The patent trolls are already on it, patenting crowd-sourcing as a method of patent defense. This will inevitably result in the shut down of Ask Patents in a stunning fit of irony.
Trouble is, what killed the patent was prior disclosure from the researchers.
The same thing has happened to friends of mine; an undergraduate's summer research presentation may have betrayed patentable inventions. An effect of the 1-year prior art rule is to force researchers to keep mum about what they're doing and to generate greater numbers of incremental patents.
If you're interested in the free flow of information, using a researcher's own publications to kill patents may not help the cause.
Prior art from other work in the past? Bring it on!
I'm not sure I understand what you mean by "prior disclosure" here. They filed a patent application: those are supposed to be public. The prior art Joel used was the documentation of the Win32 API from 5 years previously (and I'm pretty sure we've been using variable-resolution apps since earlier than 2008, and not just in Windows -- this was just an example). Could you clarify what the problem is?
App icons in OS X are interpolated between about 4 different sized images. I remember reading about this in the developer docs back in about 2002. I'd be surprised if Apple don't have pretty much the same patent on this already but registered 10 years earlier :-)
Mac OS has done this since much earlier. The classic Mac OS needed at least two different representations of the same icon, one for the desktop (32x32) and one for the application switcher (16x16). The application switcher, initially called MultiFinder, was introduced in 1987[1]. If memory serves, the various sizes of an icon were stored together in the ICNS resource of the app binary's resource fork. There was a developer tool provided by Apple called ResEdit which allowed you to view or change the icons of any application, in a nice GUI editor (of course).
I'm not 100% sure what you are confused about, but let me take a stab at it. Prior disclosure is public disclosure of the invention prior to the patent application. The one year rule states that once you publicly disclose an invention, you have one year from that date to file a patent application. Otherwise, your invention is considered public domain and cannot be patented. Does that clarify?
The problem I think you are referring to is that researchers run the risk of shooting themselves in the foot by prematurely presenting their findings unless they file an application in the process. Doing so would ruin their chances of being granted a patent.
I understand how prior disclosure works, my confusion is that nothing like that seems to have happened here. The patent was invalidated on the basis of documentation of a different piece of software with the same features released 5 years prior (albeit by the same company).
However, having read the ruling, I see it explicitly cites the prior-disclosure rule. So the bad patent system has invalidated a bad patent, but in a bad way.
I didn't think from reading it that the two groups at MS were the same - I assumed they were independent parts of the org, in which case it was fairly clear that this was an idea that various people have had without direct communication, and thus throws serious doubt on "non-obvious". I'm pretty sure that you could find plenty of other people who've had the idea.
Companies like Akamai used to (probably still do) have image resizing proxies which took something like a multi-resolution TIFF file as part of their extended CDN features. I can't remember when I first saw that in action, but I'm sure it was pre-2011.
I'm also not sure why using a researchers own publications to kill a patent would be bad in your eyes. If you think something is really, truly patentable you'd keep it under wraps until you did so, as far as possible. If you only decided later, once you'd published research and people were productizing that research - and you then patented and sued them - well, that's exactly the kind of thing that should be stomped on!
If you think something is really, truly patentable you'd keep it under wraps until you did so, as far as possible.
What this encourages though is, "If you think something is maybe patentable, you can't ever tell anyone about it until the invention is complete and fully functional." Nifty side projects occasionally become a big deal. For precision hardware, development time is often 5-10 years. For young researchers, it's critical to be able to talk about whatever it is that you do.
Completely agreed that bait-and-switch is unethical and wrong.
The 1-year grace period is there specifically so researchers can publish and patent. What counts are the publication and filing dates. File within the year and you will be in a position to earn protection.
However, the 1-year grace period isn't available everywhere. The USA has it, for instance, but the EU doesn't.
Its actually some part of governing delegated to the public. A part they can be good at. I wonder if the crowd can be authoritative, rather than just helpful. That would let them tackle obviousness, not just prior art I could see a stackexchange-like. It might be that they can tackle it now, if good methods/guidelines exist for objectively determining obviousness.
Pure Awesome. Shame there's not a way to do the same thing to existing software patents. There are plenty (~40k/year according to the article) of bad, obvious nonsense patents already out there. It'd be nice if there were a simple process to appeal and invalidate them with similar demonstrations of prior art and obviousness.
I'm definitely looking forward to seeing some Wikipedia-level-OCD focused on this site to stop the roughly 100% of bad applications for new software patents dead in their tracks. Imagine a 2014 where zero new software patents were issued.
EDIT: Incidentally, patentlyobvious.com is just a parked domain at the moment. It seems like the obvious choice for a place to host a site like this.
Can you give an example of a patent being helpful in producing innovation and prosperity in the last 20 years?
IE, a patent that was an original idea, someone else tried to duplicate in a blatantly rip-off way before the patent expired, and the ensuing lawsuit lets the patent creator (who has their patent to-market in a useful state) take the copycat off the market?
Absolutely correct. I used to work in this industry, and one of the 1st questions asked when we were thinking about testing a drug to treat a disease was "Does someone else 'own' this drug or do we?"
If someone else 'owned' the drug, then there was no point in pursuing it because the true "owner" would crush us if the drug ultimately proved useful in treating disease.
Interestingly, the people writing the drug patents use the same types of techniques mentioned by Joel...especially making claims as broad as possible.
The chemists that I worked with considered it an art form to be able to make broad claims that did not intersect with those of competing pharma companies.
There's a lot of bullshit that can go on w/ pharma patents, too, but the VCs would never fund a biotech startup if we couldn't defend "our" drugs.
While I'm sure a case could be made that IP protection is necessary in biotech, isn't the example you gave actually evidence of the the opposite -- that patents stifle innovation and research?
It depends on what you want to get out of the pharmaceutical industry.
I think it's the opposite. Since it costs so much $ to fund clinical trials and apply for regulatory approval, the lack of patents would stifle innovation because the pharma company that discovered the drug wouldn't be able to recoup their development costs if other companies were waiting in the wings, ready to sell a generic version of the same drug.
If development costs (especially regulatory approval costs) go down significantly, then I think you can argue that drug patents are an unnecessary evil. But if things stay as they are, then they are a necessary evil.
I think the idea is that then they couldn't charge exorbitant prices to recoup their investment costs. Drugs cost an enormous amount of money to develop.
Put yourself in their shoes. It's going to cost you $50MM-500MM just to test if a drug is effective and safe for treating a disease.
You sink all of that money in, with the hopes that your drug gets approved. And then, once you've done the heavy lifting, every generic manufacturer can come along and sell the drug because you don't have it patented?
You must consider though how the massive test run costs associated only exist because the FDA believed drug companies could foot the bill since they got monopoly distribution.
In a patent-free world, those regulations wouldn't hold weight. What would happen (I imagine) is independently sponsored r&d produces drugs for trial that the FDA itself puts through clinical trials on taxpayer money. You can't sell a drug without FDA approval still, but since drug companies aren't operating as super-for-profit businesses on chemical compounds to help people live, they would foot the bill to make sure its safe.
> Interestingly, the people writing the drug patents use the same types of techniques mentioned by Joel...especially making claims as broad as possible.
Systems (legal, computer, etc.) need to be developed with the full expectation that people will do everything in their power to exploit them.
I don't necessarily call that a good thing. It isn't a black and white pro that a business gets monopoly rights to distribute new drugs for 2 decades, even if most new drug research is funded for that profit.
You're assuming that patents promote innovation once they're defended in court.
Another argument would be that they promote innovation by (1) preventing blatant rip offs in the first place, (2) incentivizing individual inventors who invest their time and money with the assurance that their efforts won't be blatantly ripped off in the end by someone with more resources.
There are countless examples of truly innovative patents filed by individuals who devoted their lives to creating new things. Some of the inventors profited directly from their patents, others spent their life savings in court trying to defend their patents.
There are also countless patents for idiotic ideas, and countless idiotic patents that shouldn't have been filed in the first place. But saying all patents are bad might be throwing out the baby with the bathwater.
I have to think that the patent in Spolsky's example cost at least a couple grand to produce/submit. I also would love to see the cost per rejected patent produced by the USPTO that comes from US tax payer dollars.
Actually you do have to worry about that. Since the USPTO is financed by patent application fees it has a perverse incentive to continue promiscuously granting as many patents as possible.
You may have to worry about the incentives, but you don't have to worry about any cost in tax dollars of rejecting applications.
Unless somehow they began rejecting so many applications, and those rejections all cost more than the fees collected from those applications, so as to total more in losses than the surplus revenue of the rest of the PTO (all other patent and trademark activity, including renewals) over its costs.
Since the USPTO is financed by patent application fees it has a perverse incentive to continue promiscuously granting as many patents as possible.
I would hope that filers for patents that are rejected would still get charged a fee. If they aren't, they should be. After all, it still takes up a patent clerk's time.
1) Pretty much each and every action has a fee attached [1].
2) Each rejection has a 3 - 6 month time limit to be responded to.
3) Most applicants will fight every rejection to preserve their rights.
4) Granted patents have (admittedly higher [3]) fees due only every 3, 7 and 11 years, whereas patents in prosecution typically have fees due every 3 - 6 months. However, granted patents are not guaranteed to be renewed, which means their contribution is reduced [4, 5].
So if you want to talk perverse incentives, the USPTO actually have greater incentive to issue more rejections and collect frequent short-term fees on applicant responses, than they do to issue patents and collect infrequent long-term fees on renewals.
This myopic myth-busting of yours assumes that the number of patent applications remains unchanged regardless of how many are granted when in reality more granted patents fuels the fire of more applications. If the USPTO did the right thing and severely limited or even eliminated software patents there's no question that the number of applications, and thus the USPTO's overall budget and staffing levels, would decrease dramatically.
Apologies for the late response. I don't check regularly.
1. There is no evidence to back your thesis that "more granted patents fuels the fire of more applications". See this graph [1] of allowance rates in 2000 - 2005 (for an explanation, ask a patent attorney or agent about "Dudas") and this table [2] for number of applications over that time period. There is zero correlation with the steep drop in allowance rates and the rate of new patent applications over the relevant time periods.
2. Why do you think your idea of "the right thing" is actually the right thing, if any such thing exists? Just because the HN echo chamber agrees with you? As I have shown time and time again, the Internet, especially tech media, has no real knowledge of how patents work.
3. The Supreme Court, let alone the CAFC, the PTO and various courts, cannot even define what "software patents" are. How would you begin to limit or eliminate them?
4. None of the above entities have found sufficient rationale to exclude business method patents, let alone software patents. Contrary to popular belief, it's not because they are stupid or technically illiterate, it's because they are very careful thinkers and make the best of what they can given their boundaries.
5. The budget and staffing of the USPTO is, unfortunately, mostly independent of the number of patent applications they receive. Currently, they generate more fees than they are allocated a budget for, and the federal government siphons away the rest.
1. Thanks for the second link - both "Total Patent Applications" and "Total Patent Grants" show a similar upward trajectory, both roughly doubling since the late nineties. The first link is uninteresting - acceptance rate fluctuations don't matter much in the face of steadily rising numbers of both applications and grants.
2.3.4. I come to my opinions as a veteran professional programmer, a named inventor on three US software patents (which I now regret signing off on), a software business owner, and a long-time avid follower of this issue. I'm sure that others, perhaps yourself included, may know more about the ins and outs of the patent legal system than I do, but I submit that you and your patent-defending ilk can't see the forest for the trees. The forest is that the vast majority of software-related patents are total bullshit. Given your role here as patent apologist, I don't think I will try to make a case to you why that is; I don't think that would be time well spent.
5. Find me a chart like your second link which shows the staffing levels of the USPTO over time. I'll bet it shows the same upward trend as patent applications and grants.
I think the problem is that the proverbial small inventor is extremely rare, and if they are indeed small enough for us to take pity on them, then a patent isn't enough to protect them anyway.
Make the prior art fee scale up by number of patent applications submitted by that entity, then. First one's free. Then $100, then start growing at some small-ish exponent.
> Shame there's not a way to do the same thing to existing software patents.
What's stopping anyone from crowd sourcing a database of prior art for existing patents that would invalidate them if it came to it?
A valid defense against patent infringement is to show that the patent shouldn't have been issued in the first place, right? So a database of Ready Made Legal Defense sounds good.
THe key part is recent changes in the law have given much more weight to community/public commenting on patents, so where these may have failed in the past it is possible for them to have more success today.
Further the patent climate has become more hostile towards bad patents in recent years, more so than at in point in modern history, this will also enable more possibility for success
Your assertion remains to be seen.
There are still significant downsides to submitting prior art to the PTO that also existed when bountyquest was around.
The main serious issues are still the same, in fact:
1. If you submit prior art pre-grant, that's basically that. You don't get to argue about it, only the examiner does. If the patent issues, you've now lost a possibly valuable reference if the examiner didn't use it well enough.
2. If you challenge in any of the post-grant proceedings, you'll be estopped from suing on anything you raised or could have raised.
Showing a patent clerk prior art is much easier than showing prior art in a courtroom.
For instance at the minimum you need to bring in an expert witness to explain the prior art. Then you have to contend with the fact that they will bring in their own expert witness to explain how your example is bogus and you both get to cross-examine. etc. etc.
This is an excellent result. I worried when the idea of helping the PTO uncover prior art would be controlled by the attorneys filing the patent (which is to say they would be a filter between the examiner and the external sources) but if the examiners are going directly to the source then this will really put a crimp in bad patents being issued.
> Since patent examiners rely so much on keyword searches, when you submit your application, if you can change some of the keywords in your patent to be different than the words used everywhere else, you might get your patent through even when there’s blatant prior art, because by using weird, made-up words for things, you've made that prior art harder to find.
Wouldn't it also make sense to build up, maybe at the same time, a sort of "counter-thesaurus"?
As in: If you find prior art where the thesaurus method has been used to obscure terms in the patent, enter those as an example into a database. When another patent is looked up, individual terms that show up in the database have a "there are alternative terms for this" marker applied to them.
This might also make it a lot easier to make automatic search for prior art feasible again.
So, this sounds great and all, but I don't see any proof that it was ask patents that caused the examiner to find this, rather than the examiner's standard search?
In fact, the search history, on PAIR (look for 4-11-2013 SRFW Search information including classification, databases and other search related notes) does not say ask patents was used, nor does the search strategy (document code SRNT).
It could be the USPTO has not gotten around to noting this yet, but the only entry I see that could be related is the NPL entry, which of course, has no image available (god i love PAIR), and does not say it came from outside the search.
Don't get me wrong, I think askpatents is great, but i'm skeptical considering how examiners actually work.
The USPTO explicitly asked us to set up this site to help get the public involved in getting prior art to the examiners. (Press release:http://www.uspto.gov/news/pr/2012/12-60.jsp.) So, while it's entirely possible that the examiner might have eventually found this without us, which is far from certain, it seems reasonable to assume that they saw it on the site. Which means that even if you assume they would have found it without help - again, questionable, given the number of weak patents out there - at a minimum, we probably saved them a lot of time, which frees up more resources to look for prior art on the next crappy patent. And that reduces the odds that the clock runs out without them finding any, which leads to a lousy issuance.
Is it the examiners who post some of the requests for prior art? If so, it would be nice if their accounts or requests could be badged so other users can be reasonably sure that the research they do will be seen by the USPTO.
I'm aware the USPTO asked you to set up ask patents. The USPTO has a number of pilot programs, and i'm involved in some.
My only question is whether AskPatents was how it got to the USPTO.
I don't see why it's reasonable to assume that they saw it on the site considering most examiners still don't do that sort of thing. Most of them search the USPTO approved databases, internally, and go about their way. This is in fact, what the search strategy/et al says happened.
Given that, i think the burden falls on you to show some likelihood here.
For example, do you have anything to say that any significant percentage of examiners (IE > 10%) use askpatents on a regular basis? I would be pleasantly shocked if you did.
The rest is not the argument i am making, the only argument I am making is: "Do you have any evidence that your site is the reason the USPTO found this?"
From what I can tell, your answer is "no".
If you do, great!
No offense meant, of course, my problem is if you declare victory when you were not actually involved, this will actually make things worse - people will think they are helping solve the issue, yet, if the PTO actually isn't using your data heavily, they aren't.
Again, don't get me wrong, i think askpatents is a great idea, and i'd hope and love to find out it's actually being used heavily by the PTO. I can't find any stats on that, and given the history of prior pilot programs/etc, i'm very skeptical.
No offense taken, and thanks for the overall support!
As to:
> "Do you have any evidence that your site is the reason the USPTO found this?"
Well, the office has told us directly that examiners are reviewing the site. So, when they then reject an application, based primarily on art posted on the site, that was asked for and posted there just prior to the actual review, it makes me highly optimistic that it's not all a coincidence, although I suppose that's theoretically possible.
This is a great first step, but can we sustain this and keep more patents from being granted. Everything hinges on community involvement, so hopefully we can build and sustain a community that does this every day.
Sadly a patent examiner (in the UK at least) gets less than a day to do a prior-art search on the most complex applications in the most complex subject matter areas; half a day is probably the standard.
There are half-a-million patent applications per annum in the US alone (10 million since 1963). I'd estimate a global corpus of billions of patent applications. Prior art search extends though [in theory] to all publications.
You can see why prior art submissions are allowed from third parties. If you want to ensure only valid patents, or perhaps as few patents as possible, are granted then making such submissions is in your own interests.
Edit: Actually the gross figures are misleading, but remember these are complex 20+ page [in normal print size] documents on the whole. G06F11 - an IPC classification concerning error correction, detection and handling has ~9000 docs in the espacenet database for each of 2010/2011/2012.
> How cool would it be if Apple, Samsung, Oracle and Google got into a Mexican Standoff on Ask Patents? If each of those companies had three or four engineers dedicating a few hours every day to picking off their competitors’ applications, the number of granted patents to those companies would grind to a halt.
I'm kind of surprised this isn't happening already.
> “Sometimes you have a picture that you want to scale to different resolutions. When this happens, you might want to have multiple versions of the image available at different resolutions, so you can pick the one that’s closest and scale that.”
I read that Lance Williams invented this technique in 1983, but I remember seeing a documentary on Tron where they had to come up with a technique during the lightcycle race to reduce the Moire effect on the grid and I could have swore that they invented mipmapping to do that, even though it was a year earlier.
I had a bad attitude dealing with patents in my last job, refusing to take part on the basis that I hadn't come up with anything novel. I noticed a member of staff with views on patents in line with mine taking a more active role. He took part in the patent meetings, but instead of merely offering nothing up, took active part in finding prior art, thus preventing time wasting patents going to the office. He took special pleasure in finding prior art by current employees at the same company, preferably in the same arm.
I am impressed by Spolsky's positive attitude to do something about the problem, I bid ask patents continued good fortune!
> The number of actually novel, non-obvious inventions in the software industry that maybe, in some universe, deserve a government-granted monopoly is, perhaps, two.
One thing I don't understand. Suppose a patent examiner does not ask on Ask Patents about a patent application, but I would nevertheless like to submit prior art.
The Ask Patents site was created in partnership with the USPTO; there is an expectation that examiners will search the site while researching a pending application.
> An example might help. Imagine a simple application with these three claims:
> 1. A method of transportation
> 2. The method of transportation in claim 1, wherein there is an engine connected to wheels
> 3. The method of transportation in claim 2, wherein the engine runs on water
> Notice that claim 2 mentions claim 1, and narrows it... in other words, it claims a strict subset of things from claim 1.
> Now, suppose you invented the water-powered car. When you submit your patent, you might submit it this way even knowing that there’s prior art for “methods of transportation” and you can’t really claim all of them as your invention. The theory is that (a) hey, you might get lucky! and (b) even if you don’t get lucky and the first claim is rejected, the narrower claims will still stand.
I'm pretty sure this is not accurate. To my knowledge, the claims of a patent are considered only as a whole, not individually, so there should be no fear that this is somehow trying to patent "methods of transportation" broadly.
To make a broad patent claim, you can't have any narrowing claims on the same patent. Put another way, the maximum scope of claim by a particular patent is defined by the narrowest claim in the list.
Not a lawyer, but I'm pretty certain that you're wrong. There are 2 types of claims, dependent and independent. dependent are those that refer to a previous claim and expand on it, (2 and 3 above). Claim 1 above is independent. The dependent claims are taken together with those they reference, but the independent claims if accepted are taken independently.
This is an awesome project. I've signed up and will see what I can do to help shoot down patents!
There is one true and important point in the article that isn't supported by the example given:
> This patent was, typically, obfuscated, and it used terms like “pixel density” for something that every other programmer in the world would call “resolution,” either accidentally (because Microsoft’s lawyers were not programmers), or, more likely, because the obfuscation makes it that much harder to search.
The patent uses "pixel density" to refer to the physical size of the pixels on a display. This is a fairly common term, with over five million results in a Google search. I've used the same phrase myself for over 10 years with the same meaning.
"Pixels per centimeter (ppcm), Pixels per inch (PPI) or pixel density is a measurement of the resolution of devices in various contexts: typically computer displays, image scanners, and digital camera image sensors."
Of course here we can see where the terms get a bit confusing: "...pixel density is a measurement of the resolution..."
But display resolution these days usually refers to the number of pixels, not their physical size:
"The display resolution of a digital television, computer monitor or display device is the number of distinct pixels in each dimension that can be displayed."
Advertisements and spec sheets almost always use "resolution" this way, for example the MacBook Pro specs on Apple's site:
"Supported resolutions: 1440 by 900 (native), 1280 by 800, 1152 by 720, ..."
Forgive me if this seems like nitpicking, and I completely agree with Joel's point here: patents often do use unusual terminology to obfuscate what they're talking about.
This just isn't a case of that. The patent is using the correct term, and it even does a very good job of explaining what it means:
> A particular characteristic of display components that may affect presentations rendered thereupon is the pixel density of the display component, such as a pixels-per-centimeter measurement. It may be appreciated that such characteristics may be independent of the size of the display component (e.g., two display components of the same size may present different pixel densities; conversely, two display components of different sizes may present the same pixel density) and/or the pixel dimensions of the display component (e.g., two display components displaying a presentation with a particular pixel dimensions may do so with different physical sizes).
> There is one true and important point in the article that isn't supported by the example given
I can't actually vouch for the "true and important" part. I must confess I put that in there to try to be nicer about what I was saying. :-) And it certainly seemed plausible enough that patents would be deliberately written in obfuscated language, I just can't say it's true or untrue from my own knowledge. I know pixels better than patents.
There's a great comment here that sheds some light on this:
> How cool would it be if Apple, Samsung, Oracle and Google got into a Mexican Standoff on Ask Patents? If each of those companies had three or four engineers dedicating a few hours every day to picking off their competitors’ applications, the number of granted patents to those companies would grind to a halt.
Wouldn't the result of this be that the company who finds out that they have the earliest implementation of the code in question be able to get the patent for it?
But that could easily be for an idea that's proven to be over 17 years old (<1996 and counting up). So they'd be digging through the floor, and might fall through.
The first publisher of an invention has a year to file a patent on it. Because of the length of time taken to get to a filing, by the time this prior-art shootout takes place more than a year will have gone by and the prior art will actually disallow anyone from patenting the invention, including the original inventor.
It would be cool if there was a way to penalize high number and high ratio of rejected patents from a single company. OTOH I don't see a way that allows genuine mistakes from small shops while being immune from big bad corps acting via shell companies.
That is an excellent idea. (To clarify, you're saying a company would need to have both a high number and a high ratio rejected to be penalized.)
More generally, it seems logical to have the size of the penalty increase at a quadratic or even exponential rate, based on the number of rejected patents within a certain trailing time period[1], and scaled by the rejected/submissions ratio. So every rejection would carry a penalty, but the more that were rejected (both in total and as a percentage), the larger the penalty would be, at an increasing rate.
It might also be an idea to do this with individual claims as well as, or even instead of, entire patents, to also deal with the overly broad claims problem.
[1] Or even for all time, but with a weighting factor based on recency. Wouldn't want to get overly complex though.
How about just increasing the costs for filing and maintaining a patent as a quick way of getting rid of some of the really weak ones? Say double the fees and see what happens.
Any flat increase would just make it less likely that the big guys will have competition in their "most patents filed" competition.
Also, given we just switch from first-invented to first-filed, it is particularly onerous for those garage-based small inventors as a big corp can easily justify the cost.
What might make more sense is to tax intellectual property. We tax physical property with much less value than a "successful" software patent (ahem, Eolas?), why not tax the non-physical property as well? If you make it progressive based on the number of patents owned/controlled, it will also limit the usability of maintaining warchests of patents for controlling markets.
Perfect. Tax them heavier as they age. That way the people trying to create usable products are barely hurt, but by the last few years it costs a lot to hold on to. And you can release your patent to the public domain to release it. That's more likely to hit trolls than innovators.
An idea that might help in killing patents: Have one daily sticky post on HN (ala job post) with a short description of a pending patent and a link on where to post prior art. It is very likely some reader will be able to provide this prior art.
At 40,000 software patents per year, it would require 110 posted patents per day to actually stop them, or a significant fraction of that to make a serious dent.
One a day would be less than a 1% reduction in software patents. We need to shut them ALL down.
It would be great if others could attach bounties to certain patents. Of course then the patent examiner would have to pick "winners", answers that helped the most.
Also, like Quora, does the StackExchange system allow one to register one's fields of expertise, to have questions suggested to you? (I'm not a big participant on SE but I know it may be hiding that feature from me, because it slowly reveals features based on karma).
Anyway I'm sure this has been thought of before, just curious if it's on the todo list.
StackExchange has a "favorite tags" feature which will highlight specific tags on the front page (and for busier sites, it may even bias its display on them, though I'm not sure about that).
There is also an advanced "filters" functionality (visible at stackexchange.com/filters) which apparently lets you filter out all but certain tags, as well as have 15-minute/3-hour/daily emails with new questions in that tag. I have never used it, but it sounds like what you're aiming for.
You can post bounties on Stack Exchange if you have enough reputation to cover it. Whoever posts the bounty (not necessarily the patent examiner) gets to pick who wins the bounty. If they don't, it can be automatically awarded to the highest-scoring answer during the bounty period (one week). More details: http://meta.stackoverflow.com/questions/16065/how-does-the-b...
154 comments
[ 4.4 ms ] story [ 233 ms ] threadIf only we'd done this a decade or two ago. How about some reevaluation of granted patents?
edit: The original title, "Joel Spolsky, patent killer", was better :/
The same thing has happened to friends of mine; an undergraduate's summer research presentation may have betrayed patentable inventions. An effect of the 1-year prior art rule is to force researchers to keep mum about what they're doing and to generate greater numbers of incremental patents.
If you're interested in the free flow of information, using a researcher's own publications to kill patents may not help the cause.
Prior art from other work in the past? Bring it on!
[1]: https://en.wikipedia.org/wiki/Multifinder
The problem I think you are referring to is that researchers run the risk of shooting themselves in the foot by prematurely presenting their findings unless they file an application in the process. Doing so would ruin their chances of being granted a patent.
However, having read the ruling, I see it explicitly cites the prior-disclosure rule. So the bad patent system has invalidated a bad patent, but in a bad way.
Companies like Akamai used to (probably still do) have image resizing proxies which took something like a multi-resolution TIFF file as part of their extended CDN features. I can't remember when I first saw that in action, but I'm sure it was pre-2011.
I'm also not sure why using a researchers own publications to kill a patent would be bad in your eyes. If you think something is really, truly patentable you'd keep it under wraps until you did so, as far as possible. If you only decided later, once you'd published research and people were productizing that research - and you then patented and sued them - well, that's exactly the kind of thing that should be stomped on!
If you think something is really, truly patentable you'd keep it under wraps until you did so, as far as possible.
What this encourages though is, "If you think something is maybe patentable, you can't ever tell anyone about it until the invention is complete and fully functional." Nifty side projects occasionally become a big deal. For precision hardware, development time is often 5-10 years. For young researchers, it's critical to be able to talk about whatever it is that you do.
Completely agreed that bait-and-switch is unethical and wrong.
However, the 1-year grace period isn't available everywhere. The USA has it, for instance, but the EU doesn't.
IANAL, TINLA.
The root of the problem, the patent system itself need to be abolished.
Its actually some part of governing delegated to the public. A part they can be good at. I wonder if the crowd can be authoritative, rather than just helpful. That would let them tackle obviousness, not just prior art I could see a stackexchange-like. It might be that they can tackle it now, if good methods/guidelines exist for objectively determining obviousness.
I'm definitely looking forward to seeing some Wikipedia-level-OCD focused on this site to stop the roughly 100% of bad applications for new software patents dead in their tracks. Imagine a 2014 where zero new software patents were issued.
EDIT: Incidentally, patentlyobvious.com is just a parked domain at the moment. It seems like the obvious choice for a place to host a site like this.
That would enable "Amazon Mechanical Turk"-style killing of most patents with prior art.
And the patent office would become even less effective. Perhaps a good thing, unless you believe some things actually should be patentable.
IE, a patent that was an original idea, someone else tried to duplicate in a blatantly rip-off way before the patent expired, and the ensuing lawsuit lets the patent creator (who has their patent to-market in a useful state) take the copycat off the market?
If someone else 'owned' the drug, then there was no point in pursuing it because the true "owner" would crush us if the drug ultimately proved useful in treating disease.
Interestingly, the people writing the drug patents use the same types of techniques mentioned by Joel...especially making claims as broad as possible.
The chemists that I worked with considered it an art form to be able to make broad claims that did not intersect with those of competing pharma companies.
There's a lot of bullshit that can go on w/ pharma patents, too, but the VCs would never fund a biotech startup if we couldn't defend "our" drugs.
I think it's the opposite. Since it costs so much $ to fund clinical trials and apply for regulatory approval, the lack of patents would stifle innovation because the pharma company that discovered the drug wouldn't be able to recoup their development costs if other companies were waiting in the wings, ready to sell a generic version of the same drug.
If development costs (especially regulatory approval costs) go down significantly, then I think you can argue that drug patents are an unnecessary evil. But if things stay as they are, then they are a necessary evil.
You sink all of that money in, with the hopes that your drug gets approved. And then, once you've done the heavy lifting, every generic manufacturer can come along and sell the drug because you don't have it patented?
No rational person/company will do that.
In a patent-free world, those regulations wouldn't hold weight. What would happen (I imagine) is independently sponsored r&d produces drugs for trial that the FDA itself puts through clinical trials on taxpayer money. You can't sell a drug without FDA approval still, but since drug companies aren't operating as super-for-profit businesses on chemical compounds to help people live, they would foot the bill to make sure its safe.
It would be interesting to see if your scenario would end up doing a better or worse job with respect to drug development.
Systems (legal, computer, etc.) need to be developed with the full expectation that people will do everything in their power to exploit them.
Another argument would be that they promote innovation by (1) preventing blatant rip offs in the first place, (2) incentivizing individual inventors who invest their time and money with the assurance that their efforts won't be blatantly ripped off in the end by someone with more resources.
There are countless examples of truly innovative patents filed by individuals who devoted their lives to creating new things. Some of the inventors profited directly from their patents, others spent their life savings in court trying to defend their patents.
There are also countless patents for idiotic ideas, and countless idiotic patents that shouldn't have been filed in the first place. But saying all patents are bad might be throwing out the baby with the bathwater.
Unless somehow they began rejecting so many applications, and those rejections all cost more than the fees collected from those applications, so as to total more in losses than the surplus revenue of the rest of the PTO (all other patent and trademark activity, including renewals) over its costs.
Which Isn't Going To Happen.
I would hope that filers for patents that are rejected would still get charged a fee. If they aren't, they should be. After all, it still takes up a patent clerk's time.
http://www.uspto.gov/web/offices/ac/qs/ope/fee031913.htm
1) Pretty much each and every action has a fee attached [1].
2) Each rejection has a 3 - 6 month time limit to be responded to.
3) Most applicants will fight every rejection to preserve their rights.
4) Granted patents have (admittedly higher [3]) fees due only every 3, 7 and 11 years, whereas patents in prosecution typically have fees due every 3 - 6 months. However, granted patents are not guaranteed to be renewed, which means their contribution is reduced [4, 5].
So if you want to talk perverse incentives, the USPTO actually have greater incentive to issue more rejections and collect frequent short-term fees on applicant responses, than they do to issue patents and collect infrequent long-term fees on renewals.
[1] http://www.uspto.gov/web/offices/ac/qs/ope/fee031913.htm
[2] http://www.uspto.gov/web/offices/ac/qs/ope/fee031913.htm#mai...
[3] http://www.patentlyo.com/patent/2010/06/patent-maintenance-f...
[4] http://btlj.org/data/articles/20_04_02.pdf
[5] http://www.uspto.gov/about/stratplan/ar/2011/mda_06_01_03.ht...
1. There is no evidence to back your thesis that "more granted patents fuels the fire of more applications". See this graph [1] of allowance rates in 2000 - 2005 (for an explanation, ask a patent attorney or agent about "Dudas") and this table [2] for number of applications over that time period. There is zero correlation with the steep drop in allowance rates and the rate of new patent applications over the relevant time periods.
2. Why do you think your idea of "the right thing" is actually the right thing, if any such thing exists? Just because the HN echo chamber agrees with you? As I have shown time and time again, the Internet, especially tech media, has no real knowledge of how patents work.
3. The Supreme Court, let alone the CAFC, the PTO and various courts, cannot even define what "software patents" are. How would you begin to limit or eliminate them?
4. None of the above entities have found sufficient rationale to exclude business method patents, let alone software patents. Contrary to popular belief, it's not because they are stupid or technically illiterate, it's because they are very careful thinkers and make the best of what they can given their boundaries.
5. The budget and staffing of the USPTO is, unfortunately, mostly independent of the number of patent applications they receive. Currently, they generate more fees than they are allocated a budget for, and the federal government siphons away the rest.
[1] http://www.patentlyo.com/patent/PatentlyO2006059.jpg [2] http://www.uspto.gov/web/offices/ac/ido/oeip/taf/us_stat.htm
2.3.4. I come to my opinions as a veteran professional programmer, a named inventor on three US software patents (which I now regret signing off on), a software business owner, and a long-time avid follower of this issue. I'm sure that others, perhaps yourself included, may know more about the ins and outs of the patent legal system than I do, but I submit that you and your patent-defending ilk can't see the forest for the trees. The forest is that the vast majority of software-related patents are total bullshit. Given your role here as patent apologist, I don't think I will try to make a case to you why that is; I don't think that would be time well spent.
5. Find me a chart like your second link which shows the staffing levels of the USPTO over time. I'll bet it shows the same upward trend as patent applications and grants.
Similarly, what do you do about the "I was accidently too broad" cases, which are just as dangerous.
That's worth a chuckle.
(In truth, if it were patently obvious, this would be a much easier problem...)
What's stopping anyone from crowd sourcing a database of prior art for existing patents that would invalidate them if it came to it?
A valid defense against patent infringement is to show that the patent shouldn't have been issued in the first place, right? So a database of Ready Made Legal Defense sounds good.
Am I missing something?
Many times.
patentbusters - ditto
Peer to patent - more recent
Patexia - still exists
article one partners - which still exists and pays
I don't feel like trying to hunt down the rest, but there are/have been at least 10 or so crowdsourcing attempts, some serious, some not so serious.
Further the patent climate has become more hostile towards bad patents in recent years, more so than at in point in modern history, this will also enable more possibility for success
The main serious issues are still the same, in fact:
1. If you submit prior art pre-grant, that's basically that. You don't get to argue about it, only the examiner does. If the patent issues, you've now lost a possibly valuable reference if the examiner didn't use it well enough.
2. If you challenge in any of the post-grant proceedings, you'll be estopped from suing on anything you raised or could have raised.
For instance at the minimum you need to bring in an expert witness to explain the prior art. Then you have to contend with the fact that they will bring in their own expert witness to explain how your example is bogus and you both get to cross-examine. etc. etc.
http://www.uspto.gov/patents/stats/Reexamination_Information...
Wouldn't it also make sense to build up, maybe at the same time, a sort of "counter-thesaurus"?
As in: If you find prior art where the thesaurus method has been used to obscure terms in the patent, enter those as an example into a database. When another patent is looked up, individual terms that show up in the database have a "there are alternative terms for this" marker applied to them.
This might also make it a lot easier to make automatic search for prior art feasible again.
In fact, the search history, on PAIR (look for 4-11-2013 SRFW Search information including classification, databases and other search related notes) does not say ask patents was used, nor does the search strategy (document code SRNT).
It could be the USPTO has not gotten around to noting this yet, but the only entry I see that could be related is the NPL entry, which of course, has no image available (god i love PAIR), and does not say it came from outside the search.
Don't get me wrong, I think askpatents is great, but i'm skeptical considering how examiners actually work.
My only question is whether AskPatents was how it got to the USPTO.
I don't see why it's reasonable to assume that they saw it on the site considering most examiners still don't do that sort of thing. Most of them search the USPTO approved databases, internally, and go about their way. This is in fact, what the search strategy/et al says happened. Given that, i think the burden falls on you to show some likelihood here.
For example, do you have anything to say that any significant percentage of examiners (IE > 10%) use askpatents on a regular basis? I would be pleasantly shocked if you did.
The rest is not the argument i am making, the only argument I am making is: "Do you have any evidence that your site is the reason the USPTO found this?"
From what I can tell, your answer is "no".
If you do, great!
No offense meant, of course, my problem is if you declare victory when you were not actually involved, this will actually make things worse - people will think they are helping solve the issue, yet, if the PTO actually isn't using your data heavily, they aren't.
Again, don't get me wrong, i think askpatents is a great idea, and i'd hope and love to find out it's actually being used heavily by the PTO. I can't find any stats on that, and given the history of prior pilot programs/etc, i'm very skeptical.
As to:
> "Do you have any evidence that your site is the reason the USPTO found this?"
Well, the office has told us directly that examiners are reviewing the site. So, when they then reject an application, based primarily on art posted on the site, that was asked for and posted there just prior to the actual review, it makes me highly optimistic that it's not all a coincidence, although I suppose that's theoretically possible.
There are half-a-million patent applications per annum in the US alone (10 million since 1963). I'd estimate a global corpus of billions of patent applications. Prior art search extends though [in theory] to all publications.
You can see why prior art submissions are allowed from third parties. If you want to ensure only valid patents, or perhaps as few patents as possible, are granted then making such submissions is in your own interests.
Edit: Actually the gross figures are misleading, but remember these are complex 20+ page [in normal print size] documents on the whole. G06F11 - an IPC classification concerning error correction, detection and handling has ~9000 docs in the espacenet database for each of 2010/2011/2012.
I'm kind of surprised this isn't happening already.
This is basically mip-mapping, and was described in a 1983 paper: http://dl.acm.org/citation.cfm?id=801126.
I am impressed by Spolsky's positive attitude to do something about the problem, I bid ask patents continued good fortune!
Any idea to which two he might be referring?
http://en.wikipedia.org/wiki/RSA_(algorithm)#History
http://blog.stackoverflow.com/2013/02/podcast-43-false-facts...
How do I do so?
> 1. A method of transportation
> 2. The method of transportation in claim 1, wherein there is an engine connected to wheels
> 3. The method of transportation in claim 2, wherein the engine runs on water
> Notice that claim 2 mentions claim 1, and narrows it... in other words, it claims a strict subset of things from claim 1.
> Now, suppose you invented the water-powered car. When you submit your patent, you might submit it this way even knowing that there’s prior art for “methods of transportation” and you can’t really claim all of them as your invention. The theory is that (a) hey, you might get lucky! and (b) even if you don’t get lucky and the first claim is rejected, the narrower claims will still stand.
I'm pretty sure this is not accurate. To my knowledge, the claims of a patent are considered only as a whole, not individually, so there should be no fear that this is somehow trying to patent "methods of transportation" broadly.
To make a broad patent claim, you can't have any narrowing claims on the same patent. Put another way, the maximum scope of claim by a particular patent is defined by the narrowest claim in the list.
Lawyers--correct me if I'm wrong.
Here: http://en.wikipedia.org/wiki/Claim_(patent)#Basic_types_and_...
There is one true and important point in the article that isn't supported by the example given:
> This patent was, typically, obfuscated, and it used terms like “pixel density” for something that every other programmer in the world would call “resolution,” either accidentally (because Microsoft’s lawyers were not programmers), or, more likely, because the obfuscation makes it that much harder to search.
The patent uses "pixel density" to refer to the physical size of the pixels on a display. This is a fairly common term, with over five million results in a Google search. I've used the same phrase myself for over 10 years with the same meaning.
http://en.wikipedia.org/wiki/Pixel_density
"Pixels per centimeter (ppcm), Pixels per inch (PPI) or pixel density is a measurement of the resolution of devices in various contexts: typically computer displays, image scanners, and digital camera image sensors."
Of course here we can see where the terms get a bit confusing: "...pixel density is a measurement of the resolution..."
But display resolution these days usually refers to the number of pixels, not their physical size:
http://en.wikipedia.org/wiki/Display_resolution
"The display resolution of a digital television, computer monitor or display device is the number of distinct pixels in each dimension that can be displayed."
Advertisements and spec sheets almost always use "resolution" this way, for example the MacBook Pro specs on Apple's site:
"Supported resolutions: 1440 by 900 (native), 1280 by 800, 1152 by 720, ..."
Forgive me if this seems like nitpicking, and I completely agree with Joel's point here: patents often do use unusual terminology to obfuscate what they're talking about.
This just isn't a case of that. The patent is using the correct term, and it even does a very good job of explaining what it means:
> A particular characteristic of display components that may affect presentations rendered thereupon is the pixel density of the display component, such as a pixels-per-centimeter measurement. It may be appreciated that such characteristics may be independent of the size of the display component (e.g., two display components of the same size may present different pixel densities; conversely, two display components of different sizes may present the same pixel density) and/or the pixel dimensions of the display component (e.g., two display components displaying a presentation with a particular pixel dimensions may do so with different physical sizes).
> There is one true and important point in the article that isn't supported by the example given
I can't actually vouch for the "true and important" part. I must confess I put that in there to try to be nicer about what I was saying. :-) And it certainly seemed plausible enough that patents would be deliberately written in obfuscated language, I just can't say it's true or untrue from my own knowledge. I know pixels better than patents.
There's a great comment here that sheds some light on this:
https://news.ycombinator.com/item?id=6085888
In particular, there are reasons for some of the language that seems obfuscated. Read the comment for details and a lot of insight into patents.
Another interesting comment, with a link to a patent examiner's AMA:
https://news.ycombinator.com/item?id=6087057
(Of course there are interesting comments all up and down this thread!)
Wouldn't the result of this be that the company who finds out that they have the earliest implementation of the code in question be able to get the patent for it?
More generally, it seems logical to have the size of the penalty increase at a quadratic or even exponential rate, based on the number of rejected patents within a certain trailing time period[1], and scaled by the rejected/submissions ratio. So every rejection would carry a penalty, but the more that were rejected (both in total and as a percentage), the larger the penalty would be, at an increasing rate.
It might also be an idea to do this with individual claims as well as, or even instead of, entire patents, to also deal with the overly broad claims problem.
[1] Or even for all time, but with a weighting factor based on recency. Wouldn't want to get overly complex though.
Maybe the fee should start low and then increase by some percentage with each new application you submit.
Also, given we just switch from first-invented to first-filed, it is particularly onerous for those garage-based small inventors as a big corp can easily justify the cost.
What might make more sense is to tax intellectual property. We tax physical property with much less value than a "successful" software patent (ahem, Eolas?), why not tax the non-physical property as well? If you make it progressive based on the number of patents owned/controlled, it will also limit the usability of maintaining warchests of patents for controlling markets.
One a day would be less than a 1% reduction in software patents. We need to shut them ALL down.
Also, like Quora, does the StackExchange system allow one to register one's fields of expertise, to have questions suggested to you? (I'm not a big participant on SE but I know it may be hiding that feature from me, because it slowly reveals features based on karma).
Anyway I'm sure this has been thought of before, just curious if it's on the todo list.
There is also an advanced "filters" functionality (visible at stackexchange.com/filters) which apparently lets you filter out all but certain tags, as well as have 15-minute/3-hour/daily emails with new questions in that tag. I have never used it, but it sounds like what you're aiming for.