Torvalds Joins Anti-Patent Attack

canuck57 sent us a story about Linus Torvalds has joined the chorus of voices speaking out against software patents. Talks briefly about the recent patent releases by IBM & Sun, and notes that there are 'an estimated 150,000 to 300,000 registered software patents in the U.S. alone.'

Comments

any software patent is bad

[CAIMLAS]

I'd say any software patent is bad, simply because they're so ambiguous. It doesn't matter if the company in question is supposedly benevolent, or that they're not actively enforcing them: all it takes is a single lawyer with no scruples to cause a lot of pain. In any organization of significant size, you can rest assured they've got at least one bastardly lawyer.

Not only that, but there couldn't possibly be that many new, patentable techniques or technologies being discovered. Is it actually good practice to patent everything? While it might be "good" for open source with IBM supporting us and all, what's it do to the smaller companies that get (potentially) shafted by such absurdity? At the very least, it increases their cost of development due to necessary research.

Re:any software patent is bad

[cpt kangarooski]

Patents were created to protect inventors from having their ideas ripped off and used to profit another.

No, in fact that's what patents are intended to do: to get inventions that are reduced to practice into the public domain so that everyone can use them without paying inventors. And as implemented, we further make it difficult to get patents by requiring novelty, nonobviousness, filing before the statutory bar, etc.

Patents are intended to promote the progress of the useful arts, for the public benefit. Not to benefit inventors, though that may incidentally occur.

A good analogy is this: Imagine the public is a farmer who has a cart of vegetables he wants to take to market. He has a donkey (the inventors) but it is unwilling to do very much without some special incentive. If the farmer is willing to spend one of his carrots by dangling it in front of the donkey, getting it to move and therefore act productively, he can achieve his greater goal of getting all his stuff to market.

The farmer doesn't want to give the carrot to the donkey, however. Then he's out one carrot. But it's an okay cost if it profits him more in the end by getting to market.

Likewise, it is a bad thing to grant patents for their own sake, or for the sake of inventors. But if they are not a significant burden on the public, and the public benefits much more than we lose by virtue of encouraging inventors to do useful work for us, then it's okay.

So the problem with software patents is that the software industry seems to have been doing somewhere between good to awesome without them. There is no indication that there will be more invention in this sector by adding them, and there is a very real problem with software patents slowing down the pace of innovation in software and in getting those inventions in useful products.

So software patents don't seem to be worth it: they produce no benefit and incur great cost. We're better off without them.

Re:any software patent is bad

[Shalda]

It's not that all software patents are bad, just most of 'em. Take the LZH compression algorigthm. It was a pretty good patent back in the early 80s when it was granted. It was sufficiently narrow (one specific method of data compression) and reasonably not obvious (for that point in time). The problem is that the patent office has no concept of 'novel' or 'obvious' and that prior art means 'already patented'. Also, there's a tendancy by the patent office to let the courts decide the actual merits, whereas the courts like to defer to the patent office.

Shouldn't the real question be...

[winkydink]

why has he waited so long?

If this isn't the first time he's spoken out against them, then why is it news?

Re:Ironic

[MustardMan]

Yeah, it's almost like the slashdot editors are allowed to voice their own views and not follow a company line set out by the people who own the site. Imagine that, editorial freedom in a news site! What a novel idea!

Re:Something I've never understood...

[Kwil]

I don't think we should be able to patent processes at all.

A process is the ultimate business advantage. If you can come up with it, you deserve to reap the rewards from using it. Not from selling it to or litigating against some other group.

This is where the system breaks down. Some things are not meant to be non-freely shared around society.

Patents should return to whence they came. Physical objects.

Copyrights should return to whence they came. Expression of ideas.

Processes are neither, and therefore shouldn't be covered by either.

Re:So, how many patents has he registered?

[MyLongNickName]

I am not against software patents. I think they are a good thing. However, we seriously need to reconsider what is considered patentable. Some of the approved patents are blatantly absurd, and actually hurt commerce.

Patents are desiged to encourage innovation (as ou rightly point out). But big business has twisted gov't's arm so much that they no longer serve the interests of the people as a whole. For a ridiculous example of COPYRIGHT protection: The 'Happy Birthday' song is still protected... found this out when I wanted to add it to an app I wrote... Patents are similarly absurd.

So, like most other things I fall squarely in the middle of the two camps, and get shot at from both sides.

Re:Something I've never understood...

[Jheaden]

I have to disagree somewhat.

I don't think we should be able to patent business processes or software processes at all. However manufacturing processes should be patentable.

I'll clarify that a little bit too. The process that should be patentable would be the process for say creating a specific alloy or chemical where it is not simple. Along the lines of non-obvious to someone in the field.

What shouldn't be patentable is the how do I assemble product X

It's simple...

[Corson]

Back in 80s there wasn't so much patenting going on in Silicon Valley. In those times you needed to innovate and bring the product to the market so fast that you couldn't afford to apply for patents; even two weeks counted on the learning curve. And people became instant milionaires.

Twenty years later, after a recent bumpy ride in the IT sector, investors (who generally understand little in terms of technoloy) would not invest unless they see there is some IP protection -- a.k.a. patents. Hence, the pressure for software patent legislation comming from companies that want to positively attract investor's attention. Big sharks such as M$ shouldn't really need software patents unless everybody else moves in that direction. They also probably learned a lot from big Pharma that patent everything they "discover" and then license those "discoveries" out to smaller companies. It's a different game these days, a different kind of race that, I'm afraid, the small fish (read: open-source developers) will unfortunately lose.

Re:So, how many patents has he registered?

[Ironsides]

If software can't be copyrighted or trademarked (note I do not include patented), here is what would be legal:

I could take a companies software the day it was released, make copies of it and sell it for whatever I wanted. Think Game companies have problems now? Just wait until they can't do anything.

I could make a game called HALF LIFE 2, and sell it online and people would have to worry about buying my game vs. the original game. And the Makers of Half Life couldn't fo jack.

The GPL would become worthless as it relies on copyrights in ordeer to work. This License applies to any program or other work which contains a notice placed by the copyright holder saying it may be distributed under the terms of this General Public License.

I'll agree with you on the Patenting of software (although there might be an option for using it to a limited extent, say 3 years). But copyrights and trademarks of software are necessary.

Read what Mitch said ...

[sdanis]

Part of the problem is that the US Patent Office has been lax in granting patents, said Mitchell Kapor, a founder of Lotus Development and a prominent backer of the Mozilla browser. "There have been tens of thousands of bad software patents issued which never would have been issued if the Patent Office had actually been following its own rules," he said.

The patent laws make some degree of sense. The patent office does not. The stuff they let through is totally unbelievable.
You can either wait to be hit with an infringement case (not a lot of fun) or you can submit a reexamination request to the PTO.
The worse thing you can do is to read the patents of your competition. Once you do that, you had better react quickly. Willful infringement in the US gets you treble damages.

What a wonderful world we live in ...

Re:So, how many patents has he registered?

[jbr439]

I am not against software patents. I think they are a good thing. However, we seriously need to reconsider what is considered patentable. Some of the approved patents are blatantly absurd, and actually hurt commerce.

Patents are desiged to encourage innovation (as ou rightly point out).

I have often wondered how many software "inventions" would not exist today if software wasn't patentable. In other words, has the patentabilty of software resulted in any software "inventions" that otherwise would not have existed? My gut feeling (and that's all it is) is that there aren't many (any?) such "inventions". (inventions in quotes 'cause I find it difficult to think of any software as being an invention).

So, do software patents really encourage innovation? Are there clear examples of this? Or, are software patents just for the use of large corporations limiting competition from small outfits?

Re:So, how many patents has he registered?

[cronius]

So, like most other things I fall squarely in the middle of the two camps, and get shot at from both sides.

Please, let me ;)

Seriously though, consider this: I write a piece of software. I'm a total geek and a math wiz and it turns out I manage to create a really nice algorythm for [insert tech word here].

I want to make money off this or maybe share it to the community, regardless it's released and distributed and one beautiful morning I'm being sued for patent violations. Turns out some big company thought of this a good 10 years ago and has a patent on doing roughly the same thing my algorythm does.

What did I do wrong? Why am I being punished? Is this fair? This wasn't an obvious thing to patent, I'm just a bright kid and stumbled across it.

This is not a "nothing is perfect"-thing, this is a serious flaw. Patents (especially in software) means no one has any rights what so ever, *unless* you happen to be the first guy figuring something out, in which case you suddenly have *all* and exclusive rights.

I'm not taking about patenting obvious things: with patents no one is allowed to invent things unless they're the first one. It doesn't matter if you figured it out all by yourself, if someone beat you to it you simply have no rights.

How is that fair?

Re:Patents are ok, if they are inventive

[WolfWithoutAClause]

Patents would be ok, if they were inventive as required by the law. Unfortunately, there is no way on Earth to measure or judge inventiveness, so all a patent examiner can do is to judge whether the application is novel.

There is, but it involves going to court.

But that is the big problem.

The whole patent system needs reforming. Patents are too easy to get, can be obtained for ideas that are trivial extensions of existing ideas (particularly in the software area)- and there's little or no downside to getting such a patent. So companies just play the percentages, get lots and lots of patents.

Meanwhile, individuals are rarely well advised to get patents- a patent is just a license to sue, but individuals often can't afford to sue anyway, so then the patent isn't worth the paper it is printed on.

There's also big problems with patents in that nobody really knows what a patent covers. Remember that patent that BT 'had' on the world wide web? It wasn't a slam dunk that they wouldn't win that one, it was close. Patents don't only cover the exact invention, they also cover similar inventions. And the web was sorta similar to their patent; but the court decided it was too far, in that case, a different court might have decided otherwise. That's what makes it all impossibly complex.

To make things worse, many software patents usually don't come with a useful description of how to actually do stuff, which is sad, since software can be documented by the sourcecode and printed.

That would be stupid- a patent application has to be written so that one 'skilled in the art' is able to reproduce the invention. Not doing that might well invalidate the patent.