The law patch that shuts down patent trolls will have no effect on software patents, and vice-versa.
Software is great because it has no cost to copy, so when someone creates a new algorithm it can be put in use everywhere applicable rapidly. This pace of advancement is so awesomely powerful that many do not appreciate how much positive change it can bring or how different the world was just a few years ago. An advancement in algorithms becomes and advancement for humanity pretty quickly.
Software patents are the exact opposite of this.
I will take it one step further and claim that patents on the whole are bad. I don't think they do what most Americans think they do in terms of incentivizing inventors or protecting small business.
EDIT - Also software is already covered by copyright, why does this IP deserve two kinds of protection and most other IP gets 1 or 0 kinds of protection.
An honest patent forces a description good enough to duplicate. Better to keep your secret sauce secret.
As long as they remain secret they have protection, but the very act of outing them makes them not secret. I am aware of very few successful suits defending trade secrets.
Trade secrets cases are lost when the owner is claiming something is a trade secret when it doesn't meet the legal definition of a trade secret.
Trade secrets are useful for long-lived secrets like the formula for Coca-cola. But if someone independently develops the same thing you are trying to keep secret too bad.
Also, if the secret accidentally gets out (w/o theft) the trade secret enters the public domain.
It probably doesn't matter at all.
The usual argument for this separation is that "software patents are just math", but then aren't hardware patents are just physics (which, ultimately, is also just math)?
It feels like there's no clear line here, and whatever benefits, if any, are to be derived from patents, they apply equally to either side of it. Basically, either it's a net benefit to society, and then both kinds are valid; or it's a net drain on society, and then we should just get rid of them altogether. Or maybe set the bar for what is patentable much higher wrt to novelty and value (but, again whether it's hardware or software feels irrelevant).
Some people feel that since there is no copying cost, software favors the little guy more than with physical goods, so normal patent protection doesn't provide an benefit. These people claim that a big big evil corporation could start making a million of their widget per month and out compete them, but with software this cannot happen per their arguments. I think these people acknowledge that both physical and software patents get abused and are imperfect but feel that physical patents provide enough gain to offset their downsides.
I disagree with those people completely. I am unaware of ANY actual evidence that patents help.
I am aware of actual evidence that patents help, if enabling small companies to withstand pressure from behemoth companies is considered help.
You copyright your source code. You patent your engineering innovations.
The ease of duplicating software based inventions is an argument why patents are important for such inventions.
Many tech startup companies that would have been steam rolled by IBM, and friends, have been protected by their patent portfolios.
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.
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.
Seaborg patented element 95. That's pretty broad. However, people aren't saying get rid of manufacturing utility patents because folks are patenting elements.
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.
If you protect trade secrets, companies will never release them to the public – as in the Waymo case.
Keeping something important as a trade secret has enough risk already.
https://www.google.com/patents/US8836922
That’s what should be provided with patents, as I said in the original comment.
Patents are about telling you how to do the work, not about doing it for you.
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.
Software patents are pointless and the only reason we still have them is it takes Congress forever to catch up with innovation. Patent trolls know this and are capitalizing on the period where they can shake down others while they still can.