Join 3,512 readers in helping fund MetaFilter (Hide)


A patent on speech
March 26, 2012 2:07 AM   Subscribe

Dana Nieder explains how an ACC patent fight may substantially delay her daughter learning to speak.

The iPad app Speak for Yourself developer is being sued by their larger competitors, Semantic Compaction Systems and Prentke Romich Company. Dana Nieder explains SCS and PRC aren't interested in selling software without bundling expensive specialized hardware, which isn't currently usable by her daughter.
posted by jeffburdges (125 comments total) 6 users marked this as a favorite

 
Video of the daughter and app.
posted by lee at 2:21 AM on March 26, 2012


Anyone able to untangle what the patent is?
posted by Artw at 3:34 AM on March 26, 2012


Seems to be these two patents:

5,748,177 - May 5, 1998
Dynamic keyboard and method for dynamically redefining keys on a keyboard
A dynamic keyboard includes a plurality of keys, each with an associated symbol, which are dynamically redefinable to provide access to higher level keyboards. Based on sequenced symbols of keys sequentially activated, certain dynamic categories and subcategories can be accessed and keys corresponding thereto dynamically redefined. Dynamically redefined keys can include embellished symbols and/or newly displayed symbols. These dyanmically redefined keys can then provide the user with the ability to easily access both core and fringe vocabulary words in a speech synthesis system.

5,920,303 - July 6, 1999
Dynamic keyboard and method for dynamically redefining keys on a keyboard
A dynamic keyboard includes a plurality of keys, each with an associated symbol, which are dynamically redefinable to provide access to higher level keyboards. Based on sequenced symbols of keys sequentially activated, certain dynamic categories and subcategories can be accessed and keys corresponding thereto dynamically redefined. Dynamically redefined keys can include embellished symbols and/or newly displayed symbols. These dynamically redefined keys can then provide the user with the ability to easily access both core and fringe vocabulary words in a speech synthesis system.
posted by Gyan at 3:54 AM on March 26, 2012 [3 favorites]


Software patents are such a joke. With those patents they might as well sue every POS machine for restaurants that has a "drink" button which leads to a new menu of types of drinks or whatever.
posted by burnmp3s at 5:51 AM on March 26, 2012 [11 favorites]


We've known how detrimental software patents are for decades now. It's infuriating that these situations keep happening.
posted by JHarris at 6:02 AM on March 26, 2012 [3 favorites]


Don't all grocery store self-checkouts use the same interface? You're trying to buy some red bell peppers, so you hit the vegetables button, and then the screen changes and you choose the peppers button, and then the screen changes and you choose the red bell peppers button. That's essentially the same interface as you want to ask for a snack of cookies so you choose the food button, and then the screen changes and you choose the snacks button, and then the screen changes and you choose the cookies button.
posted by hydropsyche at 6:07 AM on March 26, 2012 [5 favorites]


Note that the suit does not necessarily mean the end of the app. Since this is essentially a medical device, it's much more questionable whether the plaintiffs would be able to get an injunction. That means the end result (if the patents are actually valid, enforceable, and infringed) could be a reasonable royalty or lost profits award, which would make the app more expensive, but would not be the end of it.

Re: POS machines and grocery store self-checkouts. The '177 patent was filed on June 7, 1995 and the '303 patent was a divisional of the same application and thus has the same priority date. These go way back and potentially predate both of those examples. The dates posted by Gyan are the issue dates, which (because these were filed before June 8, 1995) are relevant for determining the patent term, which in this case 17 years after the issue date, but which are not relevant for prior art purposes.
posted by jedicus at 6:14 AM on March 26, 2012 [1 favorite]


The Iceman Cometh with his Legal Team
posted by roomthreeseventeen at 6:31 AM on March 26, 2012


roomthreeseventeen's link quotes a price of $500 for iPad and $300 for the app. WHAAA??? If there were ever a case to be made for open source, this is it. Is the market really so limited that $300 is a fair price? I don't think so. Seems like these families get shafted by everybody along the way (my brother's family, for one).
posted by rikschell at 7:00 AM on March 26, 2012 [4 favorites]


Is the market really so limited that $300 is a fair price? I don't think so

Actually, for specialized development, expensive skills/knowledge needed for good design, and a small user base, that really could easily be a totally fair price, in the sense of "enough to cover cost of development and a reasonable profit." Of course, that's also why a healthy society pays for these things collectively rather than actually dropping all of the weight on the user and their family, but hey, that's a rant for another thread.
posted by Tomorrowful at 7:15 AM on March 26, 2012 [17 favorites]


Yes, the app is crazy expensive when you compare it to the usual $0.99 - $9.99 price point we're used to, but the dedicated AAC devices run well into the thousands of dollars, so it's still a huge price savings for those families.
posted by shiu mai baby at 7:17 AM on March 26, 2012 [6 favorites]


If there were ever a case to be made for open source, this is it. Is the market really so limited that $300 is a fair price? I don't think so.

The market responded to demand. If you want an open source version you'll have to wait until some developer needs to "scratch an itch" by creating a similar assistive device program. That's kind of one of the flaws in the open source model. There are lots of important areas where open source developers don't get too many itches (if you'll pardon the imagery).

I suppose someone could pay a developer to make the app and then release it as open source, but that's not really the open source model so much as charity, which is perfectly fine, of course, but it's not the same thing.
posted by jedicus at 7:19 AM on March 26, 2012


It just seems like a technically easy thing to implement. The hard work is testing out how to design the best UI, and it seems like a few hours/days spent playing with one of the pro dedicated models would be sufficient for even a pretty poor programmer to slap some stuff together to work the same way. Which is actually a pretty good argument for IP protection, when you think about it. Once the hard, expensive part has been done, it becomes pretty easy to copy for cheap. But it would be nice if fairness could be factored into the business model.
posted by rikschell at 7:27 AM on March 26, 2012


Dana Nieder explains SCS and PRC aren't interested in selling software without bundling expensive specialized hardware, which isn't currently usable by her daughter.

That specialised hardware seems insanely expensive: the Vantage Light is $7,495, and that only gets you the basics. With add-ons -- most of which look to be essential, rather than 'nice to have' extras -- you're looking at $10,000. You'd probably want an extended warranty too: the cheapest option is $465 per annum. They're even flogging what appears to be bog-standard USB Bluetooth dongle for $250.

I've seen devices like this in action, and they are amazing, capable of completely transforming lives, but it really does look like PRC are gouging their customers. Especially when you consider that the hardware is clunky, heavy, hard to use and, according to the post linked to by roomthreeseventeen above, has to be sent back 'at least once a year for big repairs'.

I suppose someone could pay a developer to make the app and then release it as open source, but that's not really the open source model so much as charity, which is perfectly fine, of course, but it's not the same thing.

I imagine a Kickstarter type campaign to develop an open source AAC app for cheap Android tablets would take off like wildfire, but it's a non-starter as long as SCS and PRC's lawyers are trying to sue the competition out of existence.
posted by jack_mo at 7:48 AM on March 26, 2012 [4 favorites]


I imagine a Kickstarter type campaign to develop an open source AAC app for cheap Android tablets would take off like wildfire

Unless the supporters are all users or close friends or family of users, that's still basically charity. Again, not that there's anything wrong with that. I've supported quite a few Kickstarter campaigns, including ones in which I didn't receive any direct benefits. I'm just pointing out that there's a reason the open source software ecosystem has not given rise to as much assistive technology as the proprietary competition, and it generally doesn't have anything to do with patents.

but it's a non-starter as long as SCS and PRC's lawyers are trying to sue the competition out of existence.

Actually, while they might be able to shut down a Kickstarter campaign, it's highly unlikely they could shut down a regular open source project. There are no damages to be had in such a suit, since the programmers have no money (not when compared to the cost of litigation). Even if an injunction was available, enforcing it would be impossible, since development and hosting would simply move to another country. Patents are territorial, and I doubt these companies have patent protection all over the world or the money to pursue litigation on multiple fronts.

That's part of the reason why very, very few open source projects (as opposed to open source software-using companies) have ever been the target of patent litigation. A relative handful of open source programs have left out features because of patent infringement concerns, but actual litigation is quite rare.
posted by jedicus at 8:11 AM on March 26, 2012


*considers moving to a small (preferably tropical) island nation outside the reach of US patent law and running kickstarter campaigns to develop open-source software...*

Who's with me?!
posted by kaibutsu at 8:13 AM on March 26, 2012 [4 favorites]


I'd imagine that after playing with the application most solid Android developers could replicate it, especially since Android already provides the speech synthesis. All the serious man hours go into the organizational and artistic work required to build the LAMP dictionary.

You could perhaps adopt a MAME-like route : You build and distribute an open source Android application that includes only a minimal dictionary, but offers good documentation on expanding the dictionary. Find some dedicated AAC users or parents to help maintain the dictionary if/when others start contributing. In addition, you provide a tool for extracting Speak for Yourself's dictionary, permitting Speak for Yourself owners shift their usage to Android. And presumably their dictionary could live on in piracy land if they get sued out of existence.

I'd imagine a European developer could obtain public financing for this if they jumped through the right hoops, like targeting multiple European languages. You'd also avoid some copyright issues if you focussed on building the non-English open source dictionaries first, only officially adding the English words once they existed in Spanish, German, or whatever.
posted by jeffburdges at 8:15 AM on March 26, 2012


Is it true that Apple can and will uninstall the app if the software company loses? Do purchasers at least get a refund or aer all the users in that case SOL?
posted by Mitheral at 8:18 AM on March 26, 2012


Is it true that Apple can and will uninstall the app if the software company loses? Do purchasers at least get a refund or aer all the users in that case SOL?

Based on what's happened with other apps Apple's pulled, no, there's no remote-uninstall. It just becomes unavailable to any future purchasers.
posted by Tomorrowful at 8:28 AM on March 26, 2012 [2 favorites]


I don't know about can (probably yes), but I still have a game on my iphone that I'd thought was an official release from the game company, but it wasn't. It was pulled from the app store but I'm still happily playing it three or four years later.
posted by rtha at 8:29 AM on March 26, 2012 [1 favorite]


You could perhaps adopt a MAME-like route :

The patents appears to be on the underlying method and don't necessarily require a dictionary to be present in order to be infringed. Even so, I don't think shipping it with a minimal dictionary (or even without a dictionary) would avoid infringement. This is especially true given the possibility of indirect infringement. The only purpose of a program-sans-dictionary would be to be combined with a dictionary and used to infringe the patent, and the project would be distributing it knowing that it would result in direct infringement by end users. That's a recipe for indirect infringement liability.

Is it true that Apple can and will uninstall the app if the software company loses?

It can but as far as I know it never has. There are lots of emulators, tethering software, etc that have snuck in and later been pulled from the App Store, but none have been "kill switched." That includes emulators like the original version of iDOS, which included copyrighted versions of (IIRC) Ms. Pac-Man and Dig Dug. I don't know if any apps have been pulled on the basis of patent infringement, though, so it's hard to say for sure whether Apple would resist remote deletion in such a case.
posted by jedicus at 8:37 AM on March 26, 2012


There's one sentence in the FPP link that bothers me hugely:

"Hardware profits annihilate software profits."

No, they don't. Actually making and selling stuff is a pain: you have to keep inventory, maintain your supply chain, employ workers, buy machinery on credit, etc.

These devices aren't so fearsomely expensive because the makers are necessarily gouging their customers: they are so expensive because they address a small niche market, and its makers must recoup many non-scalable investments with those small sales. Do you have an idea of the cost of just the plastic injection molds to make the casing of such a device? If iPads were sold in the same small numbers, they'd be more expensive by several orders of magnitude.

To start building the devices sold by PRC under license from SCS, the original inventors had to get investors on board. And that's where the patents surely came into play, to reassure those investors that their investment wouldn't be immediately annihilated by copycat devices within months of their market entry. Without the patents, those devices probably wouldn't have been built, and they couldn't have served as models for an iPad app, if this has been the case.

Now tablet computers offer an affordable generic device that can run software allegedly replicating those devices. This is, BTW, why it is generally misleading to speak of "software patents": the particular patents mentioned here are very much hardware patents, referring to a "keyboard". It just so happens that hardware that has become widely available and affordable only later, namely tablet computers, running an appropriate app, may be infringing those patents.

Now, what if SCS and PRC decide (or are forced) to offer a license to Speak to Yourself? well, they may as well close shop then. They'll get a steady income from the license, why should they bother themselves with actually employing people, complying withe health and safety regs, etc? They'll hardly recover the sunken costs of the manufacturing machinery, of course, and their workers' jobs will effectively have been outsourced to Foxconn in China, but that's the cost of progress, isn't it? But next time there's yet another MeFi discussion on the loss of manufacturing jobs, just have a thought about that.

Mind you, those patents only have 3-4 years' validity left in them, so SCS and PRC probably better get used to that brave new world anyway. But they probably deserve a bit more gratitude than being presented as greedy monsters who would like to keep children silent.
posted by Skeptic at 8:45 AM on March 26, 2012 [2 favorites]


There are people who could minimally rewrite the underlying application in one day, jedicus, but a large dictionary might require thousands of hours, i.e. years. You avoid the patent by avoiding U.S. jurisdiction, including the iPad's walled garden, either physically or via foreign anonymous hosting. You're problem then becomes writing the dictionary, hence the MAME stopgap while the dictionary hopefully accumulates over time.
posted by jeffburdges at 8:53 AM on March 26, 2012


I have a software patent to my name.

Software patents are massive bullshit. The company I got the patent with made a solemn promise that they'd only use their patents defensively (and has so far been as good as their word) or else I'd never have consented to it.

In particular, as anyone who's written software knows, ideas are cheap, but actually creating it so that it works is hard.

Now, I've had pretty serious debates with patent lawyers on this issue - they repeat the statement "You can't patent an idea" as if this has some meaning.

First, I spent some time reading the law on this, and this phrase or anything like it does not appear anywhere at all.

But more important, software parents are patents on an idea. You're patenting something that's purely conceptual, and you don't have to have a working model or anything like that.

There's yet another point that most of these software parents are absolutely fucking obvious to an expert practitioner, and thus should not be valid. One click to purchase? Come on. Or this patent? Come on.

One of the actual patents is here. This is ridiculous garbage. I worked on a pocket computer with context-sensitive soft keys in 1988, FFS.

Now, I have to say that it appears to be a hardware patent. So how in heck can they be battling over whether software infringes on this? The patent talks about either a physical keyboard with changing displays, or a physical keyboard with a changing on-screen display - but there's no keyboard here at all!

> But they probably deserve a bit more gratitude than being presented as greedy monsters who would like to keep children silent.

We're supposed to be grateful to people who are trying to make money? Who are selling a bulky hardware item for nearly $10K that could be replicated in software in a tiny unit for less than $1K? For people who prefer to use assault by lawyer to make sure they have control over the market, rather than competing on the relative merits of their product?

It's not like portable computers haven't been around for decades now. If "helping children" were anywhere at all on their radar, they'd be trying to come up with an affordable solution, which would mean "just software". The only possible reason to force people to buy their own proprietary hardware to maximize their profits.

This attitude, that we're supposed to grateful to people who are trying as aggressively as possible to make money out of other people's misfortunes, is pretty darned toxic.
posted by lupus_yonderboy at 9:11 AM on March 26, 2012 [6 favorites]


> There are people who could minimally rewrite the underlying application in one day,

Ridiculous. There are people who could write a hacky demo of the functionality in a day. To come up with a production program, even without the full dictionary, would take months.

People have no idea how much longer it takes to write a production program than a demo that looks similar.
posted by lupus_yonderboy at 9:13 AM on March 26, 2012 [4 favorites]


(Well, perhaps not months, looking at what it does... but definitely weeks...!)
posted by lupus_yonderboy at 9:14 AM on March 26, 2012


You avoid the patent by avoiding U.S. jurisdiction, including the iPad's walled garden, either physically or via foreign anonymous hosting. You're problem then becomes writing the dictionary, hence the MAME stopgap while the dictionary hopefully accumulates over time.

A modular dictionary is a fine feature, but my point was that there doesn't seem to be a need to keep the program and dictionary separate like that in order to avoid patent liability here. The foreign-hosted program could include a dictionary and it wouldn't make a difference, liability-wise.
posted by jedicus at 9:18 AM on March 26, 2012


Yes, it would be possible to build an app on the open source model that got round this (and any) patents. If nobody's making any money at it, it's hard to go after infringers - witness the content industries having to try to turn off the Internet to keep their business model afloat.

What's more, if it's for children (and especially disabled children), then everyone involved has few qualms.

The same sort of thinking meant there was no market for educational software for the BBC Micro, which had been designed to (among other reasons) create just such a market. Teachers copied useful software freely and anyone who tried to make money there, failed.

That's OK. Not all of society's needs can be met most efficiently (or even at all) through the capitalist model. It is usually possible, and often desirable, to organise ourselves differently.
posted by Devonian at 9:23 AM on March 26, 2012 [1 favorite]


First, I spent some time reading the law on this, and this phrase ["you can't patent an idea"] or anything like it does not appear anywhere at all.

Oh? "An idea of itself is not patentable." Diamond v. Diehr, 450 U.S. 175 (1981) (quoting Rubber-Tip Pencil Co. v. Howard, 20 Wall. 498, 507 (1874)).

More generally, the relevant phrases are "utility," "enablement," and "abstract idea."

Utility means a patent must claim something useful. An abstract idea (e.g. a mathematical formula, by itself) is not useful because it doesn't do anything. It just sits there. For example, Gauss's law for magnetism, by itself, does not tell me anything about which way is north or anything else useful.

Enablement means that a patent must enable a person having ordinary skill in the art to make and use the invention without undue experimentation. A patent that solely claims an abstract idea is not enabled because it doesn't explain how to use it.

Finally, abstract ideas are one of the judicially-created exceptions to patentable subject matter, so they're unpatentable anyway. Given the utility and enablement requirements it's not clear that it makes sense to have a separate exception, but there it is.

But more important, software parents are patents on an idea. You're patenting something that's purely conceptual, and you don't have to have a working model or anything like that.

You don't have to have a working model, but you do have to have an enabling disclosure. Now, I think that the definition of "enabling" should be tightened up, but that's a different issue.

These patents in particular seem to cover a useful device that happens to use software to accomplish its function. It's not clear that they claim an abstract idea. An abstract idea might be something like "I claim helping someone to express themselves via an assistive device." The difference there is that the abstract idea claim isn't useful; there's no practical application there. And such a patent is unlikely to enable a person of ordinary skill to make and use the "invention."

This is ridiculous garbage. I worked on a pocket computer with context-sensitive soft keys in 1988, FFS.

Well then it should be a cinch to invalidate, perhaps via reexamination, which is extremely cheap relative to litigation and actually quite cheap in general.

Now, I have to say that it appears to be a hardware patent. So how in heck can they be battling over whether software infringes on this? The patent talks about either a physical keyboard with changing displays, or a physical keyboard with a changing on-screen display - but there's no keyboard here at all!

The claims don't require a hardware keyboard. The '303 patent, for example, talks about an "integrated input and display device for displaying a plurality of keys of a displayed keyboard." Claim 9 further claims a device "in which the integrated input and display device is a touch-panel screen."
posted by jedicus at 9:39 AM on March 26, 2012


This attitude, that we're supposed to grateful to people who are trying as aggressively as possible to make money out of other people's misfortunes, is pretty darned toxic.

I suggest reading this. SCS is an IP holding company for Bruce R. Baker, a linguist who has been working on AAC since the 1980s. He pretty much appears to be the pioneer in that field. PRC is its exclusive licensee. This is a common arrangement: inventor invents stuff, gets patents and other rights, licenses its IP, including your know how, exclusively to some capitalists who put their own money on the line to translate those ideas into actual stuff, because otherwise your ideas will remain just that, ideas, and not be put to any use.

In short, without those people, AAC devices probably wouldn't be on the market, and neither would any iPad app.

Most of the time those capitalists don't just put their own money into the venture, but also that of other people, that is, they take on debt. If you are in hock for a substantial amount of money, you probably won't be very happy if somebody starts undercutting your business by 90%, no matter how beneficial this invention is for humanity. You may even become quite aggressive, as will your debtors.

Portable computers may have been around for a while, but only modern tablet computers' touchscreens appear to have made it practical to make one such device without a dedicated keyboard. Do you remember the touchscreens just ten years ago, their price, cost and sensitivity?

Neither PRC nor Speak for Yourself are charities. But charities haven't built those devices, or even an open-source app doing the same. When that happens, then maybe your indignation will be more justified. In the meantime, maybe people should be at least a trifle grateful to those actually making such stuff, even if, oh heavens, they actually would like to be paid for their effort, if only to pay their mortgages.
posted by Skeptic at 10:15 AM on March 26, 2012 [1 favorite]


And call me cynical and a Devil's Advocate, but I actually will rather be grateful to a dirty capitalist putting his money on the line to develop and produce a new medical device, even if it is sold at an insane price, than to an anonymous Internet blowhard.
posted by Skeptic at 10:19 AM on March 26, 2012 [1 favorite]


jedicus:

I've read all of these more than once before.

You can patent a purely conceptual thing that you have never built a model of. If that's not an "idea", what is?

I understand that patent lawyers use the term "idea" as some sort of technical legal term that isn't the same as the English language, but I frankly have not been able to get a straight answer as to how their "idea" does not include purely conceptual things without either a software or a hardware model.

> > context-sensitive softkeys.

> Well then it should be a cinch to invalidate, perhaps via reexamination, which is extremely cheap relative to litigation and actually quite cheap in general.

When we chatted to a patent lawyer, we were told that if we were in a patent battle, it'd cost us $100K before we even stepped into a courtroom. (He wasn't arguing against patents - he was saying we should amass an array of patents to scare off the other guy so we'd never have to step into court...)

People have been using context-sensitive softkeys since before we did - for at least 25 years now. I'm sure the defense knows all about this - if it were so easy, they'd have done this already.

I've been very interested in this, because I write software for a living. It seems to me, and many other "expert" practitioners in the field, that the courts consider "prior art" to be extremely narrow, and consider patents to be extremely wide.

Again, it bothers me deeply that make things are patented that have been in use by engineers for decades, or are completely obvious to an expert practitioner in the field. I mean, "one click to purchase" - how is this not obvious?

> The claims don't require a hardware keyboard.

Thanks, good to know... so we can go back to claiming that software patents are garbage.
posted by lupus_yonderboy at 10:21 AM on March 26, 2012 [3 favorites]


jedicus: The difference there is that the abstract idea claim isn't useful; there's no practical application there.

There are two broad families of cases I see here

1)the idea is non-trivial and manifesting the idea into functional reality is also non-trivial. Arguably, patents for such processes adhere to the spirit of the scope of patents

2)the idea is non-trivial but manifesting the idea into functional reality is trivial or just a matter of plain competent engineering. In these cases, although it is the functional disclosure which is patented, in effect, it is the idea which is being protected. In case of software patents, once a person understands the mathematical/logical underpinnings of an idea, it can be re-presented in a 'functional' garb of code by a competent coder. So the 'idea' exception doesn't have potency in these cases.
posted by Gyan at 10:22 AM on March 26, 2012


> But charities haven't built those devices, or even an open-source app doing the same.

The point of this article is that this is probably legally difficult or impossible to do.

> And call me cynical and a Devil's Advocate, but I actually will rather be grateful to a dirty capitalist putting his money on the line to develop and produce a new medical device, even if it is sold at an insane price, than to an anonymous Internet blowhard.

You think you should be grateful to someone trying to look out for their own interests? That's a very strange concept.
posted by lupus_yonderboy at 10:27 AM on March 26, 2012 [1 favorite]


So we can go back to claiming that software patents are garbage.

IIRC There's literally no person who has argued otherwise on Mefi (and possibly the entire world) who isn't a patent lawyer or similar. And while i appreciate that they show up every time to make the case, it has to said: sorry, software patents are garbage.
posted by Artw at 10:31 AM on March 26, 2012 [1 favorite]


You think you should be grateful to someone trying to look out for their own interests? That's a very strange concept.

Then you probably don't feel grateful for doctors, nurses, firemen, teachers, etc., because, after all they're all getting paid? (Badly paid, but paid nonetheless).
posted by Skeptic at 10:31 AM on March 26, 2012


Let me add that I don't think absolutely every software patent should be invalid. For example, AT&T came up with a major improvement on the linear programming algorithm. I won a prize in high school for writing a linear programming application, and I frankly barely understand this work of AT&T's.

It's clearly novel and original and has serious and obvious applications, and richly deserves a patent.

But most software patents are not in this category.

I was going to comment on the two patents I have (I thought I only had one but apparently I have two, go figure) - link to them and point out that I believe that any competent practitioner, faced with the same problem, would come up with the same solution. But, I don't think it's fair to my ex-employer who might want to use them someday...
posted by lupus_yonderboy at 10:32 AM on March 26, 2012


I'm just pointing out that there's a reason the open source software ecosystem has not given rise to as much assistive technology as the proprietary competition, and it generally doesn't have anything to do with patents.

Oh, definitely - I agree with you on the 'scratch an itch' problem in open source.

It turns out there is some open source development activity around AAC: Ithica, which looks like it's far from being a practical solution and wouldn't run on today's tablets (do they even make those Windows 'Tablet PC' things anymore?).

Actually, while they might be able to shut down a Kickstarter campaign, it's highly unlikely they could shut down a regular open source project.

That's interesting. I'd just assumed that if some bit of software infringed a patent, the patent holder could have the project shuttered. I forgot about money ;-)

And call me cynical and a Devil's Advocate, but I actually will rather be grateful to a dirty capitalist putting his money on the line to develop and produce a new medical device, even if it is sold at an insane price, than to an anonymous Internet blowhard.

I'm with you up to the insane price part!
posted by jack_mo at 10:35 AM on March 26, 2012


> Then you probably don't feel grateful for doctors, nurses, firemen, teachers, etc., because, after all they're all getting paid? (Badly paid, but paid nonetheless).

If, for example, teachers were paid a lot more, and were attempting to use the law to make sure that no one else was allowed to teach, then I would stop being grateful to them.

And I do believe that Minspeak itself might be patentable (I'm sort of out of time to look at it in detail). A lot of work went into his specific icons and dictionary (though surely a copyright is the correct way to protect this work?)

What I believe is invalid are the patents under discussion here, patents which seem to cover any virtual keyboard which changes symbols based on context. It seems to me that they're using these patents to crush any competitor, even one that doesn't use their own and probably superior icon and organization system.
posted by lupus_yonderboy at 10:39 AM on March 26, 2012


do they even make those Windows 'Tablet PC' things anymore?

Well, Microsoft seems to be staking it's entire future on them, make of that what you will.
posted by Artw at 10:41 AM on March 26, 2012


> it's highly unlikely they could shut down a regular open source project.

....because?

Suppose I went to github (a prominent open-source hub) with my high-priced lawyer and said, "This project impinges on my intellectual property." They're almost certainly going to shut that project down - because they rightfully don't want to get into such a war which would cost them a lot of money and profit them not at all.
posted by lupus_yonderboy at 10:42 AM on March 26, 2012


Let me add that I don't think absolutely every software patent should be invalid. For example, AT&T came up with a major improvement on the linear programming algorithm. I won a prize in high school for writing a linear programming application, and I frankly barely understand this work of AT&T's.

It's clearly novel and original and has serious and obvious applications, and richly deserves a patent.

But most software patents are not in this category.


Then we actually are close to an agreement. This said, you may be suffering from confirmation bias, since the patents that usually get into the news usually aren't the most deserving (press thrives on controversy, and patents are in general pretty boring, the more so the more technical they get).

Yes, a lot of crap patents have been granted, in particular in the late 90s and early 00s. Yes, patent litigation is too expensive, at least in the US (patent litigation elsewhere is rather less expensive). And yes, it ought to be easier to invalidate crap patents. Most practitioners in the field (at least the ones with a conscience), will be the first to say so. But having a conscience, and being also occasionally confronted with deserving inventions, even when software is somehow involved, one must also warn against throwing the baby out with the bathwater.
posted by Skeptic at 10:42 AM on March 26, 2012


It's clearly novel and original and has serious and obvious applications, and richly deserves a patent.

But most software patents are not in this category.


To me I think the big difference is software patents that are actually patenting some sort of novel algorithm to fit some requirements, versus patents that are basically just patenting the requirements themselves. Someone with absolutely no programming knowledge whatsoever could write a software patent for a program that displays a dynamic menu of words. In the physical world you can't just patent the concept of making a device that converts gasoline to mechanical energy, you can only patent a particular engine. And if someone figures out how to solve the same problem with a different machine then they can get around your patent. Someone with no idea how to build an engine shouldn't be able to file a patent that locks out everyone else from making engines. But even for patents of novel algorithms, I still think they do more harm than good.
posted by burnmp3s at 12:14 PM on March 26, 2012 [1 favorite]


I've read all of these more than once before.

Well you completely missed the fact that "ideas aren't patentable" is well established in patent law, so you might want to read them again.

You can patent a purely conceptual thing that you have never built a model of. If that's not an "idea", what is? ... I frankly have not been able to get a straight answer as to how their "idea" does not include purely conceptual things without either a software or a hardware model.

Suppose I invent a mouse trap by thinking really hard about how to trap mice, but I don't actually build a model because I'm really good at modeling mechanical devices in my head. Now suppose my patent contains detailed drawings with a precise bill of materials and instructions on how to build and use the trap. A competent mechanical engineer does not need to see a working model in order to build the trap; the designs and instructions are sufficient. So my claims are to a piece of hardware that has never been built yet it is not a claim to an idea.

Contrast that with a claim that simply says "I claim an improved mouse trap that traps the mouse quietly and humanely and can hold multiple mice." That's just an idea. It doesn't enable someone with ordinary skill to build and use a better mouse trap that does those things.

When we chatted to a patent lawyer, we were told that if we were in a patent battle, it'd cost us $100K before we even stepped into a courtroom.

First, this is presuming litigation rather than the reexamination route. Reexam is much, much cheaper than litigation. If the defense is based primarily on invalidity rather than unenforceability or non-infringement, it's a very attractive option.

Second, even litigation doesn't have to cost that much. It often does because wealthy companies with a lot of money on the line are the ones fighting the battle, but it doesn't have to. There are a wide range of patent attorneys who charge a wide range of rates. Not every case needs or benefits from a large litigation team.

Third, the "before we even stepped into a courtroom" part is very valuable time. During that time prior art is found, arguments are developed, and negotiations are made. Given serious evidence of invalidity or non-infringement, a patentee may well drop the case or settle. The vast majority of patent disputes never go to a decision on the merits because they tend to be dropped or settled.

People have been using context-sensitive softkeys since before we did - for at least 25 years now. I'm sure the defense knows all about this - if it were so easy, they'd have done this already.

The patent does not claim context-sensitive softkeys in general. It claims a particular application of such keys such that they display polysemous words in a particular way so as to facilitate communication. One would either need to find prior art showing that this is an obvious use of context-sensitive softkeys or prior art showing this particular use.

What I believe is invalid are the patents under discussion here, patents which seem to cover any virtual keyboard which changes symbols based on context.

No, the claims are much narrower than that. Even the broadest claims speak in terms of polysemous symbols ("also known as icons"), for example.
posted by jedicus at 12:20 PM on March 26, 2012 [1 favorite]


Suppose I went to github (a prominent open-source hub) with my high-priced lawyer and said, "This project impinges on my intellectual property." They're almost certainly going to shut that project down - because they rightfully don't want to get into such a war which would cost them a lot of money and profit them not at all.

Okay, so the project moves to hosting in a different country. If the patentee doesn't have a patent in that country, it's SOL. If it does, it has to pay lawyers in that country to threaten another hosting provider. So the project moves again. Since "moving" is essentially free, this is an untenable solution for the patentee.

Similarly, development could be done either anonymously or by developers in a country where there is no relevant patent. Really, non-commercial, volunteer open source development is quite difficult to stop via patents.
posted by jedicus at 12:26 PM on March 26, 2012


AT&T came up with a major improvement on the linear programming algorithm. [...] It's clearly novel and original and has serious and obvious applications, and richly deserves a patent.
I'm not so sure.

The FFT was co-invented (re-discovered, as they found out later) by an IBM employee. They decided not to patent it because the other co-inventor was not from IBM. This is not how things should work -- I should not be breathing a sigh of relief that the FFT wasn't re-discovered totally in house at IBM.

Can you imagine the state of "progress" that patent would have engendered?

One could argue that in this particular case, once people figured out the Gauss had invented it prior, that we'd be good to go. But it doesn't really work that way. In a world where a large corporation can patent "one-click" through force of will, and keep the patent, and another can patent essentially the working of the modern desktop and loose to a company with more money... I'm going to call major bullshit.

I do not believe in the power of "patents" over gold. Patent lawyers of course do, because they get paid either way. But I remain skeptical of their intentions, frankly.
posted by smidgen at 12:33 PM on March 26, 2012 [1 favorite]


I've seen an awful lot of solutions proposed to the patent problems, but "just let the lawyers sort it out" doesn't sound particularly wise.

We'd never be arguing over this if patent only applied against "dominant market participants" for example, i.e. maybe SCS/PRC could sue apple for distributing this application, but they couldn't sue a smaller competitor with less market access.

All such proposals miss the point that ideas and algorithms simply aren't the costly aspect of bring the product to market and should not be the part protected by law. Algorithms are in-fact nothing but equations, which shouldn't be patentable, no matter how clever.

There is a clearly costly part of software like this, namely the hierarchically laid out dictionary, but that's protected by copyright not patents. It's clear their the patent on redefining keys is simply ridiculous, no matter how much AAC experience went into the dictionary's implementation.

Amusingly, I compromised on "minimally rewrite" after editing out exactly the term "demo" from my comment, lupus_yonderboy, mostly not wanting to disparage anyone with real Android expertise, well some people are exceedingly fast. Also, even a minimal version still violates the patent, regardless.
posted by jeffburdges at 12:34 PM on March 26, 2012


IIRC There's literally no person who has argued otherwise on Mefi (and possibly the entire world) who isn't a patent lawyer or similar.

When Steve Jobs announced the original iPhone, he noted "and boy have we patented it." It's right there in the presentation as a bullet point, alongside "works like magic" and "no stylus." Later he pointed out that "We filed for over 200 patents for all the inventions in iPhone and we intend to protect them." That strikes me as fairly solid evidence that he did not believe that all software patents were garbage.
posted by jedicus at 12:34 PM on March 26, 2012


Algorithms are in-fact nothing but equations, which shouldn't be patentable, no matter how clever.

You can't patent an equation or an algorithm. What you can patent is a useful application of an algorithm or equation. An equation is not useful, and neither is an algorithm. Staring at Maxwell's equations all day long will not produce a compass, a generator, or a flashlight. The fact that computer programs are essentially mathematical proofs does not make them unpatentable because the math isn't being patented but rather the useful application of the math.

It's no different from the fact that the design of, say, the Wankel rotary engine is defined by math. A patent on the Wankel engine is not a patent on the epitrochoid or the equations that define it. It's a patent on a useful application of that shape.
posted by jedicus at 12:40 PM on March 26, 2012


When Steve Jobs announced the original iPhone...
...That strikes me as fairly solid evidence that he did not believe that all software patents were garbage.
This logic doesn't follow. SJ cared about his own interests, not society's. We're arguing about society in general -- I don't care about Apple's patents. And I don't think SJ really did care other than as a way to slow down competitors. For him it was a tool he made use of, not a social necessity.

Perhaps it's also worth noting, the two iPhone ripoffs come from two very large companies. You may want to consider why they can do this so successfully.
posted by smidgen at 12:49 PM on March 26, 2012 [1 favorite]


Jeez, a patent on the fast-foiuer transform would've been disastrous, smidgen, holy crap we dodged a bullet there.

You aren't really strengthening your case by citing an egomaniacal captain of industry like Jobs, jedicus. There are certainly bit players outside the legal system who remain pro-patent for various reasons, sometimes fantastical, sometimes financial, whatever.

There are not however many people actually doing the "important stuff" who support patents anymore. We know they serve no function except to slow and tax our efforts on behalf of the legal profession and executive cast.

We can and do implement software with nothing but equations describing the algorithms, interactions, etc., jedicus. I'm aware the legal profession hammers that distinction, but reality dumped it a couple decades ago.
posted by jeffburdges at 12:50 PM on March 26, 2012 [1 favorite]


This logic doesn't follow. SJ cared about his own interests, not society's. We're arguing about society in general -- I don't care about Apple's patents. And I don't think SJ really did care other than as a way to slow down competitors. For him it was a tool he made use of, not a social necessity.

I was responding to the claim that "IIRC There's literally no person who has argued otherwise on Mefi (and possibly the entire world) who isn't a patent lawyer or similar." (emphasis added). Steve Jobs was literally a person and, in my view, he argued that software patents are a useful tool for business and thus not garbage. This tends to disprove the specific claim made earlier. I did not claim that Jobs's approval of software patents is evidence that they are a social necessity or even that they are, in fact, not garbage. Just that someone who isn't a patent lawyer or similar has argued in favor of them.

And anyway, the fact that people on MeFi don't like software patents hardly proves that they aren't in society's interest. The social utility of software patents is an empirical question that has not been solidly proved either way, certainly not here. And even if software patents are not socially useful in their current form, that does not prove that they cannot be adequately reformed without throwing them out entirely.

You aren't really strengthening your case by citing an egomaniacal captain of industry like Jobs, jedicus. There are certainly bit players outside the legal system who remain pro-patent for various reasons, sometimes fantastical, sometimes financial, whatever.

Steve Jobs (and by extension Apple) was a "bit player?" That's an interesting take.
posted by jedicus at 1:02 PM on March 26, 2012


We can and do implement software with nothing but equations describing the algorithms, interactions, etc., jedicus.

Really? So you can write an equation on a piece of paper and it will execute the described algorithm, producing a useful result? All that from nothing but paper and graphite? I got my master's degree in computer science just a few years ago and they didn't cover that at all. I feel cheated.

Or wait. Maybe you mean that you can define a program using nothing but equations and a computer can execute it, producing a useful result, perhaps in conjunction with other physical devices to provide the input or to make use of the output. That doesn't sound very much like "nothing but equations."
posted by jedicus at 1:07 PM on March 26, 2012


I implied you should've quoted someone who wasn't planning on starting an insane patent war with half the world, i.e. a bit player.

You should probably check out Donald Knuth's Letter to the Patent Office. Any code written in Haskell is transparently equations. Any 3d model for a 3d printer is similarly just parameters for an equation. You could argue a difference in intention between whether the mathematics is rendered in code or not, but the infringing part is still nothing but an equation.
posted by jeffburdges at 1:16 PM on March 26, 2012


That doesn't sound very much like "nothing but equations."

It is very much like "nothing but equations". Software is a "set of equations" one applies to a general-purpose computer in order to make it do something specific. That is all that software is. All complete formal descriptions of an algorithm are isomorphic; that is, there is no fundamental difference between the algorithm, written down on paper with graphite, and the same algorithm written down in bits. It is the same algorithm, it is the same structure. The representation is immaterial. You don't "invent" something by copying the same data from paper to computer; you just type it in.

Furthermore, there's no reason in principle one couldn't set up a computer capable of OCRing its input from paper and executing it; this would just be a very inefficient way of doing things. In principle, one could indeed write programs on paper using nothing but a graphite pencil, and then have a computer execute them directly.

There might well be patentable inventions in the hardware used to accomplish such a feat, but you simply cannot patent the software algorithms without allowing people to patent the very ideas themselves.
posted by Mars Saxman at 1:25 PM on March 26, 2012 [3 favorites]


Jeffburdges, that letter is a prime example of someone "not getting" what the debate is about. AFAIK, nobody has ever pretended that there is anything like a "nonmathematical" algorithm. But there is a difference between an algorithm or an equation, as a purely abstract entity, and a materially useful application of such an algorithm or equation, whether it is embodied in a device or in a process.
posted by Skeptic at 1:27 PM on March 26, 2012 [1 favorite]


OK, patent layers AND CEOs of hugely litiguous companies now sueing people over the concept of "black" or "rectangular".
posted by Artw at 1:30 PM on March 26, 2012


Mars Saxman, what many people don't understand is that there is a difference between "a set of equations" and actually carrying out a succession of actions. People have this idea of patents as protecting physical "gizmos", whereas patents on processes and methods are just as old and as important as patents of devices and apparatuses.
posted by Skeptic at 1:33 PM on March 26, 2012 [1 favorite]


yeah, you need a vague flowchart and some handwaving.
posted by Artw at 1:35 PM on March 26, 2012 [1 favorite]


Artw, it's easy to be flippant about Apple's design patent litigation, but it really wasn't about "the concept of black or rectangular". It about Samsung using a confusingly similar design to Apple's. It's worth noting that the court decision blocking the Galaxy Pad's sale in Germany finally wasn't based on infringement of Apple's registered design (as design patents are known in Europe), but on the concept of unfair competition. And Samsung easily circumvented that ban with a slight redesign making the Galaxy Pad a bit more different from the iPad.
posted by Skeptic at 1:42 PM on March 26, 2012


Is there prior art in text messaging interfaces for Asian language mobile phones?
posted by BrotherCaine at 1:46 PM on March 26, 2012


t's easy to be flippant about Apple's design patent litigation

Yes, it is, it really really is.
posted by Artw at 1:49 PM on March 26, 2012 [1 favorite]


It is very much like "nothing but equations". Software is a "set of equations" one applies to a general-purpose computer in order to make it do something specific. That is all that software is. All complete formal descriptions of an algorithm are isomorphic; that is, there is no fundamental difference between the algorithm, written down on paper with graphite, and the same algorithm written down in bits. It is the same algorithm, it is the same structure. The representation is immaterial. You don't "invent" something by copying the same data from paper to computer; you just type it in.

The algorithm is the same, but putting it on a computer makes it useful in a way that the same algorithm on a piece of paper fundamentally is not.

Think of it like a CAD document that perfectly captures the design of a mechanical device. I mean truly the Platonic ideal of CAD documents. That document is, in one sense, "the same" as the physical device in that it represents the same information. But the physical device is useful in a way that the document is not. That usefulness is at the heart of the distinction between an abstract algorithm and a patentable invention.

There might well be patentable inventions in the hardware used to accomplish such a feat, but you simply cannot patent the software algorithms without allowing people to patent the very ideas themselves.

If I patent "algorithm X running on a general purpose computer such that it produces useful result Y," then I have not patented the idea itself. A computer scientist would still be free to analyze the algorithm on pencil and paper, for example, or to derive an improved version. It's no more a patent on the algorithm itself than a patent on a mechanical widget forbids someone from analyzing a simulated version of the widget on a computer.
posted by jedicus at 1:58 PM on March 26, 2012


skeptic: there is no difference when it comes to software. This is what makes software so different from other engineering disciplines: the blueprint is the product. The product is complete once you have described it in sufficient detail; there is no fabrication process - or, rather, the fabrication is handled automatically and transparently, by machines. You don't make software; you merely specify it.

When I was a kid I used to "invent" all kinds of fabulous devices by drawing sketches of them. I remember one rube-goldbergian contraption in which I was somehow going to press a record player into service as part of an air compressor. But of course I was a kid and had no idea how these things would actually be made, and so I unwittingly hand-waved across all the hard parts of the problem.

Imagine I had been able to patent this nebulous device, and then actually go after manufacturers of actual air compressors based on my patent! Some other team of engineers, knowing nothing of my "work", does all the actual hard labor involved in figuring out how to practically create such a device, and then this guy comes in out of nowhere and claims that since he thought of something similar once, he gets money? This makes no sense.

But this is exactly what happens when you patent software. If the patent is vague enough to cover an independent reinvention of a similar project, then the patent is vague enough that it does not include the actual engineering work involved in developing that project. Whoever holds the patent can extract rent from others without ever giving them anything in the first place.

The whole point of a patent system is to encourage inventors to publish their inventions by giving them a temporary legal monopoly. This may make sense in hardware, when you have to do a lot of engineering to figure out how to make such and such a thing happen - but in software, if you've done that level of engineering work, what you have is a working piece of software! And if you haven't done that level of engineering work, then what you have to offer is just an idea, and somebody else has to do the actual labor of making software out of it. Ideas are cheap, ideas are worthless; it's actual working code that means something.

The tragedy of software patents is that they destroy entrepreneurship without offering anything in return. It becomes very dangerous to write and publish software, because people you've never heard of can come along and claim legal monopoly on parts of your work, whether the piece of crap they've rammed through the patent office actually contains any useful work or not. It doesn't matter if you could have gotten any benefit from their patent even if you had ever known it existed; they can still sue you, and you're dead.
posted by Mars Saxman at 2:05 PM on March 26, 2012 [3 favorites]


You should probably check out Donald Knuth's Letter to the Patent Office.

I've read it. Prof. Knuth is a brilliant computer scientist, but he does not understand the law here and makes the same mistake made by several people in this thread. Computer programs are math, and math is not patentable. But software patents do not claim computer programs in isolation. They claim computer programs running on a computer, and a computer is made of silicon, not abstract mathematical concepts.

You could argue a difference in intention between whether the mathematics is rendered in code or not, but the infringing part is still nothing but an equation.

No, the infringing part must include each and every element of the claimed invention, which would include a computer in this case. The code standing alone on a piece of paper would not meet that standard.
posted by jedicus at 2:06 PM on March 26, 2012


Didn't we just have this conversation?

IIRC There's literally no person who has argued otherwise on Mefi (and possibly the entire world) who isn't a patent lawyer or similar.

If after "patent lawyer" you add in ", someone who is personally profiting from software patents" you'd be pretty right.

In this case (going from the reports linked so far), the worst part isn't that these companies have patents over something that is so clearly a set of ideas, it's that they seem to be using their patents over these ideas to require their customers to buy a piece of insanely expensive, outdated and apparently unreliable hardware along with the licence to use the idea. And the piece of hardware is too large for small children. And the ideas could be implemented on general-purpose hardware for a tiny fraction of the price.

Sure, the patents could be challenged, but that would cost huge amounts of money and might not work anyway. This is idiocy.
posted by A Thousand Baited Hooks at 2:07 PM on March 26, 2012 [1 favorite]


Think of it like a CAD document that perfectly captures the design of a mechanical device. I mean truly the Platonic ideal of CAD documents. That document is, in one sense, "the same" as the physical device in that it represents the same information.

You could make the same meaningless distinction between data stored on your hard disk, where it is just a bunch of magnetic states in a slab of iron, and data which is actually loaded into memory, where it can be executed. Or you could make the same meaningless distinction between data which is stored on a CD-ROM and data stored on a hard drive. Or you could make the same meaningless distinction between data which is stored on paper, in the form of punch-cards, and data which is stored on a tape drive.

But it's all data. All you are doing when you take that perfect specification from one form to another is copy bits, and that's what computers do.
posted by Mars Saxman at 2:08 PM on March 26, 2012 [1 favorite]


Computer programs are math, and math is not patentable. But software patents do not claim computer programs in isolation. They claim computer programs running on a computer, and a computer is made of silicon, not abstract mathematical concepts.

And that is exactly why the law is broken: it allows you to patent math, by simply tacking on "...and you do it with a computer".
posted by Mars Saxman at 2:08 PM on March 26, 2012 [4 favorites]


And then add further patents on "...and do it on a remote computer over an internet connection" or "do it on a computer shaped like a phone with a telephony capability".
posted by Artw at 2:11 PM on March 26, 2012 [1 favorite]


You don't make software; you merely specify it.

But you must specify it in sufficient detail. It is not enough to tell a compiler "This program sorts its input." One must define what one means by "sort" and "input."

This is an important point in patent law as well. A patent can't merely set out a goal, desired result, or a research program. It has to enable a person of ordinary skill to make and use the invention without undue experimentation. A vague software patent that requires undue effort for someone of ordinary skill to implement should fail that test.

Imagine I had been able to patent this nebulous device, and then actually go after manufacturers of actual air compressors based on my patent! Some other team of engineers, knowing nothing of my "work", does all the actual hard labor involved in figuring out how to practically create such a device, and then this guy comes in out of nowhere and claims that since he thought of something similar once, he gets money? This makes no sense.

The patent would fail for lack of enablement.

But this is exactly what happens when you patent software. If the patent is vague enough to cover an independent reinvention of a similar project, then the patent is vague enough that it does not include the actual engineering work involved in developing that project.

Your conclusion doesn't actually follow, but even so: the solution here (as I've said many times) is tightening the enablement requirement in the context of software. The patent should actually enable the claimed invention, not just vaguely describe the goal and say "the computer does some calculation and the goal is achieved." Solving this issue does not require tossing out software patents in general.

This may make sense in hardware, when you have to do a lot of engineering to figure out how to make such and such a thing happen - but in software, if you've done that level of engineering work, what you have is a working piece of software! And if you haven't done that level of engineering work, then what you have to offer is just an idea, and somebody else has to do the actual labor of making software out of it. Ideas are cheap, ideas are worthless; it's actual working code that means something.

Which is why I think enablement should be tightened up to the point that, in general, it could only be satisfied with code or nearly-complete pseudocode rather than vague flowcharts and UI diagrams.
posted by jedicus at 2:13 PM on March 26, 2012


You could make the same meaningless distinction between data stored on your hard disk, where it is just a bunch of magnetic states in a slab of iron, and data which is actually loaded into memory, where it can be executed.

Meaningless distinction? Oh good. I'm going to start a business where I sell mechanical widgets. When someone orders one of my widgets I'll just send them a CAD document. When they complain, I'll tell them they're making a meaningless distinction.

And that is exactly why the law is broken: it allows you to patent math, by simply tacking on "...and you do it with a computer".

Suppose I had a patent on "sorting numbers using a computer." Now suppose you invented a new sorting algorithm, analyzed its performance, wrote up a paper about it, etc, but never implemented it on a computer. I could not sue you for patent infringement because you never sorted numbers on a computer, as required by the patent. So how, again, is my patent a patent on math, when math can't infringe the patent?
posted by jedicus at 2:17 PM on March 26, 2012


And then add further patents on "...and do it on a remote computer over an internet connection" or "do it on a computer shaped like a phone with a telephony capability".

You're confusing obviousness with patentable subject matter. I'm assuming here that the algorithms and data structures are new and nonobvious. Taking a known or obvious algorithm or data structure and putting it on a computer, or a networked computer, or a mobile device is very likely obvious and the courts have been invalidating such patents quite frequently lately.
posted by jedicus at 2:18 PM on March 26, 2012


Which is why I think enablement should be tightened up to the point that, in general, it could only be satisfied with code or nearly-complete pseudocode rather than vague flowcharts and UI diagrams.

How would such a patent differ from copyright? People already have monopoly control over their implementations; why offer a second monopoly regime for the same class of artifacts?
posted by Mars Saxman at 2:26 PM on March 26, 2012 [1 favorite]


And lest there be any confusion: I'm not saying that these patents are necessarily valid or that the patentee is being reasonable here. There are tons of likely-invalid software patents, just like all patents, and these might be some of them. I'm arguing that prior art, enablement, utility, and eliminating the presumption of validity all make for better ways to weed out bad patents than patentable subject matter. And focusing on those issues would have the added benefit of improving the quality of all patents, not just software patents.

In this case I think the best solution is either for the patentee to buy the iPad app company and sell the app themselves or for them to license the patents to the app company. It's clearly a superior product for many people, many of whom cannot use or afford the patentee's product anyway, so it's dumb for the patentee to leave that money on the table while potentially spending quite a bit trying to enforce their patents. In a few years their patents will expire and the app company's successor will eat their lunch anyway, so they might as well enter the app market sooner rather than later.
posted by jedicus at 2:29 PM on March 26, 2012


How would such a patent differ from copyright? People already have monopoly control over their implementations; why offer a second monopoly regime for the same class of artifacts?

Because they don't protect the same thing. Copyright doesn't protect the functional aspects of a computer program, just the creative expression, whereas a patent only protects the functional aspects and not the creative expression.

The same is true of, for example, a kinetic sculpture that might have patentable functional aspects (e.g. a clever arrangement of gears) and copyrightable creative aspects.
posted by jedicus at 2:32 PM on March 26, 2012


Now suppose you invented a new sorting algorithm, analyzed its performance, wrote up a paper about it, etc, but never implemented it on a computer. I could not sue you for patent infringement because you never sorted numbers on a computer, as required by the patent. So how, again, is my patent a patent on math, when math can't infringe the patent?

The requirement of a computer implementing it makes sense for a patent enforcement perspective, but I don't see how the distinction is all that relevant when you are talking about the new piece of intellectual property that is being created and what that piece of intellectual property consists of. If someone patented a new narrative structure for a novel for example, that would still fundamentally be a patent on storytelling, even if the patent itself was only enforceable on physical books that with words printed on them that used the storytelling structure. The fact that a book containing a story is different than the story itself is a valid point but in my mind it's irrelevant when considering whether or not it makes sense to allow people to own a monopoly on a storytelling method.
posted by burnmp3s at 2:33 PM on March 26, 2012


I'm assuming here that the algorithms and data structures are new and nonobvious. Taking a known or obvious algorithm or data structure and putting it on a computer, or a networked computer, or a mobile device is very likely obvious and the courts have been invalidating such patents quite frequently lately.

So:
new pure algorithm => unpatentable
putting existing algorithm on computer => unpatentable
BUT
putting new pure algorithm onto computer => patentable?

I agree with Mars Saxman; if you can patent an algorithm merely by running it on a computer through a trivial and well-known transformation, then it's a meaningless and misleading distinction to say that pure algorithms cannot be patented.
posted by Pyry at 2:34 PM on March 26, 2012 [2 favorites]


Jedicus, here's the outcome I want:

- If I write a piece of software, on my own, from scratch, and I release this program, under a copyright license of my choosing;

- Nobody can come along out of nowhere and demand that I pay them a license fee just because they happened to write something similar once.

That's it. I don't care what you call the legal structure, this is what needs to happen. The current law does not allow this, so it is broken.

It's reasonable that I shouldn't be able to just copy someone else's code without their permission, but copyrights already protect that. It is not reasonable that the first person to implement an algorithm should be able to prevent everyone else from independently reimplementing that idea, whether they actually make any use of the first invention or not.
posted by Mars Saxman at 2:36 PM on March 26, 2012 [3 favorites]


In this case I think the best solution is either for the patentee to buy the iPad app company and sell the app themselves or for them to license the patents to the app company.

But they'll never do this, because their business model is based on charging thousands of dollars for a piece of hardware that can be replaced by an iPad. They'd never be able to charge $8,000 (minus the actual cost of their hardware) for an iPad app.

Hopefully the app maker will be able to get some kind of compulsory licence with a reasonable royalty. I'm not a patent attorney, so I'm not sure whether this ever actually happens (although years ago when I did a course on Australian IP law the lecturer said it was extremely rare), but I bet a few years of obstructionary litigation by the patent-holder could easily stretch the fight until after the patents are due to expire anyway.
posted by A Thousand Baited Hooks at 2:36 PM on March 26, 2012


new pure algorithm => unpatentable because it fails the utility requirement
putting existing algorithm on computer => probably unpatentable because it likely fails the nonobviousness requirement
putting new pure algorithm onto computer => possibly patentable if it is also nonobvious, useful, sufficiently enabled, etc

I agree with Mars Saxman; if you can patent an algorithm merely by running it on a computer through a trivial and well-known transformation, then it's a meaningless and misleading distinction to say that pure algorithms cannot be patented.

Well it's the difference between something that's useful and something that isn't. The Arrhennius equation, by itself, is not useful. But a computer may be programmed to use it to control a rubber curing machine that produces a superior rubber product, and that invention may be patented. Does that seem like a meaningless and misleading distinction? Has the Arrhennius equation been patented? Or is it a useful application of that equation.

I mean, it's not enough to have a computer that simply calculates the Arrhennius equation for random values or something. The result must still be useful (e.g. controlling a rubber curing machine), but putting it on a computer opens the door to utility in a way that is just not possible with a bare algorithm or equation.
posted by jedicus at 2:41 PM on March 26, 2012


That's it. I don't care what you call the legal structure, this is what needs to happen. The current law does not allow this, so it is broken.

You're arguing for a defense of independent invention. There are a fair number of scholars who would support you on that, but I'm not sure the lack of an independent invention defense makes the law "broken." Consider that an independent invention defense invites willful ignorance, which would lead to a lot of negative side-effects, which I could go in to if you want. It's also quite a hard defense to prove, as the copyright cases demonstrate.

But perhaps you'll be heartened to know that the recently passed America Invents Act significantly expanded the prior user defense (scroll down past the quote from the bill to see the summary and analysis), which is a somewhat restricted version of independent invention.

But they'll never do this, because their business model is based on charging thousands of dollars for a piece of hardware that can be replaced by an iPad. They'd never be able to charge $8,000 (minus the actual cost of their hardware) for an iPad app.

If all of the iPad app's customers are people who can't afford or can't use their hardware, then those customers represent a pure increase in revenue. Why wouldn't they want a piece of that?

Hopefully the app maker will be able to get some kind of compulsory licence with a reasonable royalty. I'm not a patent attorney, so I'm not sure whether this ever actually happens (although years ago when I did a course on Australian IP law the lecturer said it was extremely rare)

After the eBay case it's become much more common in the US.
posted by jedicus at 2:55 PM on March 26, 2012


jedicus: I'm not so much looking for a defense as a removal of grounds to sue. For an independent developer, there's no difference between being sued and being shut down; the cost of defense is inherently prohibitive.

I'm not sure how you can say the law is not broken, unless you believe that it is reasonable for random people to extract licensing fees from other random people, without actually doing any work or providing any benefit in exchange.
posted by Mars Saxman at 3:01 PM on March 26, 2012 [2 favorites]


If all of the iPad app's customers are people who can't afford or can't use their hardware, then those customers represent a pure increase in revenue. Why wouldn't they want a piece of that?

Because then all of the people who would otherwise buy the expensive hardware would also buy the iPad app, which (according to the article in the FPP, anyway) is at least as good and in some cases better.
posted by A Thousand Baited Hooks at 3:01 PM on March 26, 2012 [1 favorite]


Yes, plus the fact that many people who would have bought the $x,000 hardware solution will buy the $x00 iPad solution instead. It's not at all hard to see why a company with a lot invested (literally!) in hardware production would prefer no iPad action at all, no matter how large a piece of it they were offered.
posted by No-sword at 3:12 PM on March 26, 2012


I donno if it's better in all cases, but she says the iPad weights far less, and the application isn't quite so complex, which matters for a young child learning to speak.
posted by jeffburdges at 3:12 PM on March 26, 2012


There was another article linked up thread which said an older child chose a different iPad application so she could feel more normal.
posted by jeffburdges at 3:14 PM on March 26, 2012


(Er, see = imagine, of course, in this case.)
posted by No-sword at 3:16 PM on March 26, 2012


Because then all of the people who would otherwise buy the expensive hardware would also buy the iPad app, which (according to the article in the FPP, anyway) is at least as good and in some cases better.

That's a fair point. I think the answer really depends on what exactly the patentee's profit margin is, given that they're a small maker of custom hardware. It's possible that an app could provide a similar profit margin to what they get on hardware.

In any case, there's a decent chance that the end result is a royalty arrangement rather than an injunction, whether the patentee likes it or not.

I'm not so much looking for a defense as a removal of grounds to sue. For an independent developer, there's no difference between being sued and being shut down; the cost of defense is inherently prohibitive.

If your defense consists of "here's the evidence that I wrote my program before you filed for your patent," then that's really quite cheap. I mean like a few hundred to a thousand dollars cheap.

Anyway, I suppose you could require that the patentee plead in the complaint that there was no prior use. In the open source case it would be relatively easy to determine that.

I'm not sure how you can say the law is not broken, unless you believe that it is reasonable for random people to extract licensing fees from other random people, without actually doing any work or providing any benefit in exchange.

I have argued that patentees should be required to "do work" and "provide a benefit in exchange" in the form of an enabling disclosure of a new, nonobvious, and useful invention. The patent system is premised on the idea that this trade off is ultimately better for society than the alternative. The fact that some people get the short end of the stick in this arrangement doesn't necessarily mean it isn't better overall.
posted by jedicus at 3:38 PM on March 26, 2012


I have argued that patentees should be required to "do work" and "provide a benefit in exchange" in the form of an enabling disclosure of a new, nonobvious, and useful invention.

Sure, that is reasonable. But I would counter-argue that a software invention which met that standard would be indistinguishable from a piece of working source code, in which case it would already be protected by copyright, and a further layer of monopoly would be unnecessary at best and harmful at worst. So why don't we just achieve the situation you and I both apparently want by abolishing software patents and letting people protect their work with copyright?
posted by Mars Saxman at 3:47 PM on March 26, 2012


If all of the iPad app's customers are people who can't afford or can't use their hardware, then those customers represent a pure increase in revenue. Why wouldn't they want a piece of that?

Because all the people buying $9000 devices will want the $300 app instead? It's probably be a huge decrease in revenue. It might even be a decrease in profit if they've sunk a lot of unrecoverable cost into developing custom hardware with capabilities that the iPad can duplicate fully. Also, depending on how the company is run they may have a guy who's got a lot of clout who's whole division would be vulnerable if they embraced an iPad play (like IBM's failure to get into RAID tech until EMC was eating their lunch).

Personally, I've got a $6000 insulin pump on my hip that could probably be duplicated for under $100 if you could leverage some kind of commodity off the shelf hardware like a smart phone to run most of it (Looking at Animas, OmniPod and Dexcom, that may be the trend). Other than the software algorithms, the hardware is very slightly more complicated than a flow control valve or an aquarium pump.
posted by BrotherCaine at 4:06 PM on March 26, 2012 [1 favorite]


I've never seen a more clear demonstration that our patent system is doing the opposite of what it was set up to do.
posted by Mental Wimp at 5:50 PM on March 26, 2012 [3 favorites]


Wow, Skeptic and Jedicus are defending software patents!? shocking!!!!
I've seen an awful lot of solutions proposed to the patent problems, but "just let the lawyers sort it out" doesn't sound particularly wise. -- jeffburdges
But if you're a patent lawyer It would certainly be a profitable solution.
jedicus: I'm not so much looking for a defense as a removal of grounds to sue. For an independent developer, there's no difference between being sued and being shut down; the cost of defense is inherently prohibitive. -- Mars Saxman
Sure, but if a developer can't afford patent lawyers, then that developer is of no use to patent lawyers, right? So why worry about them? Developers who can't afford patent litigation are completely irrelevant.
But there is a difference between an algorithm or an equation, as a purely abstract entity--Skeptic
That every algorithm can me encoded as a recursive equation is literally what the church Turing thesis says. Which is basically foundational theory of computer science.
I've read it. Prof. Knuth is a brilliant computer scientist, but he does not understand the law here jedicus
Just like you and skeptic don't understand computer science!!!

But hey, the same two people arguing the same exact point whenever software patents come up. It's not like they ever bother to come up with any new arguments. They simply reiterate over and over again that 1) Software patents are good (Steve jobs likes them!) and that they are totally legal.

The whole argument is that software patents should not Be legal and thus the law should be changed. To the extent that software is patentable, the law is broken. The fact that software patents may be legal now has nothing at all to do with whether or not they are a good idea.


---
I imagine a Kickstarter type campaign to develop an open source AAC app for cheap Android tablets would take off like wildfire, but it's a non-starter as long as SCS and PRC's lawyers are trying to sue the competition out of existence. -- jack_mo
Which they can't do in countries that don't have software patents. Which is most of them. EU patent laws specifically excludes "Programs for Computers" from the list of things which can be patented. Look at VLC, for example. The main reason why we had such shitty 'media players' in the 90s was because all of the codecs were patented, and thus no one could make a media player that would play anything.

But VLC was developed in France, where no one had any reason to worry about it.
Now, what if SCS and PRC decide (or are forced) to offer a license to Speak to Yourself? well, they may as well close shop then. They'll get a steady income from the license, why should they bother themselves with actually employing people, complying withe health and safety regs, etc? They'll hardly recover the sunken costs of the manufacturing machinery, of course, and their workers' jobs will effectively have been outsourced to Foxconn in China, but that's the cost of progress, isn't it? But next time there's yet another MeFi discussion on the loss of manufacturing jobs, just have a thought about that. -- Skeptic
What a ridiculous load of crap. First of all, why assume they aren't already contracting out manufacture to Foxconn? What makes you think they won't in the future?

Beyond that, it reminds me of the southpark "DEY TUK UR JUUURRRRBSS!!!". People use the same crap arguments for offshore drilling or anything bad that people currently work doing. Cracking down on organized crime costs organized crime jobs too. Does that mean we shouldn't do it?

Fuck this company charging thousands several thousand for shitty software, Sure, maybe some developers have to find a job somewhere else, which is not difficult, but so would their patent lawyers. Fair trade, IMO.
Suppose I went to github (a prominent open-source hub) with my high-priced lawyer and said, "This project impinges on my intellectual property." They're almost certainly going to shut that project down - because they rightfully don't want to get into such a war which would cost them a lot of money and profit them not at all. -- lupus_yonderboy
So, they move to a different host? If an OSS project requires github to work properly, they're doing something wrong. Github isn't even an open source thing itself. Kind of annoying that everyone is using it, frankly.
When Steve Jobs announced the original iPhone, he noted "and boy have we patented it." It's right there in the presentation as a bullet point, alongside "works like magic" and "no stylus."
Patent lawyer or similar. Apple invested a ton into patents and patent royalties. They were major players in the consortium that licenses h264 and other codecs. Apple didn't care about what was best for society, they were trying to make as much money as possible, and they certainly made quite a bit. Pretty ridiculous to be complaining about Apple and Foxconn and outsourcing and loss of manufacturing jobs on the one hand and the praising jobs as an exemplar of civic responsibility on the other.
Furthermore, there's no reason in principle one couldn't set up a computer capable of OCRing its input from paper and executing it; this would just be a very inefficient way of doing things. In principle, one could indeed write programs on paper using nothing but a graphite pencil, and then have a computer execute them directly.
If you used the pencil to punch holes in the paper, then that's actually how people programmed for years. (well, not pencils specifically, they had special card punching tools. But you get the idea)
posted by delmoi at 6:23 PM on March 26, 2012 [5 favorites]


But I would counter-argue that a software invention which met that standard would be indistinguishable from a piece of working source code, in which case it would already be protected by copyright, and a further layer of monopoly would be unnecessary at best and harmful at worst. So why don't we just achieve the situation you and I both apparently want by abolishing software patents and letting people protect their work with copyright?

If someone wants a patent they have to identify and disclose the invention. But source code need not be published in order for a program to be copyrighted, and neither do the inventive aspects of the program need to be identified or explained. So there are some differences there.

And again, copyright doesn't protect the functional aspects of software, so someone would still be free to copy the invention, just not the particular creative expression of it.
posted by jedicus at 7:01 PM on March 26, 2012


jedicus: the particular creative expression of the idea is the invention, or all of it that matters. Ideas, in software, are something very close to worthless. Even if you have a complete, working example of the idea implemented in code, that may not be enough to solve your problem; it may still be quite a bit of engineering work to adapt that idea to your specific situation. This is what software development is all about. Patenting the idea is like awarding the prize for the easy part of the job. This is not like mechanical engineering, where coming up with a design is the hard part, and cloning it is easy; in software, coming up with an abstract design is generally quite easy, and applying it to a specific situation is what we get paid to do.
posted by Mars Saxman at 7:04 PM on March 26, 2012


Mars Saxman For an independent developer, there's no difference between being sued and being shut down; the cost of defense is inherently prohibitive.

Would professional indemnity insurance cover help with this?
posted by aeschenkarnos at 7:35 PM on March 26, 2012


Sure, but if a developer can't afford patent lawyers, then that developer is of no use to patent lawyers, right? So why worry about them? Developers who can't afford patent litigation are completely irrelevant.

You're confusing the interests of patent lawyers with the interests of society. Obviously policy makers should take the latter into account rather than the former. And reducing transaction costs and the waste caused by litigation is quite important to me. That's why I favor, for example, investing more in the Patent Office, tightening up the enablement standard, raising Patent Office fees (especially for large entities), and more liberal use of fee shifting. All of these things are favorable to potential defendants and are, at best, neutral for patent lawyers. Some of them are downright bad for patent lawyers (e.g. raising PTO fees and tightening standards would lead to fewer applications filed; liberal use of fee shifting would tend to discourage litigation).

I also favor eliminating the current licensing system for lawyers and substantially decreasing the cost of law school, which would tend to bring legal costs down in general. On a personal level, I've done a fair amount of free work for startups and individuals. So don't assume that I'm some mercenary just looking out for my own paycheck.

That every algorithm can me encoded as a recursive equation is literally what the church Turing thesis says. Which is basically foundational theory of computer science.

I'm quite familiar with the Church-Turing thesis. It doesn't invalidate anything I've said. Neither an algorithm on a piece of paper nor a recursive equation on a piece of paper is useful or patentable, so their equivalence is immaterial.

Just like you and skeptic don't understand computer science!!!

I can't speak for Skeptic, but I have a bachelor's and a master's degree in computer science and have worked as a software developer.

It's not like they ever bother to come up with any new arguments. They simply reiterate over and over again that 1) Software patents are good (Steve jobs likes them!) and that they are totally legal.

You're wildly misrepresenting my point about Steve Jobs's statements. I was responding to a specific claim in a specific way and even went on to make it very clear that I was not arguing that software patents were good simply because Steve Jobs found them useful.

You're also misrepresenting my arguments. I argue primarily that software patents are not special, that any problems with them are best fixed in ways other than excluding them from patentability, and that many of the arguments against them fail to understand how the law works. My point is not "software patents are legal so deal with it." The law can certainly be changed, but it should be changed for coherent, rational, socially beneficial reasons.

The whole argument is that software patents should not Be legal and thus the law should be changed. To the extent that software is patentable, the law is broken. The fact that software patents may be legal now has nothing at all to do with whether or not they are a good idea.

In this thread much of the counter-argument has amounted to "software patents are patents on math; math shouldn't be patentable; ergo the software patents shouldn't be allowed." My point has been that software patents are not, in fact, patents on math. If the counter-argument is based on a misunderstanding of the law, then I think it's relevant to point that out.

Now, if the argument is "software patents are ultimately a net social harm and no amount of reform will fix that" then that's a fine argument, but I'd like to see the empirical evidence for it, not anecdotes.

Patent lawyer or similar

I wasn't aware that CEOs were similar to patent lawyers. You'll have to define exactly what you mean by "or similar," since I assumed it meant something like "paralegals, law professors, patent examiners, and other people who derive their livelihood directly from the existence of software patents." Steve Jobs did not fall into this category, I don't think, unless you want to argue that Apple's success was due primarily to the existence of software patents.
posted by jedicus at 7:35 PM on March 26, 2012


the particular creative expression of the idea is the invention, or all of it that matters

No, it isn't, not legally speaking. Creative expression is, by definition, the non-functional aspects of the code. It's stuff like formatting, comments, the choice of variable names, the division into different files, etc. (Frankly it doesn't make much sense in the context of compiled code.)

Ideas, in software, are something very close to worthless. Even if you have a complete, working example of the idea implemented in code, that may not be enough to solve your problem

It depends on what you mean by "idea." If you mean something vague like "compare two images for similarity" then sure that's easy to dream up. But inventing an algorithm for doing so efficiently and effectively is a whole different matter. New algorithms and data structures are discovered fairly regularly, which to me indicates that they aren't easy or worthless, since if they were easy they would have all been discovered years ago and if they were worthless nobody would bother.

Patenting the idea is like awarding the prize for the easy part of the job.

Well again, the idea can't be patented. And as I've suggested, tightening up the enablement requirement would effectively require the patent applicant to have also done "the hard part of the job."

in software, coming up with an abstract design is generally quite easy, and applying it to a specific situation is what we get paid to do.

That's true of programming but it's not really true of computer science. Coming up with a new efficient algorithm or data structure is fairly difficult. They don't give the Turing Award to people for applying designs to specific situations, after all.
posted by jedicus at 7:38 PM on March 26, 2012


I'm quite familiar with the Church-Turing thesis. It doesn't invalidate anything I've said.
I realize you and Skeptic always share the same position, but I would imagine you'd be able to tell each other apart, especially since I made sure to denote who I was quoting.
You're confusing the interests of patent lawyers with the interests of society.
Actually, that seems to be you guys.

You posted 26% of the comments in this thread. Who knows how many hundreds of words. I'm just assuming you're making the same arguments you always do.
All of these things are favorable to potential defendants and are, at best, neutral for patent lawyers.
Yes "neutral" to patent lawyers. Defendants who pay tens of thousands of dollars in legal fees to patent lawyers might actually have a chance of winning! And if they can't afford it they'll just have to stop selling or giving away their software. Which is no big deal, of course, if they can't afford patent lawyers then clearly they are of no economic value to patent lawyers, and thus are totally irrelevant.

I mean, sure, maybe this app developer thought he would do something good for children with special needs and make some pocket change in the process. Huge mistake, an app that can't pay for patent lawyers is an app that doesn't need to exist. If he thinks the obvious prior art should be enough he should do the sensible thing and jack the price up to $1000 per license - how dare a developer make money for themselves without letting a patent lawyer in on the action!
You're also misrepresenting my arguments. I argue primarily that software patents are not special, that any problems with them are best fixed in ways other than excluding them from patentability, and that many of the arguments against them fail to understand how the law works
Right, rather then simply eliminate the entire problem once and for all in a simple way, we need complex legal tweaks so that patent lawyers can still money off the actual creative work of other people.
posted by delmoi at 9:58 PM on March 26, 2012


delmoi First of all, stop the ad hominem shit. Please. Secondly, my background is in mechanical engineering. I have done a fair deal of programming, and while I am not a computer scientist, I very probably know more about computer science than you know about patent law. And I don't have much of a personal stake in getting companies such as Apple get patents, because in my work I deal with "old tech" stuff such as engines and gearboxes. However, hearing people misrepresent patent law raises my hackles, because I know that stuff.

Take for instance what you say about patenting software outside the US and in particular in Europe. I happen to be in Europe and knowing European patent law is my job, and I can assure you that you are wrong. Art. 52(2) of the European Patent Convention (not EU law, which doesn't enter into such matters), says that, among other things, like business schemes, computer programs shall not be considered as inventions for the purposes of patentability. However, right after that, Art. 52(3) specifies that this only applies only to the extent to which the patent applies to such excluded subject-matter "as such". People get blue in the face argueing about this, but the way courts across Europe have interpreted this is that, while computer programs are not as such patentable, a new and inventive device or technical process does not suddenly become unpatentable merely because it involves a computer program. The two patents mentioned upthread? They'd almost certainly be allowed here in Europe.
posted by Skeptic at 11:26 PM on March 26, 2012


The two patents mentioned upthread? They'd almost certainly be allowed here in Europe.

I must qualify this pre-coffee sentence: they almost certainly will not be rejected for being "software patents". As a matter of fact, there is a corresponding European patent application that is still pending. The patent examiner's current objections, however, relate to lack of novelty, obviousness ("lack of inventive step") and lack of clarity of the claims (the examiner particularly objects to the word "polysemous"), not to the non-patentability of software.

And no, I'm not in the slightest involved with the prosecution of this patent application.
posted by Skeptic at 2:54 AM on March 27, 2012


The argument for patents (and IP law as a whole) is that without protection, people will be unable to exploit their work and without the ability to exploit it, they won't do it. So we pretend innovation is a solid, touchable thing for which some modified concepts of property apply.

Which is an easy idea to grasp, and so obvious that most people don't question it in general. We all know that being able to own things and stop others from taking them is bad, because without that we'd have everything we wanted or needed taken away by whoever had the strongest gang.

It starts to fall apart because innovation isn't a solid, touchable thing. It's information, which is as different from things as energy is from matter. You can duplicate it at little or no cost, and no change in state of the information being duplicated. How and when the analogy with actual stuff falls apart is what keeps armies of lawyers from having to make solid, touchable things for a living, and one result of this mismatch between piety and practice leads to things we want or need taken away by whoever has the strongest gang. Oops.

It has yet to be shown that the easy, graspable idea of intellectual property is actually correct in many important ways. In areas where the 'obviously right' doesn't actually apply - the Industrial Revolution after Watt's patents expired, and the pre-Clinton software industry where patents were very uncommon - we get high levels of innovation. In areas where industries entirely dependant on IP law get to make or change that law - content distribution - we don't, and when we do it is fought with a rare degree of viciousness. In areas of inherently high innovation, where a new knowledge has widespread practical uses - the early days of wireless and TV - but patents still apply, you rapidly end up with their use as weapons to spike the guns of the competition.

Every so often, the world considers changing the details of IP law. It has lots of meetings, to which those who have lots of money and influence as a result of IP law are highly motivated and able to attend.

You can see where this is leading, right? Lessing did.
posted by Devonian at 6:25 AM on March 27, 2012 [2 favorites]


I realize you and Skeptic always share the same position, but I would imagine you'd be able to tell each other apart, especially since I made sure to denote who I was quoting.

Yes, you were addressing Skeptic with regard to the Church-Turing thesis specifically, but you then said that both Skeptic and I did not understand computer science, of which the Church-Turing thesis is a part. So it seemed fair to respond to both.

Yes "neutral" to patent lawyers. Defendants who pay tens of thousands of dollars in legal fees to patent lawyers might actually have a chance of winning! And if they can't afford it they'll just have to stop selling or giving away their software. Which is no big deal, of course, if they can't afford patent lawyers then clearly they are of no economic value to patent lawyers, and thus are totally irrelevant.

I said "at best neutral." Increasing PTO fees, tightening the enablement standard, and investing more in the Patent Office are all things that will reduce patent filings and litigation. That means less money for patent lawyers and less litigation. Which means a lot of potential defendants wouldn't get sued in the first place. I don't see how you get "you don't care about small developers" from that.

how dare a developer make money for themselves without letting a patent lawyer in on the action!

This is a complete non sequitor from my arguments. I've made it quite clear that none of my arguments are premised upon making work for patent attorneys. If there were a way to have a functional patent system without any money going to attorneys that would be fantastic. Attorneys do produce value, but the overhead and transaction costs are often very high. A patent system that didn't need them would probably be better for society than the one we have.

Right, rather then simply eliminate the entire problem once and for all in a simple way, we need complex legal tweaks so that patent lawyers can still money off the actual creative work of other people.

Okay, please set out this simple solution that will eliminate the entire problem once and for all. You can start by defining "software patent" in a way that is neither over-inclusive nor under-inclusive, unambiguous, and cannot easily be gamed. I'll give you some example technologies to suggest why this might not be so easy.

1. A machine that cures rubber by heating and cooling it, controlled by hand.
2. A machine that cures rubber by heating and cooling it, controlled by a computer program using a new, nonobvious, and useful application of a mathematical equation.
3. A new kind of rubber curing machine that uses such a complex curing process that only a computer could control it, resulting in a significantly superior product.
4. A computer program for modeling the rubber curing process using new, nonobvious, and useful applications of mathematical equations.

Which of these are "software patents" and why? 1 seems like it obviously isn't. 4 seems like it obviously is. But what about 2 and 3? Does the mere involvement of a computer make it a software patent? What if the computer is doing something that a human couldn't replicate (i.e. something that requires a machine)? But isn't a computer a kind of machine? So why should 4 be unpatentable again, exactly, since it involves a machine? Why do we want to remove an incentive to develop new kinds of automation and computerization anyway?

So, since the solution is simple and easy, please provide it. Oh, and what do you plan to do about all of the existing software patents and pending applications? Are you going to revoke them retroactively? Are you prepared to pay out billions of dollars in the ensuing takings lawsuit?
posted by jedicus at 7:24 AM on March 27, 2012


the pre-Clinton software industry where patents were very uncommon - we get high levels of innovation

I'm not sure what Clinton has to do with this as such, so I assume you mention him only to indicate a timeframe. In any case: that was also the time during which Microsoft built a monopoly partially on the basis of copying the innovations of others, something which software patents may have prevented. And I'm surprised to see you call that a period of particularly high levels of innovation. After all, that was before the widespread adoption of the internet and mobile devices, before Linux and open source really took off, before wide availability of 3D graphics, just to name some things off the top of my head. What exactly is the evidence that there was a particularly high level of innovation in software before 1992?

In areas where industries entirely dependent on IP law get to make or change that law - content distribution - we don't, and when we do it is fought with a rare degree of viciousness

Don't confuse copyright with patents. They are both IP but the politics involved are very different. Copyright law is the product of decades of meddling by special interests carving out special treatment for their particular kind of content or delivery system. The copyright law is several times the length of the patent law for a reason. I would absolutely support a massive reform of the copyright law: much shorter terms (or at least significant maintenance fees beyond a short initial term), a process by which orphan works can be moved into the public domain, the end of statutory damages and criminal liability, a streamlined mechanical licensing system designed for the internet and digital media.

Every so often, the world considers changing the details of IP law. It has lots of meetings, to which those who have lots of money and influence as a result of IP law are highly motivated and able to attend. You can see where this is leading, right? Lessig did.

That's an argument for repealing just about every law on the books, since they're virtually all tainted by lobbying and campaign donations. The solution (as Lessig rightly identified it) is fixing the problem of money in politics, not deciding to throw the baby out with the bathwater simply because monied interests are involved.
posted by jedicus at 7:41 AM on March 27, 2012 [1 favorite]


You can start by defining "software patent" in a way that is neither over-inclusive nor under-inclusive, unambiguous, and cannot easily be gamed.

Just because there is no easy solution to a problem doesn't mean there isn't a real problem, or that even imperfect steps should not be taken towards solving it if they're better than what currently exists.

But I'm going to take you at your word and attempt to start a discussion on how the system could be fixed. To this end I'm going to draw from my deep wellspring of ignorance and bluesky a patent process, in the hopes that by exposing some possible solution to attack we could understand the problem better, and maybe even work towards a solution. Here goes.

Software is a set of instructions intended to be run by some computing entity, whether that entity is human, mechanical or electronic.

1. A machine that cures rubber by heating and cooling it, controlled by hand.
2. A machine that cures rubber by heating and cooling it, controlled by a computer program using a new, nonobvious, and useful application of a mathematical equation.
3. A new kind of rubber curing machine that uses such a complex curing process that only a computer could control it, resulting in a significantly superior product.
4. A computer program for modeling the rubber curing process using new, nonobvious, and useful applications of mathematical equations.


Here is my suggestion:

1: The machine is patentable, as an invention.
2: The machine is patentable, neither the computer program, nor the equation, nor the application of the equation are patentable.
3: The machine is patentable, the software is not. In this case the software need not be patentable though -- it is useless without the machine.
4: Not patentable. There is no machine.

That didn't seem difficult at all. I think in true edge cases where something might be software and thus not able to be patented, the law should err on the side of it being software. But patents in general are due for revising, and patent terms are probably too long at the moment.

What if the computer is doing something that a human couldn't replicate (i.e. something that requires a machine)? But isn't a computer a kind of machine?

We believe that the applications of that machine shouldn't be considered such in the sense of patent law. That's the whole point of the software patent ruckus. If it doesn't have a physical aspect it should not be patentable.

So, since the solution is simple and easy, please provide it.

I did, respond to that.

Oh, and what do you plan to do about all of the existing software patents and pending applications? Are you going to revoke them retroactively?

Follow your argument to its conclusion and no law that, were it passed or overturned, would cost anyone any money, should be allowed to be changed. Appeal to inertia.

Are you prepared to pay out billions of dollars in the ensuing takings lawsuit?

The "earnings" would be ruled invalid by the law, as being something improperly acquired. Nothing is taken; things that are worth money cease to exist. But they shouldn't have existed in the first place. I am no lawyer, but it seems like your argument is to simply throw your hands up and say "What U gunna do?"

In any case: that was also the time during which Microsoft built a monopoly partially on the basis of copying the innovations of others, something which software patents may have prevented.

Yes. And it should be interesting to note that most open source people, when faced with the choice between 1. a Microsoft towering over the landscape, or 2. software patents preventing that as well as whatever other damage to the software industry that entails, firmly side with 1, despite being very aware of the consequences of that choice. That is a fearsome opinion, and worthy of consideration.
posted by JHarris at 8:14 AM on March 27, 2012


(I missed my chance to link this to "what U gunna do," above.)
posted by JHarris at 8:18 AM on March 27, 2012


What exactly is the evidence that there was a particularly high level of innovation in software before 1992?

I must also add: what exactly is the evidence that the levels of patenting post-Watt went down? The opposite seems to be true: the Industrial Revolution was very closely linked to the development of the patent system, and many of the industrial conglomerates born back then (GE, Siemens, Westinghouse, AT&T, Bayer, Solvay, Du Pont, etc.) can trace their origins directly back to specific patents.

It has yet to be shown that the easy, graspable idea of intellectual property is actually correct in many important ways.

I actually have some firsthand, empirical knowledge of the consequences of an anaemic patent system. I am Spanish, and even though I've been living abroad since high school, I still have strong family ties in the old country. I've studied and worked abroad, mostly in Northern Europe.

Now, Spain is a case study of a country with rather little respect for IP in general, and above all a very weak patent system to this date. In particular, industry, even large industrial corporations, generally don't know what patents are or how to use them. Not only does Spain file very few patent applications relative to its size and GDP, Spanish companies hardly ever bother to challenge their competitors' patents.

Coincidentally, Spanish private sector investment in R&D is abysmal. That this happens despite the fact that Spain actually haves a large pool of technically- and scientifically-qualified people, and actually fares comparatively well in other indicators such as scientific publications per capita, strongly suggests that this is because the private sector simply does not know how to get a ROI from R&D. They just don't bother, because following other people's ideas is much cheaper than paying engineers and scientists to come up with their own solutions.

The outcome? It isn't just patent professionals who don't get jobs: it's also all those gifted people who could otherwise be making "useful stuff". There's currently a terrifying brain drain. Young engineers and scientists are fleeing in droves, because their skills and work are much more highly valued in other countries, which coincidentally have much stronger patent systems, than in Spain. This is why I get so passionate about this: it isn't my job which I'm worried about, I'm actually pretty safe. It's those of half my younger relatives.
posted by Skeptic at 8:49 AM on March 27, 2012



1. A machine that cures rubber by heating and cooling it, controlled by hand.
2. A machine that cures rubber by heating and cooling it, controlled by a computer program using a new, nonobvious, and useful application of a mathematical equation.
3. A new kind of rubber curing machine that uses such a complex curing process that only a computer could control it, resulting in a significantly superior product.
4. A computer program for modeling the rubber curing process using new, nonobvious, and useful applications of mathematical equations.

Here is my suggestion:

1: The machine is patentable, as an invention.
2: The machine is patentable, neither the computer program, nor the equation, nor the application of the equation are patentable.
3: The machine is patentable, the software is not. In this case the software need not be patentable though -- it is useless without the machine.
4: Not patentable. There is no machine.

That didn't seem difficult at all.


Except that you got at least #2 wrong. The machine in #2 isn't necessarily new, or non-obvious. It may perfectly be an existing machine, in which case it certainly shouldn't be patentable. What is new and non-obvious in that case is the process that the machine is carrying out.

Now consider this far-from-entirely-hypothetical situation:

Both rubber company A and rubber company B use the same rubber curing machine, provided by machinery company C. A clever boffin in rubber company A finds develops an improved process using a new, nonobvious, and useful application of a mathematical equation. Rubber company A files a patent application for that process and asks machinery company C to write the appropriate software to implement that process in the machine. Then machinery company C goes and sells the program to rubber company B, without A's knowledge.

Who does A sue for patent infringement? Obviously, company B, if A can prove that B is performing the same process in a territory covered by their patent. What about C? A possibly can't sue C for copyright infringement, because, unless there was some specific prior arrangement between them, C, which wrote the code, may actually hold the copyright. Unless there was some tacit or explicit confidentiality agreement, they can't sue them for breach of confidence or breach of contract either. And company C is certainly contributing to and benefitting from the patent infringement. Indeed, this situation is what is known as "contributory infringement".

Also, going back to jedicus' examples, #3 isn't that clear-cut either. The machine itself, though new, may be obvious, and the actual invention may well lie in the complex curing process. We go back to the same situation of #2, with the added spice that, to fulfill the enablement requirement, how the computer actually performs the control of the machine should be disclosed in quite significant detail.

As for #4, that is the kind of patent which would be very difficult to get granted, never mind enforce, in Europe these days, and I think (but I'm not so knowledgeable in US law), quite a struggle in the US as well.
posted by Skeptic at 9:44 AM on March 27, 2012


Jedicus, you seem to believe that software patents ought to be useful, and you appear to be defending some idealized idea of what software patents could be, but the fact is that software patents are a depressing mess which benefits nobody but patent lawyers and patent trolls.

Software patents have never done software developers or society at large any good. Their only function is to allow big companies to reduce competition by squeezing their rivals through patent litigation and licensing fees. They only restrict development of new technology by making it legally risky. This is all so well known in the software development community that it hardly needs to be said - I'm amazed to find someone who claims to have industry experience who has trouble recognizing it.

That's true of programming but it's not really true of computer science. Coming up with a new efficient algorithm or data structure is fairly difficult. They don't give the Turing Award to people for applying designs to specific situations, after all.

Yes, of course, but that is what universities and scientific publications are for. Patenting your new idea does nothing but guarantee that every practicioner will do their utmost to avoid knowing anything about it.

Do you realize that employees at Microsoft and Google, and presumably other large companies, are strictly forbidden from reading patents? Any patents, at all? There are separate legal departments for dealing with patents, and they are not allowed to talk directly to engineers, for fear of accidentally contaminating those engineers with knowledge about the patents. Microsoft even kept a list of patents which an employee might have been contaminated with so that they could ensure such an employee was never accidentally put to work on a related project where they might accidentally apply anything they had ever read from a patent.

If you are a researcher who is coming up with a new efficient algorithm, patenting it consigns your idea to a long, cold death. Patenting your work guarantees that no practicioner will ever take advantage of it, or even become aware of it if they can avoid doing so. In fact, people who do know about your new algorithm will go to great lengths to reject your work and find some unrelated way of accomplishing their goals! The more useful your patent is, the more likely it is that they will be able to do so. It is only the vague, fuzzy, useless patents which cannot be escaped this way.

This is the last thing I'm going to say on this topic: software patents do nothing to promote research or the development of new ideas. In fact they do the opposite: they impede the flow of knowledge, they waste developer time re-inventing wheels, and they divert capital from productive work to unproductive litigation. Whatever you think about the idealized software patents in your head, the actual state of software patents in the USA is all bad for everyone but lawyers and they ought to be repealed. They are a complete failure.
posted by Mars Saxman at 10:29 AM on March 27, 2012 [1 favorite]


Skeptic: Except that you got at least #2 wrong. The machine in #2 isn't necessarily new, or non-obvious. It may perfectly be an existing machine, in which case it certainly shouldn't be patentable. What is new and non-obvious in that case is the process that the machine is carrying out.

That situation, using a pre-existing machine in a novel way, should not be patentable. You are casting your will over a device that someone else invented, or may be public domain.

Your following example is hard to follow with the budget of brainpower I have allotted to Metafilter today, but I'll try to engage with it. Why should what A invented regarding an application for a mathematical equation be patentable at all?

Your assertion regarding Spain might be suggestive but is far from conclusive. It might be a complex issue with many causes. And anyway, investment in tech is not success in tech, and may have nothing to do with the patent system. Has anyone written on the topic?
posted by JHarris at 10:34 AM on March 27, 2012


That situation, using a pre-existing machine in a novel way, should not be patentable. You are casting your will over a device that someone else invented, or may be public domain.

All devices are arrangements of previously existing elements. Invention is casting your will (and intelligence, and imagination) on things that already existed. There's little difference in whether your invention is a device or a process. Most inventions in chemistry, for instance, are processes, not devices.

Your following example is hard to follow with the budget of brainpower I have allotted to Metafilter today, but I'll try to engage with it. Why should what A invented regarding an application for a mathematical equation be patentable at all?

Pretty much all of technology, devices as much as processes, are applications of mathematical equations. You'll be hard pressed to do any engineering, physics, chemistry or even biology without mathematical equations.
posted by Skeptic at 2:20 PM on March 27, 2012


All devices are arrangements of previously existing elements. Invention is casting your will (and intelligence, and imagination) on things that already existed. There's little difference in whether your invention is a device or a process. Most inventions in chemistry, for instance, are processes, not devices.

I know all these things, have argued about them many times in the past. I still say that an application of a machine should not be patentable, and computer software is basically a pre-packaged application.

On the "Pretty much all technology" paragraph: patents are a necessary evil, granted existance by the U.S. Constitution only to promote commerce and industry. Uses of math equations should not be restricted. This is why I argue, when we do restrict them, it should be only in terms of the creation of new devices, and not the uses of pre-existing ones, or software that runs on them.

It is true that, fundamentally, it's only a matter of degree whether we're talking about the invention of a machine or the writing of software. But patents themselves are a makeshift solution to make up for a failing of the marketplace. The line has to be drawn somewhere; I say it makes much more sense for it to be on that side of software patents than this one. Wherever the line is drawn, if it's not done carefully then patents risk being put to adverse use, like, oh, say eliminating legitimate competition or laying legal landmines for people and companies.
posted by JHarris at 10:10 PM on March 27, 2012


You can start by defining "software patent" in a way that is neither over-inclusive nor under-inclusive, unambiguous, and cannot easily be gamed.

A patent over the following processes is a "software patent":
1. Any process that could be performed by a Turing-complete computer.
2. Also, any process that a Turing-complete computer could perform when connected (in a way not covered by the patent) to one or more devices (that are not covered by the patent).

That should just about cover it. There's room to quibble about the details (after all, that's the special power of lawyers) but this would be a solid basis for distinguishing software patents from non-software patents.

With this rule, the dynamic keyboard patents would be invalid because you could implement them completely with a standard computer connected to a touchscreen display in a standard way.

The ridiculous yahoo patents from the other day would obviously be invalid.

The rubber-curing patents other than 4 could be valid, assuming that they satisfied the general requirements of inventiveness etc and as long as they didn't purport to cover a software process connected to devices not covered by the patent in ways not covered by the patent.

You wouldn't even need to use this rule to make patents invalid altogether; the rule could just be that a process covered by the rule is not infringing.

How about that?
posted by A Thousand Baited Hooks at 2:37 AM on March 28, 2012


So, since the solution is simple and easy, please provide it. Oh, and what do you plan to do about all of the existing software patents and pending applications? Are you going to revoke them retroactively? Are you prepared to pay out billions of dollars in the ensuing takings lawsuit?

Oh, and this transitional issue shouldn't be a barrier to reform. You do whatever is done in your jurisdiction every time there's a change to patent law.
posted by A Thousand Baited Hooks at 2:45 AM on March 28, 2012


@JHarris

computer software is basically a pre-packaged application

So, not just "pure maths", then? Thanks.

The line has to be drawn somewhere; I say it makes much more sense for it to be on that side of software patents than this one.

Why? Just because it suits a particular interest group? You are not asking for a line to be drawn: you are asking it to be redrawn away from where it has been for the last couple of centuries.

Wherever the line is drawn, if it's not done carefully then patents risk being put to adverse use, like, oh, say eliminating legitimate competition or laying legal landmines for people and companies.

Or discouraging investment in innovation in entire fields of technology. I am all for drawing the line carefully, and where to draw the line is a discussion that has kept agile minds busy for a while. Quick, arbitrary solutions, like your proposal, are more likely than not to be counter-productive.

@A Thousand Baited Hooks

A patent over the following processes is a "software patent":
1. Any process that could be performed by a Turing-complete computer.
2. Also, any process that a Turing-complete computer could perform when connected (in a way not covered by the patent) to one or more devices (that are not covered by the patent).


That's "any process", then. Period.

I wonder what you mean "that is not covered by the patent". Going back to those dynamic keyboard patents, the patent does explicitly cover implementing them "completely with a standard computer connected to a touchscreen display in a standard way". What you mean apparently, is that, unless the way in which the devices to which the computer is connected and the way in which the computer is connected to them are new and non-obvious themselves, then the patent should not be valid. What you have then, is not a patent on a process, it is a patent on an apparatus comprising a computer and some new and non-obvious devices connected in a new and non-obvious manner.

At least JHarris is able to recognise that he wants all patents on processes and methods banned. And that, I repeat, would have implications way beyond software.
posted by Skeptic at 3:06 AM on March 28, 2012


Going back to those dynamic keyboard patents, the patent does explicitly cover implementing them "completely with a standard computer connected to a touchscreen display in a standard way".

Yes. That's the problem.

What you mean apparently, is that, unless the way in which the devices to which the computer is connected and the way in which the computer is connected to them are new and non-obvious themselves, then the patent should not be valid.

I don't think that's what I said, although I'm having a bit of trouble working out exactly what this sentence means. Are there a couple of words missing after "unless the way in which the devices to which the computer is connected"?

In my scheme you could get a patent on a process involving an existing device connected in an existing way to a computer if the process - not including any aspects of the process performed by the computer - was sufficiently inventive.

The dynamic keyboard process wouldn't qualify, because if you take out the aspects of the process that are carried out by the computer all you're left with in terms of innovation is a touchscreen connected to a computer to be used for assisted communication. And that could well have been a patentable invention at some point in the past, before anyone thought of using a touchscreen connected to a computer for assisted communication (including at a point after both of these technologies had been invented, but before anyone used them together for this purpose). But not when these patents were filed.

If you invented a special type of touchscreen device that, together with a specified program, was useful for assisted communication, you could patent that. But you wouldn't be able to prevent others from using similar (non-copyright-infringing, of course) programs with other types of touchscreen device to achieve the same result.

The toast refreshing machine from the other thread would qualify, because the inventive parts of the process (using an element at a particular temperature, emitting a particular wavelength of light) cannot be performed by a computer.
posted by A Thousand Baited Hooks at 3:50 AM on March 28, 2012


(quoting myself:) In my scheme you could get a patent on a process involving an existing device connected in an existing way to a computer if the process - not including any aspects of the process performed by the computer - was sufficiently inventive.

I should add (to explain my "special type of touchscreen" example): you would also be able to get a patent on a process partly carried out by computer and partly carried out otherwise if the process as a whole was sufficiently inventive.

However, this patent would not cover, and could not be used to suppress the use of, a method of achieving the same result by using the part of the process carried out by computer but not the part carried out otherwise.
posted by A Thousand Baited Hooks at 4:33 AM on March 28, 2012


ATBH: I don't think that's what I said, although I'm having a bit of trouble working out exactly what this sentence means. Are there a couple of words missing after "unless the way in which the devices to which the computer is connected"?

Sorry, I borked that sentence: I meant that according to you, "unless the way in which the devices to which the computer is connected and the way in which the computer is connected to them are new and non-obvious themselves, then the patent should not be valid".

ATBH: And that could well have been a patentable invention at some point in the past, before anyone thought of using a touchscreen connected to a computer for assisted communication (including at a point after both of these technologies had been invented, but before anyone used them together for this purpose). But not when these patents were filed.

Well, that's the hitch. Apparently, according to the USPTO at the time, this particular form of assisted communication, whether using a touchscreen or a dedicated keyboard, was new and non-obvious at the time. Not everybody seems to agree (the examiner at the European Patent Office, for instance), but there the argument is purely about novelty and obviousness.

ATBH: The toast refreshing machine from the other thread would qualify, because the inventive parts of the process (using an element at a particular temperature, emitting a particular wavelength of light) cannot be performed by a computer.

And receiving a physical finger input and issuing a visible output cannot be performed by a computer either. We are still at the same point.

However, this argument is useful because it very closely reflects the current legal debate in Europe. The current status is that a "computer-implemented invention" may be patentable if it has a technical effect "going beyond the normal interaction between a computer and software", but then there is the question of whether the steps carried out by the computer itself ought to be taken into account to decide whether the subject-matter is new and non-obvious compared with the prior art. There's one school of thought (the "contributory approach") which considers that the non-technical, abstract steps carried out by the computer should not be taken into account, but there was an important decision of the Boards of Appeal of the EPO ("Comvik") which consider that the invention should always be considered as a whole and that even non-technical features and steps may contribute to the ultimate technical effect of the invention.

Things get even more complicated in the US. Basically, US law also considers that abstract entities are not useful and thus not patentable. One legal test which has been used to discriminate between "useful" and "abstract" invention is the so-called Machine-or-transformation test, which holds that "a process qualifies to be considered for patenting if it (1) is implemented with a particular machine, that is, one specifically devised and adapted to carry out the process in a way that is not concededly conventional and is not trivial; or else (2) transforms an article from one thing or state to another."

However, as the rather lengthy Wikipedia article shows, this has not solved the dilemma, not least because the US Supreme Court has consistently muddied the waters in successive decisions pulling one way or another.

In short: there is no "lawyer conspiracy" on software patents. Drawing a line between patentable and non-patentable subject-matter simply is difficult because any "quick and easy" test to separate them sooner or later collides against the complexities of real life cases.
posted by Skeptic at 5:26 AM on March 28, 2012


Sorry, the "Comvik" link should lead here.
posted by Skeptic at 6:04 AM on March 28, 2012


I'm getting sleepy and my 9.33 post was wrong, sorry - scratch that.

Well, that's the hitch. Apparently, according to the USPTO at the time, this particular form of assisted communication, whether using a touchscreen or a dedicated keyboard, was new and non-obvious at the time. Not everybody seems to agree (the examiner at the European Patent Office, for instance), but there the argument is purely about novelty and obviousness.

But the particularity of this form of assisted communication is entirely in the software implementation. Take away the steps performed by computer and you don't have anything other than "touchscreen" (or dedicated keyboard) and "assisted communication" - unless the basic idea of using a touchscreen for assisted communication was novel in 1995, which I find hard to believe.

And receiving a physical finger input and issuing a visible output cannot be performed by a computer either. We are still at the same point.

But neither of these things was innovative in 1995. They were only innovative when combined with the particular process that these patents implement through software.

The current status is that a "computer-implemented invention" may be patentable if it has a technical effect "going beyond the normal interaction between a computer and software", but then there is the question of whether the steps carried out by the computer itself ought to be taken into account to decide whether the subject-matter is new and non-obvious compared with the prior art.

That is interesting, thank you.

Drawing a line between patentable and non-patentable subject-matter simply is difficult because any "quick and easy" test to separate them sooner or later collides against the complexities of real life cases.

That's certainly true, but this is an issue for any area of law - especially areas where appeal judges have to decide questions of law that are ultimately based on technical matters beyond their expertise. Any kind of reform has to start somewhere, although sadly my Turing-machine based idea probably doesn't have much chance.
posted by A Thousand Baited Hooks at 6:25 AM on March 28, 2012


That's certainly true, but this is an issue for any area of law - especially areas where appeal judges have to decide questions of law that are ultimately based on technical matters beyond their expertise.

Well, the Boards of Appeal of the European Patent Office contain a mixture of technically- and legally-qualified members, yet are the "bête noire" of anti-software-patent activists in Europe for their abovementioned "IBM" and "Comvik" decisions. Indeed, the fact that they include non-lawyer members is often thrown at them by the anti-software-patent activists.

In general, specialised patent judges with technical backgrounds (like those of the US CAFC) are perceived as more "patent-friendly" than "generalist" judges.

So, it isn't simple.
posted by Skeptic at 7:17 AM on March 28, 2012


computer software is basically a pre-packaged application
So, not just "pure maths", then? Thanks.


I didn't say anything about pure maths, dink.

>The line has to be drawn somewhere; I say it makes much more sense for it to be on that side of software patents than this one.

Why? Just because it suits a particular interest group? You are not asking for a line to be drawn: you are asking it to be redrawn away from where it has been for the last couple of centuries.


Why? WE'VE TOLD YOU WHY REPEATEDLY. The "particular interest group" we're talking about tends to be highly knowledgeable. Just because a particular person says something that might benefit them, doesn't mean they're wrong.

Two centuries ago we didn't have computer software, or computers! Now you're strongly appealing to inertia, and you are causing me to lose sympathy for your perspective.

Or discouraging investment in innovation in entire fields of technology.

Yeah, sure, technology will grind to a halt without bloody patents.

It's all a matter of degree -- technology certainly won't end without patents, although it may be severely depressed. But the uses of patents that I see in the world now strongly recommends to me that they need to be fixed, because they're being used as a bludgeon across the technology landscape, and that's a major barrier to technological advancement. At the very least patent trolls need to be reined in.

I am all for drawing the line carefully, and where to draw the line is a discussion that has kept agile minds busy for a while. Quick, arbitrary solutions, like your proposal, are more likely than not to be counter-productive.

Yeah and I hear this a lot, the need for well-considered, measured responses. The older I get the less patience I have for this -- it is used by a lot of people for saying that things should just stay the same sucky way they are for decades, until some bigger problem comes along and distracts everyone or the whole issue is rendered moot by something else. The patent line should be drawn carefully, yes, but it definitely needs to be redrawn, and soon, and that's regardless of which company is helped or hurt by it. If you need a reason why, well, read the damn FPP. I got stuff to do.
posted by JHarris at 8:59 AM on March 28, 2012


So, if I'm understanding all this back and forth correctly, in this modern, digital age, a patentable invention is anything I can get my lawyer to convince the patent office to patent. Seems about right. The definitions are non operationalizable, because they don't have any criteria in them. Rather, they use empty phrases like "concededly conventional" and "non-trivial". They virtually invite the sort of baseless arguing going on in this thread and, I imagine in patent office filings and court cases. I'm starting to thing the whole thing is broken and we need to start over again, with different approach.
posted by Mental Wimp at 9:52 AM on March 28, 2012 [1 favorite]


It's great news for anyone highly invested in pedantry. Less so anyone else.
posted by Artw at 9:53 AM on March 28, 2012


Yeah, the ability to take out a patent and then sit on it is fundamentally broken; the only purpose it serves is to allow companies to stifle innovation for the sake of profiting from inefficient systems already implemented.
posted by kaibutsu at 10:16 AM on March 28, 2012


delmoi First of all, stop the ad hominem shit.
Sorry, you two post the same stuff in every thread about software patents, and in this case the two of you posted around 35% of all comments in the thread entirely at that point. And even more now. It's ridiculous.

All you're doing is getting into a pointless argument about the semantic meaning of 'math' that has no relevance to anything practical at all. It's nothing more then an attempt to stop focusing on the practical problems caused by software patents in order to make irrelevant philosophical points that are totally irrelevant to anything at all.
posted by delmoi at 11:17 AM on March 28, 2012 [1 favorite]


« Older Shoutmennonitenames.com generates Mennonite names ...  |  The New York Times isn't known... Newer »


This thread has been archived and is closed to new comments