[SiVal]

TL;DR: As far as I can tell, the news here is that Facebook is doing for React Native (& Yoga) what they recently did for React itself. In other words, in the past choosing React for crucial infrastructure put you in position where even if Facebook infringed on your IP, you could not sue them without losing the legal right to use React, potentially crippling your company. So, either use React and create an IP vulnerability with respect to Facebook or don't use React.

Facebook changed that licensing policy for React to a more common license. You could now sue Facebook over some IP issue without losing the right to use React. Facebook could still sue you, of course, but so could any other company. Using React no longer increased your vulnerability with respect to IP or lawsuits pertaining to Facebook.

But Facebook only removed the (claimed) vulnerability-causing license provision for React, leaving it in place for React Native.

And today, they are announcing its removal from React Native as well. Thank you, Facebook.

[Lazare]

> In other words, in the past choosing React for crucial infrastructure put you in position where even if Facebook infringed on your IP, you could not sue them without losing the legal right to use React

It hardly matters now, but as gets pointed out every time this gets brought up, this is quite false.

You would lose the legal right to use any patents which Facebook might (or might not!) have on React. You did not lose the seperate BSD license on the actual code with grants you your license to "use react".

> Using React no longer increased your vulnerability with respect to IP or lawsuits pertaining to Facebook.

That's highly debatable. If Facebook does have patents on React (and it is believed they do), then using React, or any technology that uses the same underlying technology does, in fact, increase your vulnerability to IP disputes and lawsuits from Facebook. :) The implicit patent grant in the MIT/BSD license will help, but so did the (now removed) explicit patent grant, so it's hardly clear this is a win on net.

The situation has never been as simple as people seem to wish it was. Software patents are an enormous and (so far) unsolved burden on our industry. At least if you're in countries where they are enforcable (and given the nature of the global legal system, probably even if you aren't).

[tomelders]

More importantly, you have more protection with the BSD+Patent Grant than with BSD alone.

With BSD alone, if you infringe on a Facebook patent then you are vulnerable to legal action from Facebook if they deem that you’ve infringed your patent. And regardless of wether you’d infringed or not, such action would still cost a lot of money to fight in court and would cripple most businesses.

People who claim it risks giving up their IP are ignoring the brutal truth that software patents are shit, and if you’re in the software business you’re pretty much guaranteed to be infringing someone elses bullshit patent.

The BSD+Patent Grant was designed to combat the patent trolls, but people lost their minds and now it’s gone.

The BSD licence is objectively worse than the BSD+Patent Grant. If you opposed it, then the only people you’ve helped are the patent trolls.

[jjnoakes]

> The BSD licence is objectively worse than the BSD+Patent Grant.

I'm not a lawyer, but I have heard a few claim the opposite: BSD plus the explicit restricted patent grant was worse than BSD (or MIT) with no explicit patent grant (because many believe it carries an implicit unrestricted patent grant).

[qaq]

It does not look like this was ever tested by actual court case so I guess it is hard to argue either way.

[bad_user]

Please leave such opinions to actual lawyers.

Apache placed React's BSD + PATENTS in "Category-X", banning it from being used in Apache projects:

https://react-etc.net/entry/apache-foundation-bans-use-of-fa...

> The BSD+Patent Grant was designed to combat the patent trolls, but people lost their minds and now it’s gone.

It was designed for Facebook to fight patent trolls, but not for you. In comparison with saner license like Apache 2.0, Facebook's license only applied to Facebook.

> The BSD licence is objectively worse than the BSD+Patent Grant.

That is false. Please read up on "implicit patents license".

[Lazare]

The issue the Apache Foundation had with the license was that it was incompatible with the Apache license; they weren't making a value judgement about it being good, bad, strong, or weak. Many other excellent licenses are also in "Category-X".

[jwatzman]

> Apache placed React's BSD + PATENTS in "Category-X", banning it from being used in Apache projects:

The GPL, LGPL, and AGPL are also in that list, so by itself this isn't much of an indictment of the license itself.

https://www.apache.org/legal/resolved.html#category-x

[tomelders]

Please leave such opinions to actual lawyers.

[cma]

> You would lose the legal right to use any patents which Facebook might (or might not!) have on React.

So a poker bluff that would be Russian roulette without an expensive comprehensive patent search (and researching their patents itself increases your liability; reading about all the ones that you later determine didn't cover React subjects you to triple damages if they covered any of your other tech).

Facebook could also have acquired existing patents at a later date from other companies and individuals and use them go after you if you triggered the revoke terms, so you had to search all patents ever, not just ones that Facebook filed or owned during your initial decision to use the library. You could argue you have to do that anyway, but keeping with the poker bluff analogy, Facebook could know of the patent already, have undisclosed terms that prevented React licensees from being sued by the holder, and have a pre-negotiated option to buy the rights during a licensee revoke event.

[dastbe]

So obviously IANAL, but aren't you eliding the fact that the explicit patent grant was revoked on suing facebook for patent infringement? Whereas the implicit patent grant in MIT lacks that caveat?

[Lazare]

The explicit grant was revoked in circumstances that are unlikely to matter for the overwhelming majority of users.

The implicit grant doesn't get revoked, but how it works is still unclear. This is true especially (I believe) when it comes to sub-licenseability and patent exhaustion, where the exact mechanics have yet to be worked out in the courts. If I download MIT licensed code, and then distribute it to you, you have a valid license to the copyright on that code, despite the fact that you didn't obtain it from the copyright holder. Do you also have a valid license to any patents on that code? I have no idea, but I know enough to know it's not as obvious an answer as you might guess.

In general, explicit is better than implicit, and settled legal questions are better than fascinating mysteries. These factors weigh in favour of an explicit patent grant. The terms of the explicit patent grant Facebook was using, granted, were quite negative, but on balance I think I (and my current employer) probably were marginally better off with the older license terms. YMMV. :)

[ftoo]

That last paragraph is the very definition of Stockholm Syndrome.

Facebook is an $0.5T company built on open source. Without open source, Facebook would not have been possible.

[btgeekboy]

You’re not wrong, but it seems more appropriate to applaud desired behavior (positive reinforcement) than to continue to be bitter about past undesirable behavior.

[sillysaurus3]

This is correct, for what it's worth.