John Carmack on Patents | Hacker News Reader

John Carmack on Patents

slashdot.org

64 pointsazharcs38 comments

Carmack isn't just being hypothetical here; he independently came up with "Carmack's Reverse" for rendering shadows and it turned out Creative already had a patent on it; they ended up using it to force Doom 3 to support some crufty proprietary 3d sound system.

Look at the current patent trends. The first company to paten an idea that having people interact via avatars interact in an online world managed by a server was not the first company to come up with the idea. However they are now suing every MMORPG on the market because any game in which users chat with eachother using avatars violate the patent.

Apple is doing such things as well patenting things like "a 3d desktop" they don't have to do much just describe a rough idea of one.

OOOOOO and my favorite is the company who is suing Apple, MS, Palm, Google due to their exclusive patent on operating systems. I am sure they invented OSes it just a few years ago.

Yes, it is a legal tool that may help you against your competitors, but I'll have no part of it. Its basically mugging someone.

Love it. Legal does not mean moral. Or right. Putting on a suit and using a paper weapon doesn't make it any more respectable than hiding in an alley and using a steel one.

I'm looking forward to Mr. Carmack's position paper.

Well, that post is about 3 and half years old. So don't hold your breath waiting for his position paper :).

Yeah saw that. Went searching hopefully. Disappointment. Busy guy I guess. In the end though, I'd rather he be making cool stuff than messing around with patent nonsense, whatever side he's on.

Not non-sense,m actually. Quite the opposite. This is comapany and industry-ruining stuff. To continue on oblivious to the law and the effect that industry leaders like Carmack can have on the law is nonsense.

Do you think John Carmack's time is better spent writing 3d engines or lobbying congress?

Hidden third option: building space rockets and driving Ferraris.

Touche.

I recall reading something from Carmack along the lines of "I find it horrrifying that I can't use my own ideas if someone else had the same idea first." It was a very succinct and insightful quote on a problem with patents, but now I can't find a source for the quote. Does anyone know what I'm reffering to?

Ok - So - What's the alternative?

There are some inventions that take vast amounts of investment, yet are easy to copy.

Government sponsorship of inventors, maybe?

Some expert could decide which projects are worth pursuing and fund the research and development.

Private companies could then take those inventions and build businesses around them.

There could be a large cash prize if the invention ends up being especially valuable.

This is essentially how research universities work.

The 'large cash prize' tends to be either tenure or a fat consulting job.

This is essentially how research universities work.

Research universities do not give away the technologies they develop. They sell or license them for millions of dollars. http://www.google.com/search?q=university+patent+license+mil...

I believe JesseAldridge above was referring instead to government funding R&D and then simply giving away the technology the way the daguerreotype photographic process was given away. http://en.wikipedia.org/wiki/Daguerreotype

Instead of Daguerre obtaining a French patent, the French government provided a pension for him. In Britain, Miles Berry, acting on Daguerre's behalf, obtained a patent for the daguerreotype process on 14 August 1839. Almost simultaneously, on 19 August 1839, the French government announced the invention as a gift "Free to the World".

I'm not sure the bureaucracy of this would be any better than the current (flawed) patent system... In fact, I'd venture it would be much, much worse.

At least in the current system you're protected, and potentially rewarded - even if everyone thinks it's a bad idea (which is often the case for break-through innovations).

The current system is not well tuned to current technology. If you patent the steam regulator, the idea was that you would sell then far and wide... In the industry now it's used to lock up a technology for exclusive use.

Maybe a patent should coffer the rights to an invention - and the patent is also opened up for auction by the govt to "n" organisations (where n is somehow determined to maximise profit), with the profits going to the patent holder.

I think the problem is in the details. It's far too easy to get a patent. And on trivial 'inventions'.

Paul Graham has literally got two patents for the exact same thing. 'Somebody hit submit twice' indeed.

Patents should be examined by people knowledgeable in the field and have a strict process for being granted.

And there should be a much higher activation energy required for taking someone to court for violating a patent.

I think the problem is in the details. It's far too easy to get a patent. And on trivial 'inventions'.

Yeah I agree - I want to dig deeper when people say "patents are evil".... By this they could mean the current system needs to be tweaked, right through to totally abandoning the concept.

A quick shout-out to the US Patent Examiners - these individuals have technical degrees just like us, and do their best to weed out the crap patents they can. In fact, they are judged (for better or worse) by the number of patents they successfully reject.

They are also now some very overworked individuals. They have only hours to review patents in a diverse set of topics. There is only so much research they can do - a patent written by a master claims writer, no matter how frivolous, is very difficult to reject outright.

If you want to improve patents in the US, you can start by asking your congressperson to increase funding to the USPTO.

Or ban software patents altogether. That would probably be cheaper.

In the meantime, strike small blows here and there against bad patents by helping examine them:

"Help the USPTO find the information relevant to assessing the claims of pending patent applications. Become a community reviewer and improve the quality of patents."

http://www.peertopatent.org/

Your question assumes that "the more inventions, the better for sociaty". While this may have been true for the last five hundred years, maybe it's time to reconsider the assumption?

This may sound foolish but please consider: What is more important? To create pills that enable old men to have sex or to solve today's energy problems? The later is not just about new technologies but to overcome the old established structures. Or consider medical treatment: Do we really new technology to heal very uncommon ills for a rich upper class when a large part of the sociaty cannot afford to have well-known basic medical treatment? The later is a social problem, not a technical one.

Another assumption of your question is that patents are the only way to secure a large investement and that there are no alternative barriers to entry for others. However, that is clearly wrong. Do you have the money to manufacture automobiles, for instance? Even if you're capable to invent a new kind of car, do you have the money to advertise it properly?

The third assumption of your question is that we may need a completely different system. However, patents seems to work somewhat fine for the material world. It just doesn't work for software. So, maybe we should not look for an alternative but for a different, a better kind of patent system?

I'm not saying I have an answer. I just think one should be careful about either/or questions.

Do you have the money to manufacture automobiles, for instance?

Capital for new risky ventures, including automobile manufacturing ventures, is available from this source: http://en.wikipedia.org/wiki/Venture_Capital

Venture capitalists tend to be more willing to provide capital if one's venture has patents.

To create pills that enable old men to have sex or to solve today's energy problems?

This is more of a moral judgement than one around the efficacy of patents. Are you arguing that patents shouldn't be available at all, or they should only be granted to those that are a suitable fit for society?

This sounds dangerous as I can't really think of anyone suitable to make such a judgement. Even in the later case, you still need an enconomic basis for the patent - or more correctly, the R&D required to gain the patent.

Even if you're capable to invent a new kind of car, do you have the money to advertise it properly?

I don't need to invent a new kind of car - if I came up with a valve arrangement that increased efficiency by x% that would be patentable, and I would profit from it.

If this invention, say, could be retrofitted to existing vehicles and reduce emissions - then governments might mandate it's use - and I'd profit from that.

I don't think a body such as the government should be guiding research. It's a slippery slope to bureaucracy. What you can do is put in place the right framework - e.g. all cars must comply with an emission standard by X - then you create the economic basis for the right inventions to follow.

Short patent lifetimes. Long enough for you to get out and establish yourself in the market, but short enough that it wouldn't kill a competitor to wait a bit. It would also dramatically reduce the cost of licensing a patent since the waiting time for it to expire would drop so much.

Perhaps a good amount of time would be one or two years?

Hmmm... Interesting... Seems like a prisoner's dilemma situation to me. If I was to patent, it would be to my benefit, but it would also hurt the 'other ' I was playing the game with (supposing he didn't patent).

I suppose Carmack's position paper would advertise to the world that he is in favor of no patents, thus in this 'multi-round' game he is providing others with the ability to collude with him to keep patents out of his field.

When a product is covered by more than one patent, the patent-holders end up in a tragedy of the commons problem. The commons is the person with the product. He has money he can pay. Each patent holder has a veto over the product and wants to extract as much as they can. One hold-out can ruin it for everyone and every single stake holder often ends up worse off as a result.

The cases you are referring to - where there are blocking patents - generally are resolved with mutual licensing agreements. Of course, there are many exceptions, especially when there are patent trolls involved.

In these cases, though, I really wonder the comparative cost of innovation versus licensing battles. Software is still an amazingly open field - given that licensing battles can run into the millions, if that money was invested in R&D could an alternative (superior) method be found?

One of the methods that patents were supposed to encourage innovation was by providing reasons to innovate. If you are forced to explore new options and develop your own IP that might be patentable, is that such an awful thing?

Software patents often leave zero room for innovation. Let's say you patent pushing email to a hand held device AND pulling email from a hand held device what are the options? Showing a screen shot of another device that's not hand held? It's still either push or pull. etc.

You can also try to release as much info as possible to flood patent trolls with prior art.

I know of other game developers who essentially have public domain game engines / helper libraries. Aside from sharing, the rationale is that they can re-use their code on the job, without having to re-solve problems they've already solved.

Who has to shoulder the burden of proof on prior art?

I work at a commercial software company, for which I've filed and been granted some patents. I do so even though the pittance my company provides as a "patent bonus" doesn't begin to compensate for the paperwork, writing, iterating with a lawyer, etc. involved in the process.

In principle, it's possible one of my patents will either defend my employer from litigation, or be a poker chip in some licensing deal in the future, but honestly the odds of that are small. My selfish reason for writing patents is that they provide a way of publicly recording my technical contributions. An alternative venue is academic publications, but that's inherently a crapshoot. Granted patents have your name on them, are freely available unlike a lot of CS journals, and usually provide better technical detail about the problem and solution than a typical 5-page academic CS paper.

In what sense are academic publications more a crapshoot than patents? I hope you don't mean that because they might be rejected! That would completely convince me of Carmack's correctness.

Until we can remove the leverage that intellectual property holders currently have, they will fight ruthlessly to retain 'their' property.

Compare the economically optimum copyright length (http://arstechnica.com/old/content/2007/07/research-optimal-...) with the actual changes in length within the US (http://en.wikipedia.org/wiki/File:Copyright_term.svg). This increase in terms coupled with the ease of patent filing surely is purely marginal decision making.

Uh, effect -> affect.

The truly elite (or at least, those who think of themselves as such) know that "effect," when used as a verb, is (sort of) a synonym for "cause."

It's clear from reading Carmack's sentence that he means affect (alter/disturb, in this case negatively so), not the verb form of effect/to cause. (Not that that justifies a grammar flame on HN, but it was at least an accurate flame.)

Patents should be valid for one year only, to give a head start to whom had the idea and invested developing it.

That willl give them enough time to build a brand (iPhone, Viagra) then it is all open season and everybody can copy and improve on that idea to get an edge, investing their own resources (Zune, Cialis) whether they fail or not is up to the market.

You can't blame people for you being a great researcher and a poor marketer.

People will reward who can be both.

Reply on news.ycombinator.com