions.
Engineers (LWN readers) don't object to the existence of software patents, really. They object to businesses using them against technical people.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 26, 2011 21:10 UTC (Tue) by tao (subscriber, #17563) [Link]

Speak for yourself. I certainly do oppose software patents. And hardware patents. Because I oppose government imposed monopolies.

No, I object to software patents

Posted Apr 26, 2011 21:49 UTC (Tue) by david.a.wheeler (subscriber, #72896) [Link]

I read LWN, and I object to software patents, period. Patents - including software patents - should benefit society. Since software patents do not benefit society, they need to be abolished, immediately. Software patents are an industrial-era construct that make no sense in an information age.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 26, 2011 22:04 UTC (Tue) by coriordan (guest, #7544) [Link] (2 responses)

I object to patents being used against software developers. It doesn't matter if the attack comes from a business person or a technical person.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 21:28 UTC (Wed) by coriordan (guest, #7544) [Link] (1 responses)

Actually, maybe our points of view can be merged.

Software developers don't like being attacked by businesses (with patents, for example). And software developers do like interacting with other engineers.

I think we agree on that much. And then I'd add that once someone (a programmer or whatever) acquires a patent and starts using it aggressively, they switch from being considered an engineer to being considered a business.

If you go with that logic, then we're saying the same thing when you say that software developers simply don't like patents being used by businesses (because any aggressive patent holder is a "business"), and when I say that software developers simply don't like patents being used against them.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 21:38 UTC (Wed) by nybble41 (subscriber, #55106) [Link]

That's an interesting argument. Unfortunately, it amounts to saying that engineers are OK with software patents only to the extent that said patents are guaranteed to never be enforced. In which case, of course, they might as well not exist. Why add needless complexity?

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 7:59 UTC (Wed) by mpr22 (subscriber, #60784) [Link]

With (1) and (2), you sounded like you were a civilized person who I might or might not have agreed with. (3) proves you a troll. Go away.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 11:11 UTC (Wed) by Seegras (guest, #20463) [Link] (1 responses)

I certainly object to government granted monopolies if they are detrimental to society. Which patents are. Always. All of them. They were established as an experiment in the 18th century, but as one found out they didn't work (see for instance the history of the steam-engine: One huge case of why patents are a total failure), it was too late. People started believing in the superstition that patents were somhow "good".

The only thing that helps is to abolish them completely. In the meantime, a ban on patents on mathematics (wich includes software) would at least be the lesser evil.

Seegras
(LWN reader and System Engineer)

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 16:28 UTC (Wed) by Wol (subscriber, #4433) [Link]

To the steam engine, add the aeroplane - the first aeroplane to fly did so in the US, yet within fifteen years the US was so badly behind it had to buy nearly all its warplanes from the British and the French.

And add pharmaceuticals !!! most patents and research go into incremental research aimed at keeping old drugs "in patent", and not towards research for new drugs.

Add the light bulb - Edison's patent was for a carbon-filament bulb - filed LONG after he had seen a British tungsten-filament bulb. Which one was it that took over the world, again?

Pretty much all patents have been a disaster - not all I'll admit, but most have.

Cheers,
Wol

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 26, 2011 22:16 UTC (Tue) by fest3er (guest, #60379) [Link]

My opinion is that patents technically apply to tangible inventions--those things one can see and feel. Copyrights technically apply to intangible inventions--those things one cannot see or feel.

A frombus made of aluminanium with a trans-attached widget allowing one to daisy-chain a set of frombi would be patentable.

A software algorithm is intangible; it is entirely virtual. It takes no tangible form in the universe. It is comprised solely of bits and bytes which are, themselves, intangible. Software is an idea and can only be copyrighted. One may be able to touch the paper source code is printed on. One may be able to read the source code or even a printable representation of the compiled software. But these forms are not the form software takes when it is executed. What about firmware? you ask. The object containing the firmware can be patented, but firmware itself remains an idea.

Money and politics, of course, skew logic.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 4:23 UTC (Wed) by kevinm (guest, #69913) [Link] (1 responses)

This is a clever approach. By showing that a particular language executes only algorithms expressible in terms of mathematics, we can then prove that a given algorithm can be expressed in terms of mathematics, simply by implementing it in that language.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 28, 2011 17:16 UTC (Thu) by jd (guest, #26381) [Link]

Essentially what Alan Turing argued with his Computability argument. It's also a keystone in formal methods, which all rely on everything in a computer program being reducible to a mathematical form.

(Sadly, nobody has figured out a formal method that can be used in any practical way. Z and Object Z are nice, I like them a lot, but I dare you to find someone who designs their code using them. Since all code is maths and nobody has trouble coding, there must be a way to exploit the mathematical aspect of code without running into issues of horrible complexity.)

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 7:17 UTC (Wed) by feth (guest, #69696) [Link] (8 responses)

There is prior art, but 1+1 is not obvious in any case!

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 9:03 UTC (Wed) by mpr22 (subscriber, #60784) [Link] (3 responses)

1+1=2 is obvious to the school-educated non-mathematician (because they have a concrete definition of those terms which precludes it being non-obvious) and to the ordinarily skilled mathematician (because they're familiar with at least the summary of the number-theoretic proof).

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 11:12 UTC (Wed) by anselm (subscriber, #2796) [Link] (2 responses)

1+1 could also (less obviously) be 0, if you consider the group consisting of the set {0,1} and addition modulo 2.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 23:38 UTC (Wed) by bjacob (guest, #58566) [Link] (1 responses)

2 is just 1+1 by definition. In arithmetic modulo 2, 1+2==2==0.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 23:39 UTC (Wed) by bjacob (guest, #58566) [Link]

er, i meant 1+1==2==0

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 9:20 UTC (Wed) by dgm (subscriber, #49227) [Link]

Hey, cool id sixty-nine-six-ninety-six! But 1+1 is obvious once you know what "1" is and what "+" is. As is 1+1+1, 1+1+1+1, etc. once you know what induction is and how to omit parenthesis.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 9:57 UTC (Wed) by loevborg (guest, #51779) [Link] (2 responses)

True. "1+1=2" may have been true in all past instances, but what guarantees that the rule will yield 2 and not 85 in the future? What makes it true that "+" really means plus and not, say, quus?

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 27, 2011 14:12 UTC (Wed) by dgm (subscriber, #49227) [Link] (1 responses)

One word: definition.

1+1 (pat. pending) - Mathematics, Software and Free Speech (Groklaw)

Posted Apr 28, 2011 17:19 UTC (Thu) by jd (guest, #26381) [Link]

That response just BEGS someone to do a Filk version of "Tradition" (the first song from Fiddler on the Roof) for mathematics. Anyone want to have a go?

Patents don't require working devices.

Posted Apr 27, 2011 21:08 UTC (Wed) by ccurtis (guest, #49713) [Link] (1 responses)

I have to admit I stopped short of completing this rather lengthy tome, having made it to this part:

"Where is the requirement that the disclosure in a patent specification has been tested? Where is the requirement that the invention has been implemented? I don't find this in Quinn's explanation. [...]

"I have to ask, is this kind of disclosure sufficient to meet the societal bargain of a grant of temporary monopoly rights in exchange for disclosure? I don't see any requirements that the invention is implemented let alone tested. How could this disclosure be deemed sufficient? We are not even sure there is a working invention."

Part of the trade in expressing a patentable idea is just that -- to release the idea. The reason a feasible idea is sufficient instead of also a working device is so that the patent holder is granted time to either work out the details, or to protect themselves while shopping the idea out to someone they believe can bring the device to market. For evidence of that one only needs to look at the numerous patents for perpetual motion machines that have been granted (see this wikipedia article for examples).

Whether the author is hinging the remainder of the text on this argument, or is arguing that the patent system needs this fundamental level a reform, case law will not help; this action has to come from the legislature. It is likely also to be bound in all sorts of international WIPO and Free Trade style treaties preventing a relaxing of the standard as well.

I would want to argue that in a world where everything is faster - where an idea can cross a continent and be fabricated in less time than a rider on horseback can traverse a state - that there's been a fundamental shift in the time value of the idea versus the potential societal benefit. But to counter this I could point to pharma and the mandatory FDA approval process, where the original time value balance has actually worsened (getting approval versus crossing a continent or ocean). To this end, the extended protection period may be justifiable; however, this argument would simply lead to slowing progress via additional government regulation of software through some now-fictitious "Software Security Board" instead of increasing the rate and the freedom of the information. :-(

All that aside, I wish the author well.

Patents don't require working devices.

Posted Apr 28, 2011 17:28 UTC (Thu) by jd (guest, #26381) [Link]

Patent offices, at least in other countries, haven't accepted perpetual motion machines (or time travel devices) for some time. It's my understanding that the US no longer grants them either and that part of the justification in so doing is that there is a-priori knowledge on the part of the patent office that a working invention does not exist.

It's also been my understanding that you have to have prototyped at least the key parts of the invention even if you have not prototyped the invention as a whole. There has to be a minimal implementation. Certainly in companies I've worked at in the US where they were actively patenting inventions, that was a requirement the patent lawyer stated had to be fulfilled.

It would seem Groklaw has the same understanding, so if a patent lawyer and a paralegal come to the same conclusion, I can only conclude that that is the legal requirement. Doesn't make it the requirement in practice, merely the requirement as set down by law.

The US patent office, again according to my understanding, does NOT evaluate patents but waits for others to do so on their behalf. If this is correct, there may be many technically illegal patents in the books. If the system relies 100% on challenges, the law merely describes what can be enforced rather than what can be filed.