News for hackers. Stuff that matters.
Re:Explaining React's license by Scaevolus on Saturday August 19, 2017 @02:19AM (#15051972)
Almost all of Facebook's "open source" code has the same PATENTS file, which makes them toxic if you can imagine a future where Facebook infringes one of your patents: https://github.com/search?p=5&q=org%3Afacebook+filename%3APA...

This includes things far outside the React ecosystem:

    - Flow (JS type checker, like TypeScript)
    - prepack (JS optimizing transpiler)
    - a bunch of Android/iOS UI/debugging frameworks,
    - all their GraphQL libraries (server/client)
    - their machine learning work (mostly targeting Torch)
    - Reason (statically typed programming language that transpiles to JS/OCaml)
    - ZSTD, a highly competitive compression algorithm.
Disclaimer: I work at Google on Kubernetes (which is Apache licensed). This is my personal opinion. I am frustrated with software with trivially incompatible licenses.
Re:Explaining React's license by IanKelling on Saturday August 19, 2017 @05:09AM (#15052461)
> I am frustrated with software with trivially incompatible licenses.

It's not a copyright license, and it's not incompatible with any other free software license. I'm frustrated with the misinformation spread about it.

If this patent license/grant get's revoked, you are back to simply using the BSD license with no patent grant. I've read so many people say "you'd have to stop using react if you sued facebook", uh, no, you'd have a bsd license with no patent grant like you probably do with tons of other free software your company uses. Clearly, people should be complaining about that if they are complaining about this, but the misunderstanding and misinformation is really strong. If you believe software patents are universally bad, like many people including me, then it is clearly better using the MIT/BSD license alone, which gives you zero patent rights, you are simply infringing and waiting to be sued. I have no problem with it. https://www.gnu.org/philosophy/software-patents.en.html.

Re:Explaining React's license by jnbiche on Saturday August 19, 2017 @08:43AM (#15053048)
Your explanation really doesn't jibe with the explanation given in Facebook's post. If what you say is true, why do they make a point about this patent grant protecting them against patent trolls? There's something tricky about this that I can't figure out, but I really have a hard-time believing they're acting in good faith here.
Re:Explaining React's license by vonmoltke on Saturday August 19, 2017 @11:04AM (#15053763)
As cited at multiple other places in this thread, Facebook's own FAQ on the licensing explicitly states what GP just did.
Re:Explaining React's license by jnbiche on Saturday August 19, 2017 @11:55AM (#15054039)
If we use the "minimalist" interpretation like you and others propose here, then I'm stumped by Facebook would bother. It's such a tiny, ultra-specific advantage for Facebook. Why throw this gas on the fire for such a minimal advantage? Just leave it alone.

I suspect this must be coming from Zuckerberg. He's pissed off that some patent troll, somewhere, was using Facebook open source, and he issued an edict. That's the only way this makes any sense at all.

Re:Explaining React's license by Scaevolus on Saturday August 19, 2017 @05:27AM (#15052510)
A patent grant is a license.

The patent license is indeed independent of the copyright license, but it's troubling enough that multiple legal teams (Apache and Google, at least) don't want to entangle themselves with it.

I'm doubtful that any licensing scheme can significantly drive patent reform. Lobbying and legal precedents hold that capability.

Re:Explaining React's license by IanKelling on Saturday August 19, 2017 @05:36AM (#15052539)
> multiple legal teams (Apache and Google, at least) don't want to entangle themselves with it.

Citation needed on Google. Apache isn't "not wanting to be entangled", they reject EVERY license which gives stronger protections for user freedoms than apache license alone, like the gpl, lgpl, cddl, mpl, etc, etc, etc. Google has no problem using the gpl for linux and lots of licenses and projects that apache would reject, so you can't just lump them together like it's the same thing.

Re:Explaining React's license by sanxiyn on Saturday August 19, 2017 @06:02AM (#15052597)
DannyBee is on this thread. That is good enough citation for me.
Re:Explaining React's license by IanKelling on Saturday August 19, 2017 @04:31PM (#15055428)
Fair enough. And my point about why Apache avoids stronger licenses actually applies to google too, but they are just more flexible about it on a case by case basis.
Re:Explaining React's license by Teknoman117 on Saturday August 19, 2017 @03:57PM (#15055256)
Just so I understand, as I've never really mulled on how software patents mix with open source licenses, when you say "you'd have a bsd license with no patent grant", who are you waiting to be sued by? Facebook and/or others? Could you theoretically release an open source project under something like BSD that infringes on a patent you hold and then sue someone for patent infringement that uses your project?

I suppose mainly what I'm asking is that by having this patents clause, is Facebook asserting that some or all contents of react are patented by Facebook? Would they be able to sue you for using react if you did something to get your patent grant revoked?

Re:Explaining React's license by sanxiyn on Saturday August 19, 2017 @03:50AM (#15052235)
I am very happy to report that ZSTD removed PATENTS clause: https://github.com/facebook/zstd/pull/801
Re:Explaining React's license by hobofan on Saturday August 19, 2017 @05:02AM (#15052438)
RocksDB too.

That really makes it seem like Facebooks stance is: "As long as we have a few really huge projects that infiltrate most companies' codebases, our 'smaller' projects can drop the patent clause".

Re:Explaining React's license by jnbiche on Saturday August 19, 2017 @08:45AM (#15053055)
What the hell? None of this makes any sense. What is Facebook up too?
Re:Explaining React's license by rockdoe on Saturday August 19, 2017 @04:31AM (#15052344)
This is huge.
Re:Explaining React's license by markyuckerberg on Saturday August 19, 2017 @05:56AM (#15052581)
Welcome to the "folks who miss the forest for the trees" club.
Re:Explaining React's license by coldtea on Saturday August 19, 2017 @06:36AM (#15052678)
Welcome to the over-reaction party.

3 years from now, most startups and companies will still be using React as they did for the last 5 years that the same terms held.

Re:Explaining React's license by markyuckerberg on Saturday August 19, 2017 @09:35AM (#15053259)
>> Welcome to the over-reaction party

Hmmm.. in that case the people who actually need this advice even more would be the maintainers of ZSTD who seem to have hastily changed the license after seeing the backlash from the Apache policy change. Care to stand by this opinion and leave your comment on the thread which discusses the license change over on GitHub?

https://github.com/facebook/zstd/issues/775

Re:Explaining React's license by diakritikal on Saturday August 19, 2017 @05:26AM (#15052509)
What's the difference between say the PATENTS file/grant in these Facebook projects and the PATENTS file/grant in Go which Kubernetes is written in?

Not being inflammatory, genuinely interested as a user of both Go and React...

Re:Explaining React's license by Scaevolus on Saturday August 19, 2017 @05:35AM (#15052534)
Not a lawyer, but as I understand it:

The Go and Kubernetes patent licenses terminate if you file litigation specifically regarding those projects-- if you sue anyone for violating some garbage collector patent in Go, you lose the patent licenses Go granted.

Facebook's patent license terminates if you sue Facebook, subsidiaries, or "any party relating to the Software" for any patent infringement. It doesn't have to be related to React.

Re:Explaining React's license by diakritikal on Saturday August 19, 2017 @07:05AM (#15052743)
Thanks, very interesting difference.

Google's grant protects (or tries to protect) specific tech, Facebook's grant protects (or tries to protect) as an umbrella, the organisation.

Well - can't comment on how effective it will be as a disincentive to sue, but I do feel at the same time if it does succeed in that goal it will almost certainly act equally as a negative factor in adoption; because it's my opinion that this is creating a soft-walled garden around Facebook open source tech.