Fog Creek Software
Discussion Board




Software Patents

What is your opinion about software patents? I don´t think that they are totally evil or totally good, but I don´t can find the middle term in this question.

Camilo

Camilo Telles
Monday, April 19, 2004

They're pretty much 100 percent evil... the basic reason for a patent system is to provide an incentive for creators to invent things.  The software industry wouldn't change the least bit if patents were 100 percent done away with.  Can you name a single good software patent?

blah
Monday, April 19, 2004

Is this a software patent? Search in google. patent 5446747

Camilo Telles
Monday, April 19, 2004

Google is great!  Simply searching for that number and nothing else, the first returned link is the patent you speak of.

Hope they've got their algorithm patented :)

Wayne
Monday, April 19, 2004

I wouldn't mind software patents if they had shorter terms than patents for physical inventions... but as it is, when a company can buy the rights to an ancient compression algorithm and use it to extort licensing fees for any company that wants to be able to write to a popular image file format that happens to use said compression algorithm, long after the algorithm itself has been made obsolete... then I start experiencing comprehension failure.

Sam Livingston-Gray
Tuesday, April 20, 2004

As always it depends and as always it can be boiled down to the smartness of the people in the patent office. If they're stupid then we get stupid patents like the Amazon one click to buy patent. On the other hand the RSA crypto system (which haven't been cryptoanalysed yet!) is a very justified patent and now that it has run out (did in 2000 I think) we can now all use it because we know of it.

Peter Monsson
Tuesday, April 20, 2004

I know that Wayne was speaking tongue-in-cheeck, but of course Google holds numerous patents.  The patent for the Pagerank algorithm (number 6,285,999) is exhibit A:

http://www.google.com/search?q=google+pagerank+patent

(The patent automatically pops up at the top of the searches.  Spooky.  <g>)

This seems like a very well-deserved patent -- it was ingenious at the time and created a lightyear advance in searching technology.  (When combined with Google's many other proprietary searching algorithms.)  FWIW, Google wouldn't be quite the company it is today if everyone (Yahoo, Microsoft, etc.) had free dibs to use the algorithm and create Google knock-offs.

Of course there are many examples of stupid patents, but every so often there's a good one too.

Robert Jacobson
Tuesday, April 20, 2004

Any patent can be circumvented, evolution never stops so that idea that you patented today can become outdated in a matter of hours.

With this said, let's say I create a software and some company has a patent on a part of my idea implementation (this implementation I invented myself, without awareness of the already existing patent), that's OK. Even the telephone was invented almost at the same time by several inventors on different continents. What can the patent holder really do?

How are patents enforced in the IT world, does the patent holder want the other party to just stop using their technology in the future? they want a part of the profit from sales that have been made of products using their patented technology? does the patent holder want the "infringer" to recall all their product and replace it with the product that doesn't infringe upon their patents?

Naturally, if all negotiations fail and common sense doesn't kick in, the best case scenario is to simply change that part that is patented already and keep on doing things that are more productive than "patenting ideas."

It appears that in the whole process, there are alot of times where negotiation can kick in. Hopefully, it doesn't come down to "my lawyer is better than your lawyer"

Whisper
Tuesday, April 20, 2004

Software patents are useless because there isn't any software that would have not been developed anyway if software patents didn't exist.

Software can be protected via copyright and trade secrets.  A third form of protection in the form of patents provides no added incentive for creation, and only serves as a *deterrent* to creation when others find it nearly impossible to develop anything without infringing on somebody's patent.

T. Norman
Tuesday, April 20, 2004

Google has done very innovative things with their search capabilities, but they would have done those things anyway without patents because that's what it took to become the most popular search engine.  If they come up with a good idea they're not going to decide against implementing it because a patent on the idea gets rejected.

In some cases a particular company or individual may decide against something because they can't patent it, because it's an idea that's easy to copy.  But if it's that easy to copy without seeing the source code and other details, it's probably going to be easy to implement from scratch -- which means somebody else will bring the idea to fruition even if you don't.

Patentability simply is not a driving force in real software development (although it may form the basis of those pure intellectual property companies that don't develop anything).

T. Norman
Tuesday, April 20, 2004

What the EFF thinks of bad software patents (via Boing Boing):

http://www.eff.org/Patent/20040419_eff_pr_patent.php

has
Tuesday, April 20, 2004

Suppose tommorrow I have written this really innovative garbage collector algorithm that not just beats the pants of anything in existane today. No, it is about 2 orders of magnitude better than the best before.
Now what are my options appart from a patent? If I try to put it in a product it will be decompiled in hours and replicated in days, since it is really elegant.
I feel this has to worth something, right? I mean, if it where anything but software I could probably be a billionaire just by patenting some better canopener or something.
So what should I do? Write a paper on it and settle for some scientific attaboys?

Just me (Sir to you)
Tuesday, April 20, 2004

"What is your opinion about software patents? I don´t think that they are totally evil or totally good, but I don´t can find the middle term in this question."

As Sir pointed out, one needs some sort of incentive in order to create something. For some, that incentive is, indeed, the "attaboy" from his peers. Others won't even bother unless there's some sort of financial return.

So, for the benefit of the money-oriented crowd, you got the patent system. Which, theoretically, is a good idea. Unfortunately, since humans are not perfect, it's also very easy to abuse.

I could be wrong, but I think the reason why you find it so hard to form an opinion on this is because the idea of patents is not the problem. The problem is an ineffective patent system that seems helpless against abuse, and, of course, good ol' human greed.

Paulo Caetano
Tuesday, April 20, 2004

You might patent your garbage collection algorithm, but you won't make any money of it; It obviously infringes on ~200 patents you are not aware of (and will only be aware of when the patent holder sues you). These patents will all seem obvious to you -- you reached your algorithm independently, based only on published, unpatented material -- but it doesn't matter. (Much the same as anyone that comes with something  close to your algorithm, independently, will have to pay you).

P.S: Why on earth do you think a patent will protect you from the "dcompiled & reproduced" thing? It almost never does for real processes/inventions - someone will find a way around it.

And .. on the other hand ... why do you think anyone cares about extremely elegant solutions to problems?

Ori Berger
Tuesday, April 20, 2004

Ori, don't focus on the specifics, but on the general problem:
1> should it be possible to protect a sofware "invention"? If your answer to this is no, all the power to you, but then I'd like the remainder of the discussion to go with those that answer yes.
2> Still here? OK, so how should we protect the "invention" from being hijacked? Any other ideas besides patents?

Just me (Sir to you)
Tuesday, April 20, 2004

A good patent?  How about U.S. 4744028, for Karmarkar's linear programming algorithm.

I'm not much into linear programming, but I am told that this is decidedly nontrivial and represents a significant improvement in the state of the art over conventional simplex techniques.

David Jones
Tuesday, April 20, 2004

JustMe - I think Ori's point is that you will be buttfucked by corporations anyway, so what's the point? Your victim will just drag out the legal process until you can no longer afford the lawyer fees.


Tuesday, April 20, 2004

Blank,

I know what you are saying, but can anyone offer an alternative?
Couldn't an alternative history be that one of the big players snaps up my patent for their MAD arsenal, leaving me a reasonable $ amount in the process?

Just me (Sir to you)
Tuesday, April 20, 2004

Originally the courts said that algorithms were not patentable and I think, for the reasons others mentioned, this makes sense.

How about a program though?  What if you were the first to invent the spreadsheet- a complete and inobvious invention.  Don't you think you should be entitled to patent protection for something like that?

name withheld out of cowardice
Tuesday, April 20, 2004

The concept of a spreadsheet was already in use for decades (on paper).  If you can patent your spreadsheet program it should be limited to the specific collection of algorithms used in your implementation.  Others should be able to create their own spreadsheet in their own implementation.

There are a few good patents out there, but those 0.1% of useful innovative patents aren't worth the damage caused by the other 99.9%. Especially since most or all of those innovative ones would have been created anyway.

NoName
Tuesday, April 20, 2004

By "patent" I meant "software patent".

NoName
Tuesday, April 20, 2004

I don't think "software inventions" should have any more protection than "literary inventions". re the spreadsheet example, what if you had the most wonderful idea for a book, e.g. Neuromancer, that defines a whole genre? Aren't you entitled for some IP protection? I don't see much of a difference.

Copyrights? All for it, for a limited time (20 years should be enough incentive). Patents? Not for software in general, and definitely not the way the USPTO awards them.

Re karmarkars algorithm, ask yourself this: Would have karmarkar NOT published it if he wasn't awarded patent protection? It didn't seem to stop any of his predecessors, especially Danzig, whose contribution to the field far outweighs anything Karmarkar had done.

Just me: You are aware, I hope, that in this day and age, registering a patent eventually costs $20K-$40K, and that's before you try to litigate if someone infringes on your patent and refuses to pay up (and -- what do you know, he might just come with prior art; Those $40K do not guarantee _anything_).
Also, you can only win this game if you don't produce anything (see, for example, eOlas) because otherwise, any one of the big players can bury you with your own violations. Have you examined IBM's coutersuit against SCO? Please do; Regardless of whether SCO is right or wrong in this case, IBM can do that against _each_ and _every_ software producer out there.

There is no natural inherent right to invention, regardless if you thought about it first (or, more likely, believe you've thought about it first). There is a system in which a temporary monopoly is granted to an individual in order to give an incentive to share knowledge and promote the state of the art. Software patents don't seem to have this incentive effect.

Ori Berger
Tuesday, April 20, 2004

Knuth is another example.  His "Art of Computer Programming" which advanced the state of the art in many aspects at the time, was published before software was ruled as being patentable. Copyright protection and fame were sufficient incentives.

NoName
Tuesday, April 20, 2004

this isn't new, patents have had these abuses for a very long time, before computers even really existed.

mb
Tuesday, April 20, 2004

Ori and Noname,

do you believe the abolishment of all forms of IP should be general, or just limited to some fields?
I am fully willing to admit that the current reality of software patents is not good, but you are offering nothing else but an anarchistic law of the jungle type of environment, which I am also not too keen on. No other alternatives?

Just me (Sir to you)
Tuesday, April 20, 2004

Drop the "IP" uber-term. It puts too may unrelated things in the same place.

Trademarks, good (they are abused sometimes, but they do much more good than harm).

Copyrights, ok (should be shorter, IMO - and much more limited, but the general concept is ok).

Patents, depends - software patents are pure evil. I don't feel knowledgable enough to comment on other forms of patents, but I suspect in some areas they are justified. The patentability of software is new (~20 years or so), and has done NOTHING good for the community at large, it seems.

Trade secrets - ok, between parties having an agreement. If one wishes not to patent something so that it won't become public, that's fine - but one can't expect protection from reverse engineering or independent research in that case.

Just Me, software has been patentable in the US for the last twenty years or so, to various degrees. Can you point at even anecdotal evidence that it has promoted the "state of the art" in anything? e.g., large sums money spent on software R&D that would not have been spent otherwise? All evidence I'm aware of is to the contrary.

And I repeat again: I don't believe there is inherent right to anything because one did it first. "But I saw it first" doesn't even work in the kindergarten. A patent is a compromise meant to increase the probability that someone will "see it" at all in the cases "seeing it" requires great effort. It is not meant to reward the first to shout "saw it!" (with $40K or so to spare).

Ori Berger
Tuesday, April 20, 2004

Oh, and if you think Software Patents are justified, please explain why the following aren't justified (or, if they are justified, imagine a world in which they are enforced):

Recipe patents (e.g. patent "adding wasabi to beverage"?)
Movie plot patents (e.g. patent "asteroid crashing into earth")
Book/Novel patents
Makeup patents (e.g. patent "mixing concealer with lipstick to modify concealer color")

Many software patents are more ridiculous than this, and look more like a highway route patent (e.g., patent "getting from SFU to Berkeley by way of 101")

etc.

Ori Berger
Tuesday, April 20, 2004

Ori,

as I said before I am not out to defend the current state of software patents, so do not turn me into that strawman every time.
Is there a place for things that require a lot of upfront investment, are thruly novel, but once distributed are extremely easy to duplicate and recreate functionaly in wildly varing flavours? If so, then what form should this take?

Just me (Sir to you)
Tuesday, April 20, 2004

For that, you will have to define "large upfront investment". One man sitting in his own kitchen does not, in my book, a "large upfront investment" make (wrt to recipes). Similarly, one man sitting in his own study with a piece of paper and a PC is not a large upfront investment -- and most software patents essentially reduce to this kind of effort.

Can you give a concrete example of what you have in mind? I tried, but the only examples I can come up with are those worthy of _copyright_ protection (e.g., writing a book), not of _patent_ protection.

Ori Berger
Tuesday, April 20, 2004

And the years that this guy have studied, you experience and a-ha moments? This is not a "large upfront investment"?

Camilo Telles
Tuesday, April 20, 2004

It's important to bear in mind that, while business patents might have problems (and they do), we should not just blindly support open source calls to reject them completely.

Open source fans essentially want to remove all protections for IP and creators, including copyright.

So, yes, business patents have problems, but I'm not sure what the solution is.


Tuesday, April 20, 2004

Neither an "a-ha" moment, or years of studying, nor experience is worthy of any protection, IMO.

Because, it sometimes happens that what takes one person years of studying, and 8 years of real-world experience, and a once-in-a-bluemoon a-ha moment -- is trivial for others.

The real question, is -- I believe, the only one relevant to the mandate that the patent office legally has -- whether those years of studying and/or gaining experience and/or "a-ha" moment would not have happened if a patent wasn't awarded.

And nameless, please educate yourself. Neither the open source movement, nor the free software foundation (FSF) wants to void copyrights. On the contrary - the FSF works to educate _against_ copyright-less practices like placing software in the public domain, or using a BSD-style license. Copyleft _is_ a form of copyright, see, e.g., [ http://www.gnu.org/licenses/licenses.html#WhatIsCopyleft ]

Ori Berger
Wednesday, April 21, 2004

Ori: What is the fundamental difference between hardware and software that allows EE graduate students (and others, I'm sure) with little or no resources to patent 'stuff' and make money?

&#1588;&#1587;&#1740;&#1576;
Wednesday, April 21, 2004

I don't know enough about what's happening in the hardware patent arena these days - but, if they can do that without an inherently necessary multi-million dollar investment and expensive experiements, they shouldn't be awarded patents. The food recipe analogy is appropriate here too - there's no reason to award a patent to a chef for a food.

The question, again, is: Would the community at large suffer without this patent? A patent is a LEGALLY SANCTIONED MONOPOLY, for crying out loud. Why do people want to award these in a free market so easily is beyond me.

The blue led is a good example for why patents are a necessary evil in some cases - based on media reports, it seems that the company (don't remember name) wouldn't have given the inventor (don't remember name) a free hand and financed 15 years of salary and not-so-cheap experimentation, if they didn't believe that, _IF_ at some point something does come out of this work, then they will be able to control, even if for a limited time, the production of those leds.

Ori Berger
Wednesday, April 21, 2004

Ori, you are basically arguing against potential large monetary reward as an incentive for innovation risks, aren't you? Are you against multipliers? Do you believe reward should be directly linked to "time spend"?

Just me (Sir to you)
Wednesday, April 21, 2004

The point is that if the expense and effort to create something is trivial it will be done anyway without a patent.  Thus the patent is not necessary to provide an incentive for such creations.

The reward for innovation risks should be reserved for those risks that are large enough that they need the reward of a patent to bring them to fruition.

NoName
Wednesday, April 21, 2004

Just me: NO. Please re-read my blue-led post.

I'm saying patents should be reserved for innovation RISKS (emphasis on RISK, not on innovation), which would not _otherwise_ be taken. The blue led fits this. No software patent I'm aware of fits this bill. Neither do business process patents - and neither do some chemical process patents and some hardware patents.

The market should reward taking risks, and -- in the cases in which the market punishes taking risks to the extent that _community rewarding_ risks will not be taken (what originally led to the introduction of copyrights a few hundred years ago), only then should the government be involved, possibly through granting a temporary monopoly, e.g. a copyright or a patent, or through other ways.

Ori Berger
Wednesday, April 21, 2004

NoName,

would that reserve patents for things only if they aren't being done, but once they are done, well, then you don't get the patent since, well, it's done anyway? Bit of a catch 22, no? Or will it be more of a concours style setup. Hear ye, here ye: patents for 2010-2015 period will be awarded to the following realisations only ...

Seriously though, then how much of a reward should a patent be? Just enough to cover the investment? Hey, but why focus on results? Maybe we should just reimburse expenses directly, right?
Ah, the grant dwelling researchers wet dream. Kick out the whole private sector by negating their potential to go beyond commodity production, and have it all tax financed.

Excuse me for not being thrilled.

Just me (Sir to you)
Wednesday, April 21, 2004

Ori,

how do you measure risk? When do you measure risk? At the onset, or at the finish line? Who decides?

Just me (Sir to you)
Wednesday, April 21, 2004

The judgment of whether something should be allowed to be patented is inherently subjective.  It will be up to the Patent Office to determine what is patentable.  Somebody somewhere will always disagree with them -- but if they would just base their judgments on the principles which they were founded ("promote the sciences and useful arts"), we wouldn't have so many trivial patents cluttering the landscape. 

Instead, their judgment is based on the presumption that a patent is valid when applied for, and the patent is to be granted unless the examiner has a compelling reason not to grant it. No consideration is given to the danger of a wrongly granted patent. It should be the opposite, with the applicant having to make a strong case and impress the examiner into believing that the creation is new, nonobvious, and would not have been created without the incentive of a patent.

NoName
Wednesday, April 21, 2004

"and would not have been created without the incentive of a patent"

PO: 'Dear Just me, please prove to me that you would not have created this socalled invention without the incentive of a patent.'
Just me: 'That is easy dear PO. Everyone will attest that I am a lazy moneygrabbing bastard that would never lift a finger without there being some potential profit involved.'
PO: 'I see. Well ok then, I see no further objections. Patent granted.'
Just me: 'Thank you dear PO. Now if I may be so bold as to draw your attention to my competitor *ugh* I mean collegue next inline. We all know he is a goody two shoes that just works for the "cause", don't we. It is obvious he would have researched that stuff even if he would have been told by our dear Lord himself that he would die poor as a rat without any recognition.
PO: 'Is that so. Yes he does look like a bit of a radical indeed. Obviously under those circumstances he is completely disqualified from any patent whatsoever. Thanks for the intelligence, Just me, and have a nice day.'
Just me: 'I believe I will, dear PO, I believe I will'

Just me (Sir to you)
Wednesday, April 21, 2004

Since risk is not measurable a-priori, you estimate it by correlating the past. That's supposedly where the patent 'test' came from ("1.non obvious to someone skilled in the art; 2. novel; etc.").

And the question isn't whether *YOU* wouldn't do it. It's whether *NO ONE* will.

Ori Berger
Wednesday, April 21, 2004

Of course, many would have done it blessed with the hindsight of success.

Just me (Sir to you)
Wednesday, April 21, 2004

Oh well, what can I say. You have convinced me beyond the shadow of a doubt that without a patent-like incentive, no one would ever try to achieve anything.

Like going to school (what? You mean no one promises that I'll get a fruitful job after spending all this time in school?) or publishing a recipe (what? you mean anyone would be able to make my yet-unknown world-famous apple tart at home without paying royalties?!?)

I've kind of lost interest in this discussion, as it seems at least one of the sides insists on not considering historical evidence, and would rather hypothesize in a "wouldn't it be nice if ...." world.

Ori Berger
Wednesday, April 21, 2004

Ori,

Let me be very explicit. I agree 100% with you that the current software patent system is a farce that does potentialy a lot more harm than good. But at the same time I do not agree that somehow everything should be up for grabs.
I don't see how copyright can be succesfully applied to the software field.
Sadly I do not have a good answer as to what would be a >good< system.  The "risk-based" and "would it have happened without" systems briefly put forward here do not convince me.

Just me (Sir to you)
Thursday, April 22, 2004

The point of a patent to me is that if you are a small entity, and you enter a market where there is no competition, you should have your product licensed by competitors...not stolen.

I guess there is truth to patent abuse and you will have to defend if you infringe on others. But if you develop a great product that (morally) ought to result in others licensing, or acquiring your company. Instead just copy and steal your work...that really sucks.

Its valuable property. I couldn't call my company Microsoft...it has value. I don't see how software is any different.

Some guy
Thursday, May 13, 2004

*  Recent Topics

*  Fog Creek Home