4 ms·
Software patents should have the same requirements as real patents: you have to provide the full knowledge required to create the concept that is patented in th
by kuschku 9y ago
Software patents should have the same requirements as real patents: you have to provide the full knowledge required to create the concept that is patented in them.
You want to patent sticky scrolling? Then put the entire source code required for it into the patent.
You want to patent your AI tech? Then put the entire code into the patent.
Additionally, we should make sure that trade secrets are not protected in any way, and decompiling code to find out what it does, and replicate it, is legal, as long as it doesn’t violate patents.
These measures would provide reasonable measures for companies wanting to protect their secret sauce, but also ensure that this technology is all public and available for research, and that after a few years it becomes available to competitors.
Otherwise we’ll end up with Google, Amazon, etc piling up knowledge and patents, without ever giving anything back to society.
And if you now try to argue with Google being "good", please take a look at ReCaptcha, where they take captcha associations given by users, but never publish the resulting recognition data (which then could be used by open source projects or researchers), or look at Android, where Google has now closed everything they could.
- CalChris 9y ago> Then put the entire source code required for it into the patent Some do. But patents already have an enablement requirement. This means that they must teach in sufficient detail for a person skilled in the art to carry out the claimed invention. That's what the whole Detailed description of the invention section is. Skip the claims and read the description. You might want the code; me, I want the design doc. That section should be clear and if it isn't, chances are it isn't enabling. If you don't enable, you'll probably get an office action (rejection) and if you don't (lazy examiner), you'll definitely lose when the patent gets challenged.
- woah 9y agoThat's a nice thought, but most software patents are written to be as broadly applicable as possible, not to be a good guide to implementation.
- CalChris 9y agoIn fact, most patents are written to be as broadly applicable as possible. The enablement requirement applies to all patents. Seaborg patented element 95. That's pretty broad. However, people aren't saying get rid of manufacturing utility patents because folks are patenting elements. https://www.google.com/patents/US3156523 https://www.google.com/patents/US3156523
- curun1r 9y ago> we should make sure that trade secrets are not protected in any way That's a little extreme. Trade secrets still need protection from outright theft (e.g. Waymo's allegations against Lewandowski) by those that get them as a part of their job. Similarly, trade secrets acquired through any other illegal act (breaking and entering, hacking, corporate espionage, etc) should also be protected. I agree that trade secrets arrived at through reverse engineering should be okay, but that's very different from saying that they should receive protection in any way.
- kuschku 9y agoI disagree with that. Because then companies just keep new developments as trade secrets – I want to push them to patenting everything, which would require them to open up the entire technology to the public a few years later. If you protect trade secrets, companies will never release them to the public – as in the Waymo case.
- curun1r 9y agoIf you don't patent, you lose the ability to go after others if they independently create the same invention or reverse engineer your product. And if you fail to properly document your use and someone else patents it, you could lose the rights to your invention altogether. Keeping something important as a trade secret has enough risk already.
- kuschku 9y agoYet, the waymo case exists, and waymo hasn’t patented or published anything, and probably won’t ever publish it.
- curun1r 9y agoThat's false. The Waymo-Uber case contains patent claims, though 3 of the 4 were dropped at the judge's suggestion. So it seems that Google has patented quite a few aspects of their self-driving vehicles. The article I found didn't list the patent they're still pursuing against Uber, but here's the 3 patents Google decided not to pursue: https://www.google.com/patents/US8836922 https://www.google.com/patents/US8836922 https://www.google.com/patents/US9285464 https://www.google.com/patents/US9285464 https://www.google.com/patents/US9086273 https://www.google.com/patents/US9086273
- pnw_hazor 9y agoAll US patents have the same standard. Among other things, they must enable one of ordinary skill in the art to practice the claimed invention without undue experimentation. For SW patents, if your invention uses machine learning but is not inventing a new kind of machine learning, there is no need to describe how the ML package works. It is a black box. Similarly, a gadget patent does not need to describe the metallurgy necessary to make aluminum sheet metal or how to make plastic. The same standard applies to biotech patents. Though, some biotech patents require proof or disclosure of experimental results (e.g., gene sequence listings, etc.) this is because some bio inventions would be very difficult to reproduce without the extra information. Trade secrets are not protected from independent discovery. Reverse engineering is generally fine, though it may violate your license or the DCMA depending on the details.