5 ms·
Ask HN: how can we establish prior art to fend off future patents?
As part of building a spreadsheet, we've come up with a pretty neat algorithm to run a recalculation in parallel, taking advantage of multiple CPU cores etc. We're don't like software patents, but have had some experience with the patent process in the past. Looking at the algorithm, our gut feeling is that clever enough to be patentable but not so clever that we don't think other people aren't quite likely to come up with similar ideas should they put their minds to the same problem.
If we don't patent it, our worry is that someone else will independently come up with the idea, patent it, discover us, then sue us. The best defence against this that we can see is to make sure that we can prove that we had the idea before they patented it -- that is, to be able to establish priority. What we're wondering is, what's the best way to do that?
We've no worries about making the algorithm public -- it's pretty clever, but not that clever. But proving the date on which we made it public might be hard.
Any ideas?
[EDITED TO ADD] It occurs to me that there's another question here -- do we need to prove prior art, or just priority? That is, do we need to show that the algorithm was already public and known at the point that someone else tried to patent it, or do we just need to show that we had created it (even if we'd not shared it with anyone).
I guess being able to prove that we'd published the algoritm and thus it was prior art is a subset of the case where we can prove that we had thought up the algorithm, so making it public is safer.
- mryan 16y agoPost it on Github? Also, if emails are admissible in court in your country, you could email a description and copy of the algorithm to a company that provides "email notary" services.
- gpjt 16y agoYes, that could work, though of course we might then have to prove that GitHub's records of when the repo was created/populated were correct -- I like mindcrime's suggestion (peer of yours) of combining that with a PR blitz, though. Re: a notary -- we're in the UK, and I believe email can be admissible in court, but its admissibility can be contested on what amount to common-sense grounds -- for example, I doubt we'd be able to use a record of an email from our own server to prove we'd sent a message with a given content on a given day unless we had a lot of backup evidence. Interesting point about an email notary -- perhaps we could even use a regular one and get a statement of the algorithm printed on dead trees notarised? I guess the question then has to be, do we just have to prove that we had developed the algorithm before the other person had the idea, or do we have to prove that it was already public?
- mryan 16y agoInterestingly, the dead tree option did not occur to me for a moment. Once it is printed, the problem almost goes away - you can do the publicity blitz, but also lodge the copy with the notary. Wrt your second question, I am not sure - I think the latter. For something to be considered "prior art" I believe it has to be publicised.
- mindcrime 16y agoSubmit it to a journal for publication? Present it a a conference of some sort? If you can't get it published in a traditional journal, write it up and submit to arXiv? Publish it on a website where it will get crawled by the Wayback Machine? Write it up and see what it takes to get it indexed by CiteseerX? Even better, try to turn publishing it into an opportunity to gain some PR: send press releases far and wide about it, and put an implementation of it in a GitHub repo. Shout it to the world. IANAL, YMMV, HTH, WTFBBQ, etc...
- gpjt 16y agoHey, thanks! I like the last idea -- I'm sure a PRWeb release ($200 or so) would get picked up by enough random press-release-republishing semi-splogs to make the publication date pretty easy to prove. And a GitHub repo would give us another proof-point.
- prodigal_erik 16y agoRMS once pointed out that prior art is more effective if it doesn't surface until a court case: http://www.linux.com/archive/feed/57167 http://www.linux.com/archive/feed/57167 Basically, if a examiner knew about the other work but granted the patent anyway, the court must presume the examiner was competent and the patent is valid, and therefore that the other work isn't relevant prior art. But if potential prior art wasn't known to the examiner, the court has to actually evaluate it. Judges don't share patent examiners' quota-driven incentive to rubber-stamp everything under threat of their jobs.
- gpjt 16y agoThat interesting, thanks -- I didn't know that.
- sagacity 16y agoI too am not a lawyer but another possible way to do this, to the best of my knowledge: File a provisional patent application (with maximum disclosure) with the USPTO - this will give you a priority date on official, government record. Make sure you obtain a certified copy from them. Then let it lapse (i.e. don't convert to a full application within the stipulated time period of one year). I guess that should do it. Disclaimer (again): I'm not a lawyer :-) Edit: Added: I don't know how effective/legally valid this is today, but there also used to be something called a sealed cover method. You disclose the full invention on paper, get a government approved notary to notarise it, seal it up in an envelop and send it as a registered mail to yourself. On receipt, don't open it - just keep it safe. In case of any later dispute, you'll have two government agencies establishing your date of invention. - HTH 2nd Edit: Added: The sealed cover method, if valid, would probably be the least expensive one. :-)
- prodigal_erik 16y agoHas the "mail to yourself" thing been done successfully? I have trouble believing a million-dollar court case can be decided upon the tamper-resistance of a cheap paper envelope. And all prior art must be published somewhere in order to have any effect on patentability, so the envelope could only establish a priority date.
- Shooter 16y agoIt is most definitely NOT a valid method.
- sagacity 16y ago> so the envelope could only establish a priority date. As I understand, this is exactly what the OP is after. :-) If OP has a provable priority date, a future patent holder can't sue the OP. Or can he still? I guess only a patent attorney can tell.
- gpjt 16y agoThanks for the ideas! Re: posting -- I guess I would worry that the "attacker" would say that although the letter proved that I posted myself something on a given date, if the package was not tamperproof then it could have been anything. OTOH the idea of filing a provisional patent sounds very plausible, albeit perhaps expensive.
- PaloAlto101 16y agoThere's a great ebook, Patent Pro Se, that gives practical advice on protecting you IP, and it is now free to download...goes over everything from Provisional Patent apps to the Mail-yourself-a-letter-method. May be a good resource for you http://www.carrferrell.com/free-resources.php http://www.carrferrell.com/free-resources.php.
- gpjt 16y agoCool, sounds useful!