5 ms·
nit-pick: Elastic Search is not open-source (anymore)
by karussell 6y ago
nit-pick: Elastic Search is not open-source (anymore)
- cperciva 6y agoIndeed, that should say Open Distro for Elasticsearch.
- erezsh 6y agonit-pick: Only according to the OSI, whose credibility within the community isn't what it once was.
- karussell 6y agoAlso according to many people in and outside of the community. OSI or free software definitions are meant to give the contributor the same rights which they don't have for a "source available" license. Under a "source available" license ES would have never gained the same popularity nor commercial success. (Don't understand me wrong: ES can do what they want with their code but advertising it as "it is the same for most people" is untrue)
- erezsh 6y agoI'm not saying Elastic's model is the open-source ideal. But a product that can't find a financial model won't exist either. ES may not give their contributors the same rights on paper, but in practice there is no discernable difference, other than that the contributors aren't allowed to directly compete with them (iirc).
- Daho0n 6y agoSo open but not free.
- DetroitThrow 6y agoThe epipen design might have been "open to public inspection", but I think most descriptions of it are accompanied by the term proprietary rather than open because of the nature of it being closed for them to use exclusively. In contrast to the patent on insulin being made open for all to inspect, use, and build upon - SSPL does not smell very "open source" if it's similarly opposed to forking or being the foundation for something someone takes in a different direction.
- DetroitThrow 6y ago>other than that the contributors aren't allowed to directly compete with them (iirc). No, the wording in their SSPL is much more broad than that, such that creating any logging facet would put you in legal ambiguity at best - the worst case, which unfortunately now two licensing lawyers I've watched be adamant about this being the realistic interpretation in court (/dev/lawyer and one who advised our and other software companies), is that you really can't be in compliance without releasing your entire software stack. I totally agree with the idea that a business becomes before project/idealism considerations though, if you're a company trying to finance an OSS project you need to think about your financials/survival before any licensing purity - which I don't think anyone is criticizing them about in this scenario anyways. That being said, you can make changes to your project for your business but do so on terms that the internal and external community don't find disingenuous, don't characterize those whom are upset with a withdrawal from previous promises(1) made to them as conspiring to spread misinfo, or even to have taken many in the community's suggestion for dual licensing under copyleft license and commercial license that provides a hedge against large cloud providers. Anyways, if Elastic's response to going back on a promise is to gaslight their own (reasonably) begrudged customers and contributors about how the company going public shortly is the real little guy/victim, then I think most already know whether to trust their lawyers or the company. 1: https://web.archive.org/web/20200120104750/https:/www.elastic.co/what-is/open-x-pack https://web.archive.org/web/20200120104750/https:/www.elasti...
- bosswipe 6y agoIt is not Open Source™, but it is open source.
- DetroitThrow 6y agoWait, I'm pretty sure you're that guy from the original "Elastisearch doubling down on [gaslighting the OSS community]" thread! The same guy who defended the Elastisearch exec by claiming that by turning an OSS project into proprietary software was equivalent to companies open sourcing previously closed software, despite the former benefiting from contributions+customers which had been misled by communications about it not being or becoming proprietary! Just curious - why do you invent your own meanings for terms that have different meaning for the vast majority of readers? This is a genuine question since I don't think the doublespeam comes off as either sincere to those in the community or helpful to those learning about the subject, who are still only just a Google search away from finding that no one else besides those benefitting from taking away user rights describe such projects as open source: https://opensource.org/node/1099 https://opensource.org/node/1099
- bosswipe 6y agoI'm definetly not the same guy from whatever thread you're remembering. I just feel like the new license doesn't take away any of the freedoms I care about, the source is open, I can run and modify it without issue. I can even run it on a AWS instance. I want more open source companies to figure out sustainable business models so we can have more open source software. Amazon taking all open source revenue streams will lead to less open source software. I feel like OSI plays all these word games that are in bad faith, calling SSPL "proprietary" which is just wrong. How do you know that for "the vast majority of readers" open source means exactly the 11 point definition that OSDI came up with? (I realize that by replying I'll probably garner another -4 HN points, but I'm willing to spend the karma since you said it was a genuine question.)
- DetroitThrow 6y agoDon't worry, I won't contribute to negative internet points, this is a substantial response. I agree no one should be religious about OSI's standards, but open source colloquially means "open for people to modify/fork/use" - I think the FOSS movement has created a similar colloquial definition even if FSF isn't the end all be all for free software. The differentiator "free as in freedom, not beer" is a great and uncontroversial litmus test for the community. OSD 6 I think presents something similarly uncontroversial and closely tied to the colloquial definition of OSS (unlike other parts of their directive) which is "No Discrimination Against Fields of Endeavor" for licenses. It's not very open if one can't fork the project should their vision diverge, which is explicitly what both of Elastic's reasonably both paid and unpaid proprietary licenses will restrict - it keeps the software closed to their domain rather than fully open. This is in contrast to open copyleft licenses which only require that you return contributions that you make to the codebase - something very comparable and that open source businesses have successfully worked with is dual licensing AGPL. With it I have no restrictions on how I use the software, but if I consider it a business need to begin running it, I have to be *open* about what I'm running. It's not going against the broadly almost anyone's understanding of OSS to say "fork it, do whatever" or "fork it, just be open" - but it is going against most people's definition when it's "fork it, but you don't use it if you fall into a category that competes with our monopoly". Patent rights are an apt comparison - just because a patent is available for the public to view doesn't necessarily mean it's open for the public to expand on without paying a proprietor! And what you say about the SSPL just being okay to run for your own purposes isn't true either - I've been advised by two lawyers at this point that section 13 I believe is so broad that it's viral enough to require you basically publish everything that you run in a stack because of how open they leave the language to interpretation, and their stipulation that it's "not that broad" in a FAQ does not supersede the language used in a license - Google attempted that defense early on in their battle with Oracle about how Java namespaces were previously advertised as open. It turns out the law prioritizes licenses over box stickers every time, and I don't think any business redistributing ES in any form can guarantee they're complying with the license (which redistributing in Elastic's SSPL's terms could be a logging facet for an application that's just backed by ES). Anyways, I totally back companies making great software proprietary or closed, it pushes our industry forward - I just find it abhorrent the argument being out forward is that anyone can redefine "proprietary" if they're up against a massive company (even when they're a commercial and not so tiny business themselves) and especially if they're gaslighting the broader community about how they're doubling down on open without any irony or explanations about what the majority thinks open means in contrast.
- m463 6y agoI think GPL software can be USED for any purpose, but has restrictions on redistribution. Many of these other "open source" licenses have restrictions on who can use and how the software can be used. They're basically created out of fear to prevent competition.
- knocte 6y agoYou see it as "restrictions" because you must be a developer. GPL licenses mainly benefit the user of the software, and for the user these are not restrictions but rights (e.g. the right to see the source).
- DetroitThrow 6y agoGPL licenses greatly benefit contributors as well! Not only does that mean improvements are required to be introduced into the public domain for other contributors to build upon further, but it also allows contributors to take their own direction with an (open) fork. Much of the benefits the user has here is also what benefits good contributors ironically, in contrast to licenses like SSPL which restrict contributors from freely and openly diverging (and perhaps competing with for the benefit of users) from a proprietor.
- rimliu 6y agoHow does seeing source helps someone who is not a developer?
- anoncake 6y agoFree software is not about seeing the source code.
- diffeomorphism 6y agoSame way free speech helps someone who is not a journalist or right to repair helps someone who is not a mechanic. If there is some free software I like, but it is is missing a feature I want, I can pay anyone to make these changes for me. If the software is proprietary, I can try asking the owners. They say no. The end.