The US Patent and Trademark Office (USPTO) has asked the software community to help advise it on how to properly handle software-related patents – a move that could represent the first steps toward software patent reform. The agency published its request for comments in the Federal Register on Thursday, explaining that it wants …
The primary problem is funding. The USPTO does only a tiny fraction of searching for prior art when reviewing software patents. And god forbid they trust the applicants.
Several Apple patents are clearly shown to be prior art in videos from the 1970s Architecture Machine Group under Nicholas Negroponte, yet the USPTO appears to have ignored these. For example: page flipping as part of my Master's Thesis; The Spatial Data Management System (SDMS) 1978
Many other examples of current software patents can be seen to have appeared in the 1970s at MIT and other universities and research groups.
The USPTO gets pressed over and over by rich companies to award patents, and they "let the courts decide" with the results we see today.
The system is broken.
They are so fed up with all the patent shenanigans that they want to find a way out.
More funding does not solve the fundamental uncertainty of what can or cannot be patented. It also does not solve the problem created by an exponentially growing body of software patents. It is infinitely easier to write some new code that to create a new pharmaceutical compound and prove that it cures an illness, therefore software patents can never work as pharmaceutical patents do (the latter being the poster boy for a working patent system). As a result, the work to create potentially patentable software may be less than the work to check for prior art or related programs -- and the USPTO might have to hire more and better software people than those in industry. Such a system simply cannot work.
Time to give up on software patents. Software companies need to compete on price and quality, not by protecting 'ideas'.
They can very easily. Invalidate ALL software patents and be done with it.
Re: Underfunded USPTO - this is EASY.
I get to use everything everyone else makes, for free, within 5 minutes of it being patented.....
And everything that I patent, costs everyone else a million dollars a day in royalties to use, for ever.
Once we figure out that all patents in existence, are actually mine, because I lodged them.... under all my pseudonyms, everything will be all right.
Being that all the huge corporations love playing Cloak & Dagger with their patents, nothing will change. Let's don't forget who put software patents in place to begin with.
It's a nice gesture though, pretending that the USPTO cares and all.
Don't you know that the tragedy of the commons applies not only to all digital bits but all ideas as well? Every thing must be owned by one single person or entity or the world will end. Ironic how most of the people that believe this don't care about this world anyway and are only looking forward to the next one.
If we all roll over and play dead that will surely be the case.
But be a man and make a stand if you feel strongly about this.
If you're in the USA, respond to the request for input and send copies of your response to your senators, rep and the white house.
Tragedy of the commons? How do you use a JAVA instruction so many times the supply of that JAVA instruction runs out?
I share your point of view that the idea of the tragedy of the commons is taken too far, but it rarely ever begins to apply in the cyber world.
Sarcasm on my point is pretty lame and far beyond novel at this point in these forums I agree but my ridiculous post was that to some ideological right wing idiots even the cyber world must be owned and we all made to pay a tax to some wealthly ahole to use. Obviously some major open sources initiatives have proven that when all of humanity owns the software still some pretty awesome software can be written and that the tragedy of the commons notion is just something right wing tools use to control the conversation in all realms.
Tragedy of the commons
You keep using this term, I do not think it means what you think it means. I don't believe you are referring to either Garrett Hardin nor Aristotle as right wing idiots. So I can only assume by ideological right wing idiot you mean the wealthy, homophobic, gas-guzzling suv driving, control what you do authoritarian and not the wealthy, dope smoking, hybrid driving, control what you have authoritarian that is presumably the left wing. Either way if you follow the money going into campaign contributions I think you'll find ideological idiots contributing to both sides of the authoritarian divide.
I fully admit that I am unaware of any recent conversations where any one wing tool has used the tragedy of the so called commons in any realm so I may just be speaking from ignorance but I'd welcome something more concrete and not just an assertion. My apology to others for tangent but I'm genuinely flummoxed by the comparison. Perhaps the meaning is that like the "commons" software can't be owned and may require an upfront fee for use and a given amount of maintenance so it's a bit like a road use tax that gets paid on car registrations and petrol.
Re: Tragedy of the commons
Without getting any farther off on a tangent basically just saying that the belief that one entity has to own something or else it will degrade is horse crap (as shown clearly in software by the open source movement). As well the belief that it is necessary to monopolise ownership of an idea in order for the possible profit motive to motivate implementation is also patently (ha pun) false in many cases (software patent trolls often being nothing but law firms as opposed to development shops is a perfect counter example). The whole right wing thing was because in my homeland of the USA there are an awful lot of people that honestly believe this, including some in some positions of power and indeed the whole culture has been corrupted by the greed is a virtue to be cultured lie (personally could accept a necessary evil in some cases but not a virtue). Humanity doesn't necessarily prosper the most by monetizing everything as much as possible. Indeed when you set up a virtual land rush house of cards you get yourself in exactly the mess the USTPO is in where you start discouraging everyone but lawyers from doing any real work (lawyers real work haha).
Commendations to cousins overseas?
It is a tricky area - in the UK law tends to follow precedence but this verges in to new territory both in the nation and international spheres.
So, judgements need to consider present licensed users (an international factor impacting on national factor and vice-versa?)
And I rapidly conclude: incoming imaginary headache = terminate
I've yet to see a truely new idea in a software patent
I think they all hide out with the Unicorns.
The problem is that the USPTO only considers older patents as "prior art", meaning all "prior art" must be patented (surely it was?). Until they get over this hurdle and recognize that some (all?) software has some unpatented prior art they will never "get it right".
One could probably get a patent on various ways (methods) of obtaining a square root (there are lots), including one called the "odd integer method", that was implemented on a computer in the 1960's (exercise for reader, which one?).
Lots of the "reform" needs to be in the attitude that the patent office takes. One story is that they didn't hire "computer science" majors because the patent office doesn't do science. Perhaps there should be a "public comment" period on proposed patents so we can all contribute to the "prior art" knowledge base.
As I write this my cube-mate is working on a patent. Go figure!
In the US, prior art must be published and available to the public.
Herby, in the US, prior art also includes “defensive publications” (late 1960s to mid-1980s) and “statutory invention registrations” (currently). Thus, (for example) had the suits at Xerox been sufficiently concerned about their PARC inventions being patented out from underneath of them, they could have filed defensive publications at the time to establish prior art to prevent others (as well as preventing themselves) from obtaining patents on their inventions.
Re: In the US, prior art must be published and available to the public.
> Herby, in the US, prior art also includes “defensive publications”
That may be true. However, the vast majority of prior art is neither patented nor published in papers in the software realm. Most of this body of knowledge is obvious to those skilled in the art but evidently not those at the USPTO.
cart before the horse
skelband, if something is neither patented nor published* and available to the public, then by the USPTO’s definition, it is not prior art. The ability to write code to perform a particular task could well be within the ability of someone skilled in the art, but it is not necessarily obvious that someone has already done so unless there exists proof such as a patent, a publication, &c.
* — “Published” is not necessarily synonymous with “printed”: see §2128 of the Manual of Patent Examining Procedure (http://www.uspto.gov/web/offices/pac/mpep/s2128.html).
Re: cart before the horse
> skelband, if something is neither patented nor published* and available to the public, then by the USPTO’s definition, it is not prior art.
What you say makes sense but the body of all human knowledge is not encapsulated in papers and patents.
The USPTO cannot rely on these resources alone to determine what is prior art.
See John Smith 19's comment to see his echo of my view on this.
The irony of this situation is that really cutting edge , technical innovation may succumb to more stringent constraints than that which is pretty straightforward and obvious since it is likely that there is a larger body of relevant material available to the USPTO of this type of patent idea.
fun with modals
skelband, the USPTO does not claim that the body of all human knowledge is encapsulated within papers and patents. However, it is perfectly clear that the USPTO can rely on these resources alone to determine prior art, because that is exactly what they do. I don’t think that anyone claims that the USPTO is infallible, or that their examining procedure is the crown of Creation; they have tasks to do, set by the contemporary combination of Constitution, Congress, and the Court, and a finite budget with which to accomplish those tasks. That budget limitation is probably the main reason for their narrow definition of prior art. (And to address Herby’s point, perhaps one reason that they don’t hire more “computer science majors” is that the pay for being a patent examiner pales in comparison to the pay for becoming someone more “skilled in the art”.) Short of Congress legislating that no software-related invention will receive patent protection, the best (and least expensive) defence against such patents is establishing prior art via §2128 electronic publication.
How much s**t have large corporations gotten away with over the years?
The Lisa and Mactinosh interfaces were blatant copies of the Xerox Alto in their core ideas.
Which should not be too surprising given Apple hired several people from around the project.
If "prior art" really means only stuff in previous patents then it's pretty clear why the system is f**ked.
Historically those tricks of the graphics art (for example) were passed on by working in the labs like Xerox or MIT, or later in books. Books like the "Graphics Gems" series started documenting them (partly) to try and stop this BS. But a lot of the early graphics patents could have been (and might have been) lifted wholesale from "Principles of Interactive Computer Graphics," (either book would do).
But the other puzzle is that patents are not meant to be granted if the idea is "obvious to some one versed in the art". Changing a copy of the screen image and swapping the fully changed version with the current one is obvious to someone "versed in the art," but that's been patented.
So let's see what happens. It should be clear the US software patents system is FUBAR.
"US PATENT OFFICE SEEKS PUBLIC INPUT ON SOFTWARE PATENTS' FUTURE"
Solution: Throw software patents where they belong - in the bin.
So I'm NOT the only crazy one here!
"how to properly handle software-related patents"
Simple answer: Don't allow them.
Software Patent's Future
Nasty, brutish and short, to paraphrase Hobbes.
"to paraphrase Hobbes"
I don't recall a stuffed tiger ever saying that...
stupid business people
I was recently approached to give a quote to company for making an iPhone app.
I laughed them out of my in-box when the bottom of the email said they wanted us to work with them to register a patent for it.
I have no idea if that would fly or not, but there is no way as a non-US citizen or resident, i would go near that stuff....
no software patents
Given that programming languages are a language you shouldn't be able to patent software any more than you could patent speaking.
In theory you could have something generate every possbile algoritm anyway, whether it be with monkeys and keyboards or another computer. So simply discovering yet another combination of keywords, operators, etc isn't genius enough to patent.
Re: no software patents
Agree, it's a bit like patenting the sentence "Once upon a time there was.." and then having companies fighting over patent infringement regarding "was a Queen" and "was a King".
Re: no software patents
"In theory you could have something generate every possbile algoritm anyway, whether it be with monkeys and keyboards or another computer. "
I have the patent for that.
Re: no software patents
I have patented this too, "Once upon a time" along with rounded corners, rectangles, air, water, dirt and all other modern inventions.
Re: no software patents
Don't be daft, of course that sort of thing isn't patentable. "Once upon a time there was... ...on a touchscreen" however IS!
Re: no software patents
In theory you could have something generate every possible algorithm anyway, whether it be with monkeys and keyboards or another computer.
Wouldn't Alan Turing own all software patents then? Since any program can be represented by a turing machine which in turn can be represented by a number......
But we wouldn't know when to stop. would we?
Compel the submitter of a patent to disclose how many competitors' apps and how many non competitors' similarly-capable apps the submitter tried to use. Force the submitter to disclose whether or not the alternative programming or processing apps fully, partially, and unacceptably recreated the purported novel features claimed in the patent filing.
Get the submitter to agree under oath and risk of finding of purjury and punishment of forfeiture of the right to in the future submit a patrnt by self OR proxy, and to disclose all parties already involves in the patent being submitted, as well as unnamed sponsors of it, to prevent doing an end run around the filing ban punishment.
If my my purported wonder app can be reproduced in Alpha 5, ms access, Borland paradox, filemaker, et al, thenthe ONLY legitimate claim I can or should be allowed to claim would be the actual style or design, rendering my application null, and steering me to a frickin COPYRIGHT in most cases. If lucky, i might qualify for a restricted utility or function patent if the competing apps are too burdensome to provide to a regular user the output or functionality sought or claimed.
**I** am willing to accept those conditions or limitations, all in the name of leveling the playing field so poor devs and ordinary people can try to make a living without fear of being screwed over by corps or shotgunning lawyers blasting out illegit scare letters.
Re: One Approach
I posted almost the same idea.
Copyright, not Patent, is what is appropriate for software.
Few ideas for reform
The best solution would be to simply abolish software patents, but since we haven't a hope of getting that, there are a few reforms that would render them less problematic.
1: Err on the side of refusal. Apply financial penalties to the applicant for an invalid patent, for applying, and to the approver for approving it. The penalty would be proportional to the number of patents the applicant has, and how many previous 'offences' the applicant has committed. [Reason: Discourage "chancer" patents ]
2: Implement a situtation where unasserted patents are invalidated. An assertion could however simply be a communication to the allegedly infringing party, informing them of which product allegedly infringes which patent. Damages would not be due where the alledged infringer promptly removes such features from their products. [Reason: discourage submarine patents]
3: Charge a fee for patents, proportional to the number of patents an applicant already has. For example if an applicant has 10 patents, charge £10 for the next one. If another applicant has 1m patents, charge £1m for the next one. [Reason: discourage "salami" patenting].
4: Invalidate patents that the owner is not actively using. [Reason: Discourage "non-manufacturing-entities" ]
5: Establish terms proportionate to the softwares lifecycle - perhaps three years. [Reason: 20 years is ridicules]
Both for #1 and #3, suitable measures would have to be taken to prevent a loophole where the company simply spawns loads of subsidiaries or something like that.
Re: Few ideas for reform
Personally I think that #1 and #5 are the most reasonable and likely to be acceptable. I would also add
6. patents must be provided with a minimal code working example, as the idea of a patent is that it is of something that can actually be produced, and while for other disciplines it may be economically non-viable for a working sample of the invention to be produced. With software working versions are much, much easier to produce.
It would also serve to automatically narrow overly broad patents.
Set higher standards for trivial patents
1. The USPTO should set higher standards for non-trivial patents. Most patents we read about are very trivial.
2. The USPTO doesn't do much in the way of searching for prior art; they figure that can be done later if someone launches a dispute.
True, so make the dispute mechanism regarding evaluating prior art simple and free. If someone sends in prior art, there should be no need for lawyers or courts to get the patent tossed.
3. It should not matter if the patent or the body disputing the patent is or is not present in the USA. There should be no national preference.
Re: Set higher standards for trivial patents
1. Congress doesn't give them a big enough budget to do the research.
2. Again, Congress doesn't give them a big enough budget to do the research. It takes money and manpower to do the research, and if you raise fees, you discourage people from applying.
3. Basic American judicial law states that in order to file a suit for something, you must be AFFECTED by the law or policy in question. In other words, it must have harmed you in some way or the judge is required to throw out the suit as frivolous.
Needs to return to public domain within 5 years.
The ownership as it stands is far too long.
Don't complain here. Complain *there*.
US commentards should check the Federal Register (You'll need to use the FedReg link in the article. The URL is too damm long) notice. It gives full details of when and where. BTW You need to register by Feb 4th. While it may be webcast that does not necessarily mean they will accept questions from the web.
If you think software patents are too long/broad/stupid say so.
Likewise if you think they are too short/narrow/difficult-to-get say so too.
I think the argument is quite a lot like file sharing of copyright works. Some think they should be free, some think a $150k fine is not enough. Most would pay what they feel is a fair price if they felt most of the payment was going to the author of the work, who is providing the core creative bit after all.
The many different aspects around software patents Vs normal patents (and patents for methodologies WFT is that about?) suggest there can be many different viewpoints but I feel there is a core consensus that can be found and the majority could live with. I don't think that's what exists now.
Re: Software is not patentable - its copyrightable.
I would never have thought it possible, but I actually agree with Eadon! Wow :)
Re: Software is not patentable - its copyrightable.
The problem is that you can also write down the design for a new fuel injector, formula for concrete, process for producing a drug, etc. And given the heavy automation in those industries, we're almost at the stage where you could produce said items by purely writing them down and the automated processes would then just produce the new item.
The difference is that with software finding out if it works or doesn't work and whether it's fit for purpose takes at most weeks (excepting non-obvious bugs), but for the other items it takes years.
How to solve USPTO underfunding...
Drop software patents.
Software built into hardware should be patentable ie BIOS and such
Software that is built into the OS or required for a function should be patentable if it is unique and new ie protocols such as Ethernet and video codecs
All others should only be copyrightable ie a copyright on slide to unlock and a max fee chargeable as being standard ie all copyrightable items available for a fixed fee of say $1 for a large block with some being able to be withheld such as the look and feel of the UI, ie Apple UI can not be copied but similar is allowed, after all a copyright on a book about Wizards doesn't stop others writing about wizards but it does stop them stealing Harry Potter as a character to use in their book.
> Software built into hardware should be patentable ie BIOS and such..
Eh? How does that work?
How on earth do you define that? Does it include a device driver?
Perhaps you mean any software on flash. How about all the stuff on my SSD?
Perhaps you mean any software on ROM. Does anyone do that any more? How do you upgrade it?
Perhaps you mean FPGAs. In what way is it different?
Why make the distinction anyway? How does it help?
What to do about software patents.
Take them out the back and put a bullet through their head. Quick and painless.
Alternatively, if you must have software patents, then do the following.
1. Drastically reduce the time they are valid for.
2. Force those who apply for them to submit a working example in the ISO standard language of their choice.
3. Limit the degree to which they can be transferred.
An alternative to 2. would be to give a limited time period in which the working example must be submitted or the patent is invalidated. The reason for a working example is to prevent the expansion of claims when somebody comes up with something similar. It is also intended to make it difficult to have open ended claims since it must be encapsulated in the code.
Re: What to do about software patents.
> An alternative to 2. would be to give a limited time period in which the working example must be submitted or the patent is invalidated.
I would argue that should be a requirement for *all* patents.
Like trademarks you should have to use them or lose them.
They should not be transferable, loaned, bought or sold.
Also don't allow larger companies to take them out, they're for small inventors to protect their ideas not for big companies to whack their rivals.