I didn't stalk you: you include a link in your HN profile to your Github page, I only clicked on it. Consider removing the link if you don't want people visiting your Github profile, and learning what you engage in.
Many of your statements are imprecise/incorrect.
1. Patent grants are good. Many licenses, including ASLv2, MPL, CDDL, not only include a patent grant, but also a “weak patent retaliation” clause. A minor reprimand if you adopt a project and then decide to sue THAT VERY SAME PROJECT for patent infringement. That’s a pretty fair deal. Some licenses like CDDL are quite relaxed, e.g. they even give you a 60 days grace period after you initiate the litigation to migrate away from the software under dispute. Others like MPL not only remove the patent grant, but also the copyright license as well — which depending on the project could be assumable.
2. Facebook could’ve done the same (simply by adopting ASLv2, like they did with RocksDB when they wanted to allow Apache Cassandra to use it), but they decided to introduce a “strong patent retaliation” clause, where they protect all of their patents by immediately revoking all current (hypothetical) or future patent grants of ALL SOFTWARE with that license (which also includes Jest, immutable.js, draft.js, Flow, Infer, etc.).
3. As I said in my second article, aside from Palantir, no other corporation has adopted Facebook’s “BSD-3 + strong patents retaliation clause” since 2014 (when it was first released) – based on my examination, which I linked to in my previous comment. Meanwhile, FB claims wide adoption of this license would decrease “meritless patent litigation”. If it were true, companies should be running towards this license model. So how do you explain that no one else (aside from Palantir) has adopted it, over the course of 3 years? Something doesn’t add up. I’m sure they will explain at some point.
4. You do seem to complain a lot about my style. But with regards to content, aside from you trying to lecture me in several areas where you make mistakes yourself, your main point seems to be that the scenarios I paint are conjecture. Well, yes. They are conjecture. You see: people need to imagine possible outcomes in order to take decisions and act accordingly. If it had already happened, it wouldn’t be a blog post, it would be news.
5. Yes, I am an ASF contributor. I use GNU, ASLv2, MIT, MPL, CDDL-licensed software, and that also makes me versed (not an expert, of course) in the licensing field. I am impartial to licenses as long as they work for me and the purpose of the software I'm building. Is there anything wrong with that? Please stop making camps in your head. There is no React vs. ASF conspiracy nor battle taking place. Everyone is entitled to have their own opinion, and to study the consequences from their own angles.
6. Large companies who have adopted React, like Amazon, Microsoft, Airbnb, etc. have enough manpower to migrate away if they need to file a patent infringement lawsuit against Facebook. For them, adopting React could be a good deal. Startups don’t tend to enjoy that amount of manpower; I have already answered this several times.
I could be wrong in many of my statements, I acknowledge that. But the truth is that the license creates legal provisions whereby the scenarios I envisioned would be possible, even if several conditions would have to align. Hence, what is left? Trust that FB won’t go after you? Goodwill statements from FB’s team? Sorry, mister, but those are worthless. Especially when there are legal terms mediating.
We could keep going on forever, but my time is finite so I’m going to close this discussion with you here. Good luck!
Edit 1: changed the order of some points for better flow. Edit 2: minor changes in PS.
P.S.: BTW - you will find that OSS developers who are affiliated to some Foundation (Mozilla, Apache, Eclipse, etc.) tend to be versed in licensing, without needing to be lawyers. Everybody can understand the terms. You only have to be a lawyer if you're going to enforce them.
> Patent grants are good.
So you're not upset with Facebook providing a patent grant but with the terms of that patent grant, fair enough.
> Facebook could’ve done the same
The projects using this patent grant are almost 100% sponsored by Facebook. They're not tied to any particular ecosystem or using Facebook services. If Facebook makes the use of the relevant patents contingent on you not suing them, I'm having a hard time feeling upset about it.
Yes, this isn't as permissive as Apache2 but "you get to use all my stuff as long as you don't sue me for stealing your ideas" sounds fairly benign.
> If it were true, companies should be running towards this license model.
This is literally an argument from popularity. Despite all appearances, IT can be mindnumbingly slow to adopt new ideas, especially ones involving politics. Most people don't care about patents -- heck, as can be seen from 90% of the people commenting on React's patent grant, most people don't understand patents.
In the JS ecosystem MIT and BSD are among the most popular licenses. The reason they're popular is that they're short and don't contain much legalese. Does that make them good licenses? Hell no, but it makes them appear easy to understand and simple. Adding a patent grant to that makes it "not simple". It's also something you have to do intentionally.
But that's beside the point. Let's just assume Facebook is wrong and having this be the standard wouldn't bring about world peace and the death of patents or what have you. Let's go with that. How does your assessment that this is evil and bad follow from that?
> Well, yes. They are conjecture.
Thank you. Then stop wrapping them as quality advice to trick people into reading your opinion.
> Please stop making camps in your head. There is no React vs. ASF conspiracy nor battle taking place.
You tried to dismiss my opinion by portraying me as a React fanboy. I pointed out why that is bullshit and you're throwing bricks in a glass house. Now you're accusing me of being the one who tries to make this about camps?
> Startups don’t tend to enjoy that amount of manpower; I have already answered this several times.
Your entire point is that startups shouldn't use React because bigcos might acquire them. Now you're literally arguing that bigcos don't have to worry about migrating away from React. Doesn't that seem entirely contradictory?
Sure, you could argue that bigcos will still factor in this cost into acquisitions but if they already pay that cost for their own projects it can't be so prohibitive to result in bad deals. And how do you rationalise Microsoft actively encouraging outside developers to use React to build on their platform? Surely this would make any resulting acquisition unnecessarily costly.
> Hence, what is left? Trust that FB won’t go after you?
Are you worried Facebook will go after you? Are you worried Google will go after you? Are you worried Apple will go after you? Software patents are already a minefield. You're far more likely to be sued by a non-operating entity.
A patent lawsuit can easily kill almost any startup. You seem to be obsessing over how many nukes you're going to be hit with while operating at a scale where being caught in the blast radius of even a single one would end you.
Oh, wait, this is completely irrelevant. The patent grant does not terminate when you get sued. It doesn't even terminate if you countersue. The patent grant ONLY terminates if you proactively sue Facebook or sue (or countersue) any user of React over React (including Facebook). If Facebook sues you and you countersue Facebook over your smart fridge patent, you still get to use React.
> but my time is finite
So is mine. But protecting people from unsubstantiated FUD is worth it.
> you will find that OSS developers who are affiliated to some Foundation (Mozilla, Apache, Eclipse, etc.) tend to be versed in licensing, without needing to be lawyers
Good, as every developer should be. But if your mischaracterization of how the patent grant can terminate isn't based on a lack of comprehension or absentmindedness, what am I to read into that?
PS: I meant what I said. You should have written an article entirely about the issues you have with the BSD+Patents issue from an aspect of open source purity. Without the conjecture and clickbait angle to it. I would have read that, too. I would probably have disagreed with some of it even if I would have upvoted it. But at least it would have provided the grounds for an honest discussion.
The argument is valid to be had. There's no point to dressing it up as if it were about startups.
You're not an idiot. You know very well that your article won't convince startup developers to drop React, it will only influence non-technical people who are risk-averse and see your headline without understanding any of it.
Instead of playing this game and claiming you "don't want to start a flame war" after writing an obviously inflammatory article, please salvage those ideas you buried in it and write that article properly.
No, your article is literally FUD. "I don't know what their goals are with this, so let's assume the worst". What could that be? Nobody knows, so let's assume it's really bad and you don't want it. F, U and D.
> I analysed 75+ projects from 35 companies (link: [1]), and none uses this license model. What's going on?
You're feigning surprise that no other company than Facebook provides a Facebook Open Source Patent Grant along with their BSD-licensed projects?
The only major licenses that contain patent grants are Apache 2 and GPL. As I'm not going to touch the GPL debate with a ten foot pole, that only leaves Apache 2 or no patent grant whatsoever. Turns out most projects just don't bother with patent grants.
There are good arguments for Apache 2. BSD+Patents seems to be a good compromise if you want a patent grant, but not one as unconditional as Apache's. This is not nefarious. This is literally more than the nothing 99.999% of non-GPL, non-Apache projects give you.
The only thing the PATENTS file does is remove the Heisenberg-esque "maybe a court finds this license implies some kind of patent grant or not" limbo the BSD would otherwise put the project in and replace it with a fixed set of legal certainties you can give your informed consent to or not.
> As I argue in my article, this kind of license may hurt the Open Source industry, more than it will benefit it.
I thought the point of your article was that startups shouldn't use it if they want to be acquired, oh wait, that's just the clickbait title to sell your opinion on.
Your entire "it's bad for OSS" argument boils down to a slippery slope of chilling effects. This is entirely conjecture. By your logic Microsoft would never have published ReactXP and React Native Windows. Google, Twitter, Amazon are actively using React in their products. Your argument falls flat if you even bother looking at the reality.
You defend your article by saying you don't want to be involved in flamewars because you're not a frontend developer. You're using intentionally provocative language to make a point about something you admittedly have no stake in.
Your article is trying to argue about open source purity but abuses a clickbait premise to garner attention in places like HN.
> So just stay away of problematic stuff.
Again, this shows you don't understand patents. ALL software is "problematic stuff". BSD+Patents, like Apache2 or GPL, provides certainty about a subset of patents. All other licenses provide only a vague sense of "probably the author won't sue me because that kind of fishing seems obviously unethical".
If you want to write an article about why React should be using Apache 2, write that fucking article. Don't try to cover it in startup-relevant buzzwords.
> 4) And as a result of the above, my stance is that I prefer to use a stack that grants me any patents unconditionally, or maybe with weak patent retaliation like the ASLv2 license, that's OK, i.e. I sue the holder over patents covering only the work I am using, I lose the patent grant for the work I'm using. Simple and symmetrical
Yes, you're an ASF contributor. Maybe you should lead with that instead of pretending you're impartial.
The world would be a better place if everyone used Apache 2. But what does that have to do with startup valuation? Stop spreading FUD.
> Would you rather relinquish your IP assertion rights with this "BSD-3 + strong patent retaliation" model, or would you rather pay $99/dev for a commercial license on React and not give Facebook any control?
Let me repeat what I just said: you don't understand patents. React works because it's open source. If React were closed source, it would be about as relevant as Sencha Ext. Not very. Try forking Preact/Inferno with a $99 "I bite any patent lawsuits levelled against you over this" license and see what happens. People don't care about patents.
> Would he have used LAMP if it entailed relinquishing any rights to initiate patent litigation?
Probably yes? He created Facebook as a one-off stupid thing in college. He might have migrated to something else afterwards. He might have open sourced that. Who knows? Who cares?
Again, none of this seems relevant to the headline: React supposedly being bad for startups because it hampers acquisition. You address this in a single section and the entire argument is a single pie-in-the-sky fictional scenario. None of your "empirical" claims back this argument up.
> React fan
Rich, coming from an ASF fan. Based on your logic I should assume you don't like BSD+Patents because it's incompatible with the ASF's purity requirements.
If you spent more time internet stalking me you would have noticed most of my work goes to Apache 2 licensed projects, not React. But that would not make a very good _ad hominem_.
I would love it if React switched to Apache 2. I would also love to see a lot of MIT/BSD projects switch to Apache 2. If it doesn't happen, I'm not losing sleep over it and neither should you.