AWS forked my project and launched it as its own service

Just don't repeat Microsofts smooth talk, make-your-own and ghost strategy ;-)

It is possibly even worse than just forking.

I know its not required by licensing - but "legal" doesn't always mean "right".

Probably Amazon's legal department doesn't let him say much more, but then his statement sounds unconvincing and doesn't serve the purpose of taking responsibility and assuring the comunity of their good intentions.

I always find these messages weird, because in the end it's one engineer like you and I who looked at some open sourced stuff and decided to use it, and perhaps it didn't really do what they wanted so they forked it, and it ended up being used in whatever product they were working on, and in the grand scheme of things it was not about a big trillion dollar company being evil, it was about how engineers do their work nowadays.

"At Serge’s trial Kevin Marino, his lawyer, flashed two pages of computer code: the original, with its open-source license on top, and a replica, with the open-source license stripped off and replaced by the Goldman Sachs license." [1]

[1] https://www.vanityfair.com/news/2013/09/michael-lewis-goldma...

If it's not in black and white then it's not part of the license. Spirit isn't defined.

To paraphrase Theo de Raadt, if you're not happy for your code to be used in a puppy mulching machine then don't license it under a permissive license.

I don't understand. Wasn't the project released under a license that explicitly grants anyone the right to freely use it as they see fit?

https://www.linuxfoundation.org/membership/members/

This is something we are very aware of, and discussions about this cross my desk weekly.

One big problem of the many problems I face regarding such proposals is the people who are the best at writing open source who need the money the most are not the people who are good at writing grant proposals and are good at sucking the money out of such funding systems.

Email me at atwoodm@amazon.com, and tell me what the internal code names for your projects were, and were you were in the org. I'll see what I can do.

> 3. All advertising materials mentioning features or use of this software must display the following acknowledgement: This product includes software developed by the <organization>.

There's a good reason why we no longer use BSD-4 anymore.

https://www.gnu.org/licenses/bsd.html

> The result is a plethora of licenses, requiring a plethora of different sentences. When people put many such programs together in an operating system, the result is a serious problem. Imagine if a software system required 75 different sentences, each one naming a different author or group of authors. To advertise that, you would need a full-page ad. This might seem like extrapolation ad absurdum, but it is actual fact. In a 1997 version of NetBSD, I counted 75 of these sentences. (Fortunately NetBSD has decided to stop adding them, and to remove those it could.)

Or are you talking about how if you download the Chrome extension and extract it you can see him referenced in NOTICES.txt? https://twitter.com/maxibanki/status/1317071448322789376

Thanks for informing us with the clarification. The original tweet was 11:16 UTC. This HN thread was submitted 11:23 UTC. That twitter reply showing the acknowledgement in "NOTICES.txt" was later at 11:54 UTC.

https://twitter.com/maxibanki/status/1317071448322789376

I don't usually read through files like ~/.config/chromium/Default/Extensions/bhdnlmmgiplmbcdmkkdfplenecpegfno/0.0.1_0/NOTICE.txt (though perhaps I should).

Obviously there is no legal requirement, but would it be that hard for Amazon to include a "forked from..." or "built off of...", etc. to the announcement and product pages, if it really is heavily based off of another work?

It seems to me it would just require mentioning it in some CREDITS.txt or whatever, which the other lincenses also do.

Somewhat relevant: https://opensource.google/docs/thirdparty/licenses/#wtfpl-no...

> He's not complaining that Amazon forked his code with Apache license.

The poster was indicating what part of Amazon's behavior he wasn't complaining about, not asserting that he wasn't complaining at all.

> What did AWS do wrong here? Were they supposed to know this guy's unwritten expectations?

I suspect you're either autistic or a lawyer being obstinate. Human society is full of unwritten expectations, we learn these quickly as a child or face social consequences. No where in the law is it written that you must say 'please' and 'thank you' but it's also expected and people are less likely to do things for you again if you don't.

So consider this situation now:

_A person (the dev) did something nice for someone else (a trillion dollar company) and they didn't bother to say thank you._

It is like fair use: They guy that uses google drive to backup youtube in its entirety is not doing anything illegal. He just demonstrates that he cannot deal with freedom.

There's a number of things that are strictly speaking legal, but still considered rude. Often, the reputation of a person or business is based at least in part on whether they do the legal bare minimum, or if they hold themselves to some level of higher standard.

I also think there's a difference between attribution because a license requires it (commonly buried several links/pages deep in some obscure "Here's a laundry list of ALL the open source packages we used to build this"), and acknowledging that a _specific_ library powers the core of a new product. I don't know of any license that marks that line.

> It seems like you're arguing that people should get both the good will of having an open source license and the profits of a closed source license. Which comes across as disingenuous.

It isn't. Linus Torvalds and all kinds of other people are paid to work on the Linux kernel, sort of like how the op suggests, and if that hadn't happened Linux would probably be a shadow of what it is now.

The alternative is to have some guy slave away for another's profit, and eventually burn out (which happens to tons of open source developers).

A specific example that comes to mind, one of the main core developers of Webpack (Sean) got hired by Microsoft to essentially just work on Webpack on MSFT paycheck (I think now he branched out a tiny bit into other adjacent areas as well, but he is still one of the main Webpack contributors). I think it worked out great partially because he was not the sole creator, but just one of the few, and he was the only one who went to MSFT, so the other core devs still had some level of control they could exercise on their own without anyone being able to tell them otherwise.

Interesting idea for a license.

Everyone can use and modify it, but if a large company with a market cap above $100m or a company wholly funded or owned by such company decides to utilize this project as a for-profit service, then said large company must hire me at for no less than $175,000 in 2020 value.

Sure the author can't require payment after the fact, but there are no open source licenses that prevent the author from soliciting payment from a downstream user. This is all fair game.

>You seem to want all of the good of open source, with none of the downsides.

If you do it right there aren't downsides. As far as business is concerned, the point with choosing any license is to create a win-win situation for all parties involved.

Unlike other types of work, this isn't necessarily true elsewhere (eg in the US & UK), but I understand moral rights originate from France and German and are especially strong there.

(I see that page now mentions "Credits: CloudWatch Synthetics Recorder is based on the Headless recorder. " - is that new? )

I love how these reckonings and syntheses are coming.

That is special. That is wonderful. It is a place for idealism. And honestly, it's worth getting angry about when people (and companies) don't respect it and improve upon it.

The money changers have fled from their high seats in the temple of our civilization. We may now restore that temple to the ancient truths. The measure of the restoration lies in the extent to which we apply social values more noble than mere monetary profit.

Happiness lies not in the mere possession of money; it lies in the joy of achievement, in the thrill of creative effort. The joy and moral stimulation of work no longer must be forgotten in the mad chase of evanescent profits." [1]

[1] https://avalon.law.yale.edu/20th_century/froos1.asp

We have not seen such an egregious abuse of open source software with any other company. Most companies that make use of a large amount of opensource actually contribute a fair amount back.

Like hiring on the core developers or making a large amount of code fixes and feature development contributions.

Amazon does neither of these things.

Take a look at any highly successful society or community. There is a large amount of gifting and selfless behavior

When companies couldn't find free software to use, they just wrote it themselves, typically, unless it was something big and way outside their domain.

Rather than being beholden to a licensing agreement, it would be nice if OSS had a license for an "enterprise-level donation" that was mandatory for for-profit use. It would be a one-time cost, so it would be easier to push through the accountants at lots of big companies, and companies could feel free to use the software at their leisure. Authors/contributors could choose to charge another fee for upgrading to a new major version, opt to end support of an old version or not and so on and so forth.

AWS has appeared to be doing a better job recently (from what I can see) in that regard. It's all around good PR. They lose nothing by thanking the maintainer/community for the work, even though the license doesn't require this. On the other-hand they build good-will.

Given the AWS response linked above, it would appear AWS recognizes this, and so maybe they will better accredit the work.

Sure, if attribution is a requirement then the natural thing to do is to turn it into a legal requirement. But I don't think that is the discussion here.

It comes down to how we want to treat open source. In order to encourage open source, I believe giving credit, even if not required, is courteous. Corporations are not monolithic entities that are perfectly defined. People work on these corporations.

Every contract, every law states what we see as the bare minimum required not to be actively harmful. They're society's skeleton. But bodies are more than bone, and societies are more than people doing the bare legal minimum.

Look at what the law requires of parents, for example. Food, shelter, clothing, school attendance, a lack of physical abuse. But parents who do the legal minimum and no more are awful parents, and awful people. But more laws wouldn't help. What kind of law could guarantee love? What kind of police could enforce it?

Community spirit is not something that can be expressed in a contract. Acting like people should have foreseen a particular asshole and tried to defend against them contractually is victim-blaming. The actual solution is for assholes to hear from the community that the behavior isn't welcome.

Some of us think copyright is unfair and we want to use it as little as possible. That means using MIT or BSD licenses.

That doesn't mean we are against attribution. We are only against the use of coercion to get attribution.

We can say "It would be nice if you give attribution" without saying "I'm going to use my legal rights to coerce you into giving me attribution"

In the same way that we must continue to steer our car even in the presence of guard rails, we must continue to act morally in the presence of rules.

I'm beginning to question this. The proliferation and commoditization of F/OSS is what made SaaS business thrive, and made it so that integration and polish is the only avenue left to make a buck, leading to our paltry attention economy, oligopoly, and platform lock-in by network effects. This after decades of personal computing striving to liberate users from mainframes. F/OSS is also drying out - when was the last time you used a piece of software that truly achieved something useful on its own rather than solving a perceived problem that only exists because of the idiosyncratic nature of the web and cloud stacks? Meanwhile, maintainers of popular F/OSS get nothing in return.

Changing incentives, standards and cultural norms absolutely is possible within a capitalist system. In fact, it's required. Otherwise, capitalist economies quickly descend into oligarchies with skewed markets that favor those with all the capital.

If profit were the only motive without any other rules in play, that wouldn't be capitalism at all. We need interventions in order to preserve a healthy system. To suggest otherwise is to defend an ideology that isn't capitalism itself.

What is the alternative? Develop everything yourself in-house? That's not just expensive, it's dumb, because you'll get worse/less reliable software in general.

It looks like Matt Asay, the lead for the open source and marketing team at AWS, has already reached out and said he's looking into it (and thanked Tim for the contribution).

I think there's generally a cultural norm to recognize an individual's contributions in general, especially when freely given.

If the comments on here largely echoed that sentiment and demonstrated that it was a cultural norm, expect AWS (and others) to be more likely to adhere to it in the future — it costs almost nothing, but there's definitely a value in having a positive reputation.

We do have the capacity as a community to define and uphold such cultural norms. Laws and licenses are not as binary as code.

That’s pretty much unrealistic, given how many pieces of oss a project uses.

Definition 2 of exploitation: "the action of making use of and benefiting from resources."

Definition 1: "the action or fact of treating someone unfairly in order to benefit from their work.".

I don't like being accused of bad faith though, I do put a lot of effort into strong-manning positions I disagree with.

I really do think we need a couple of new licenses along these lines to become popular enough to be generally accepted. Having said that, I struggle to really grok the language in these licenses; it could be that I'm just more familiar with MIT, BSD and the like, but with the Polyform licenses I'm always left with questions about what I can and can't do.

We ran into that, sort of, although the only one it ultimately affected was the author of the free software that we wanted to use. We were a small company doing development for a major Japanese software distributor. He wanted a virtual CD product that would make images of your CD-ROMs and let you use those images. We were using zLib to compress the images, but he said it was too slow. zLib has settings to sacrifice some compression to gain speed, but they would not gain enough speed while maintaining the minimum amount of compression he wanted.

I found a zLib compatible library that some grad student at a nearby major university had written in assembly for speed. It was very good, handily beating zLib, and he even used the same assembler we used (Watcom). The library was GPL, which was not acceptable to the distributor we were writing for, so I contacted him about a different license.

We agreed on some reasonable price to use his library in that product, and some reasonable larger price for a license to use it in all our future products. But then he started worrying--our products were for Windows. What if we caught Microsoft's attention and they bought us, and his code ended up available for all of Microsoft to use.

He wanted to negotiate a license that would cover all of the kind of thing to make sure his code could not end up at Microsoft or some other big company. Negotiating that would require our CEO's involvement, and probably bringing in outside lawyers. The CEO did not have the time for that nor any interest in it, and told me to figure something else out.

I did. I went back to zLib, and I added a slider to our UI which went from 0 to 100, labeled something like "Faster ripping" on the 0 end and "Smaller Images" on the 100 end. If the slider was set to N when ripping an image, I used zLib at maximum compression setting on N out of every 100 sectors ripped and stored 100-N out of every 100 sectors ripped uncompressed. The distributor was delighted with this. (I have never been able to decide if I should be proud of this solution or deeply shamed by it).

25 years later, and nothing from then ended up at Microsoft or anywhere else, and that grad student lost out on several thousand easy dollars that I'm sure would have been very nice and useful for him to have.

People should be able to comfortably use software within the bounds of the license without worrying about the author coming along and then shaming them for not complying with an additional set of implicit constraints.

Yes, it would have been nice for Amazon to acknowledge the original author. But given that they are not obligated to, it's unfair to act as if they are at fault for not doing so.

If attribution is needed, put it in the license.

If there is, it's a space that you can minimize using licensing terms, and as a user I will assume that the licensing terms are chosen by the authors in a fashion that best represents their interests. Especially since so many template licenses exist that address this exact problem.

> I don't like this idea that you have no right to complain about something if you haven't specifically banned it in a legally enforceable license.

I think it's unfair to imply that the post you respond to represents that idea, if that's what you're doing. Although worded frankly, it's a constructive suggestion for what proprietors can do to prevent this. It's not a new problem.

(edit: corrected amount, thx jimhi).

I really hate this happens to people. :(

This is fine of course. We have many competitors.

Maybe the developers would like to say thanks to barista but lawyers said to the developers to not talk to the barista no matter what since it's safer to say nothing than risk saying something wrong.

It's not befitting a large enterprise like AWS to tell their prospective customers that they are just wrapping their infra around free software in their marketing copy. If that single sentence even slightly impresses upon .1% of their potential customers that they ought to spin it up on their own infrastructure, obviating their need to pay for the service, that's plenty of disincentive to add the message.

If trust and reputation were enough, we would not need licenses. The GPL is great and had the impact it did exactly because it forced people to do things.

I said that in the '90s, because it was obvious even back then, and it was drowned out by the "We're gonna destroy all the evil proprietary software giants and all code will be free forever, yeah!" chorus. It's nice to finally be vindicated.

Giving attribution (and monetary compensation) to the lawyers who make open source licensing function is also a thing.

Maybe there is a way to define "bad" practices in OS outside of license. Similar to how monopolistic practices are banned in business because they are unhealthy for the broader ecosystem. Just asking

That's due to fact that AGPL's _user interaction_ clauses can be too vague in legal terms, in a way that many internal use cases could be litigated as a user interaction over a network.

It has been discussed before here:

https://news.ycombinator.com/item?id=16636963

If fact someone should come up with a well written licence that adds conditions when the licensee revenue is over a certain (large) amount, like say royalties. Everyone should be happy with that.