The author posted this in a recent discussion on GitHub [0]. Thought I'd share it to get more opinions on the matter.
Personally speaking it irks me a little when different terms are made ownership by someone. OSI is doing great work but is it considered an authority when it comes to defining what open source is? In other words, can open source exist without OSI? I certainly think so.
For example, what does it mean when Sentry says their licensing scheme is not acceptable by OSI [1]? Does it change the meaning of open source?
Yes, OSI is considered the authority on this. Who else? They invented and popularized the term, after all, back in the early dot-com era.
Of course, they can't do it alone. The question is whether you should support OSI's efforts or not. Do you want "open source" to have a clear meaning or do you want to drift into meaningless like "agile" or just "open?"
It's the same as any other standard. If everyone stops following the standard then it will fail, but is that really want you want?
Open source was not invented by the OSI; they claim that, but you can easily find press releases going back to the 80s using the term.
But yes, on the popularity bend, you're correct -- the OSI basically was a giant PR campaign, started by Netscape to publicize their going-out-of-business sale as a good thing, actually. Most importantly, they specifically wanted to separate themselves from the FSF project's four freedoms, which they viewed as antithetical to business -- or rather, antithetical to any business who wanted to buy the charred remains of Netscape, like AOL. So they started with Debian's DFSG, and whittled it down until they had something they could sell.
In my opinion, people argue about terminology too much rather than the freedoms and the society we want to build, but that's just the internet, sometimes.
OSI being in charge of open source is like that Indian tribe that sold Manhattan to the white man for 28 dollars. They are always cited as being idiots but they were selling another tribe's land and were happy to get paid for it. Likewise people like Eric Raymond were only too happy to facilitate the businessification of free software for a little time in the limelight.
It's about one organization appropriating years of work like they've been the ones doing it all along and corporations being all too happy to accept it as such because it is good for business.
Free Software's only relationship to Open Source is that you can relicense OSS with the GPL just by kissing it (as long as you comply with any minor strangeness in the OSS license.)
Open Source tried to steal the energy of Free Software and steer it in a business friendly direction, but they didn't steal the term "Open Source" any more than Free Software stole the term "Free" from "free beer." These are just terms of art, certifications by two very established and influential certification bodies, and wittering on about whether somebody said they were "going to open the source and let people have a look at it" in 1973 is as irrelevant as the fact that plastic polymers are organic but not "Organic."
Are you all replying to something the parent said or something? I am not talking about something so simple as the term "open source". In my eyes open source is free software just in a shinier package. (We do seem to be largely in agreement you and I by the way)
AS far as I know, OSI didn’t appropriate anyone’s work. If you’re going to make an accusation like that, you should explain specifically what you mean. what work? Where do they claim ownership? Who had ownership before? How do you know?
open source didn't want to separate itself for the 4 freedoms, it wanted to separate itself from the toxic and cultish free software culture. to put it more charitably, free software is ideological and utopian whereas open source is pragmatic and realistic. either way these differences are cultural and produce the exact same type of software. in almost every single case an open source piece of software is also free software and a free piece of software is also open source.
An additional question might be, "Do we want to allow a tiny group of people to declare that they, and only they, can pick a generic term like open source and then dictate that their definition is the only definition?"
This is not "Kleenex" or "Champagne". This is a generic term.
You could say that about many standards, though. Nobody living was around when they decided what "meter" and "kilogram" mean. And the adjustments to the exact definitions are decided by tiny groups of experts. You don't get to vote on that.
But someone who decided on their own definition of "meter" would be causing confusion, so nobody wants to do that.
Open Source isn't a generic term, it's a weird term. The only reason it seems generic or natural to say is because of the OSI, which makes it more like Kleenex than champagne is like Kleenex.
Coraline Ada Ehmke's stated reason for attempting to join the OSI board was to try to change the definition of open source to permit restrictions on fields of endeavor (specifically, getting her "ethical source" license vetted as open source). The OSI still has emotional if not legal authority over the definition, even among people who aren't satisfied with that definition.
According to this press release they did not win the trademark for ‘Open Source’, only ‘OSI Certified’ (which would be a wholly better definition for people to use, however, it’s a term that might need explanation or further context when used, so it’s not often said).
No organization can strictly define "open source" except to say which definition they use. A more valid way to approach the issue is to point out that a license is or is not not approved by the OSI, the FSF, or some other person or group. "Open source" has always been and will always be as fuzzy a term as "the Metaverse" and "handsome".
For most developers and even a fair number of users open source isn't a fuzzy concept but a very concrete technical term describing a very particular kind of license.
Being under an open source license can provide a huge boost to a project or product. As such it is a powerful marketing tool and companies try to exploit that by systematically undermining the the most widespread definition of open source.
The danger of course being that we end up with a term as nebulous as Metaverse.
I would even go further and argue that we should cement that definition into law, i.e. you can't call a product open source if it hasn't an approved open source license. The EU has created precedent for that by protecting the names of certain foods and beverages with regard to their origin, manufacturing and characteristics, because it's an area that faces many of the same issues.
I am aware of the history of the term and how it was taken up and popularized by the OSI. It has served an important and useful purpose. That said, they can't camp on a general purpose term. It isn't like "Champagne", which is a region with ~clear boundaries.
It wasn't a general purpose term when it was coined. That only occurred later, as companies jumped on the bandwagon while not actually permitting the user freedoms that are goal of open source software.
(There's a side argument of whether "Champagne" should be a protected term. The physical region has clear boundaries, but that doesn't mean that the beverage should be defined in terms of the region. If I am opening a combination diner/antique shop, I may advertise it as "Hamburgers and China", even if I am neither near Hamburg nor China. I do not need to follow the lead on calling champagne "sparkling wine" and name my ill-conceived business venture "Ground Beef Patties and Exquisite Ceramics".)
"For most developers and even a fair number of users open source isn't a fuzzy concept but a very concrete technical term describing a very particular kind of license. "
A particular kind, yes. The OSI definition? I'd want to see something that showed objectively that this is what people think. I saw some projects I support come under fire for not being open source recently- I've been. a user and supporter of FOSS software for a long time. I didn't realize that making exclusions on who can use your software wasn't open source. If a developer doesn't want their code used in a weapons system, it doesn't meet the OSI definition.
I know it now, but then I think that leaves the software to use a term that's much less accurate. I don't think that happens because they have a profit motive and want to exploit the term open source. In fact I've seen licenses that are essentially open source except they try to stop people from making a bunch of money off of community work, including those currently controlling the project.
I spent half of my life around Open Source and Free Software. The only cases where people questioned the OSD definition of "Open Source" I've ever seen were either projects that wanted to benefit from the positive clout the term has gained over years without actually doing their part in return, or people who were clearly confused on what this term actually means (usually in relation to Free Software), tried to explain it to someone else and when pointed out they're wrong they didn't want to admit their mistake and went "but OSI can't dictate what this term means". Never seen it used differently in an actual, honest and informed way anywhere that matters.
> I didn't realize that making exclusions on who can use your software wasn't open source.
Really? How? Isn't that incredibly obvious? If my license says "this software can only be used by Jane Doe from Boston", is it FLOSS? Or even "only by people who pay me $2000 every month for using it"? After all, it's just a (large) exclusion on who can use my software.
If a license tries to stop people from making money, it's clearly not "essentially open source" in its very nature. In its essence, "open source" means "everyone can use it, and no limitations can be placed other than this particular set of acceptable restrictions". How can you expect that placing additional arbitrary restrictions still gives you the right to call your thing "open source"?
"Isn't that incredibly obvious? " I just asserted that it is not. That was my entire point. I'm not sure how you missed it.
Your assertion that everyone who doesn't agree with you is a liar is problematic. I recommend really thinking about it and the ramifications. You are saying millions of people all agree with you or they are dishonest. Or possibly confused and they will agree with you when the confusion is cleared (unless they are dishonest). I can say I am not convinced and I'm not a liar or a person who is defensive and embarrassed by my not knowing this previously.
I was not aware of this provision that the software must be available for any use until this year (I've been using/involved with FOSS software for about 22 years - unfortunately a good number of years less than half my life) and it didn't occur to me as being necessary because to me saying something like, "this is open source except for military use" or "this is open source for non-profits" seems just as reasonable as "you can't modify this code and distribute binaries unless you distribute that new code". For people willing to agree to the terms, it is open. For the people unwilling to agree it is not.
I've never been really involved on the legal end of it all. In business contexts I used whatever was approved and the personal projects I've done, I've always used an MIT license because it was the easiest for me to understand.
I found out about this on github when I saw some issues opened on a project, because it was described as open source but the license didn't fit the OSI description. I found it to be interesting but also disappointing because it seems to be an issue of "comply or be attacked". There were links about why the vitriol was justified, but I read through it all and wasn't persuaded that the anger directed at the project owner was appropriate.
A license by its very nature restricts things. It's just a matter of which things and I think there is room in this space to be a good actor and have a different opinion. As the GPL says, "To protect your rights, we need to make restrictions..."
I don't see a long chain from "I don't want anyone to keep you from reading and modifying this code" to "I don't want someone to use this code to create an environment that takes away the human rights necessary to read and modify this code".
The fallback position I've seen is the one repeated here - "Everyone already knows and agrees with this except bad actors." and I just doubt that's true. Too many people make software and it's too large a group for me to take at face value any claims regarding a majority understanding. I'd want to see something objective that shows the work was done to know this rather than it just being an opinion or reflective of anecdotal evidence.
I did not. I just fail to see how it isn't obvious to anyone who gives two thoughts about how the term "open source" is being used, so I wanted you to explain it.
> For people willing to agree to the terms, it is open. For the people unwilling to agree it is not.
I'm sorry, but this doesn't make any logical sense. You can easily extend that to "for a person willing to buy the copyright from me, it's open; for a person unwilling or too late to buy it is not". How many people can you exclude until such a license stops being "open" in your book? This approach dilutes the meaning of the term so much that it completely ceases to be useful. I honestly can't understand how does that make any sense to you.
> Your assertion that everyone who doesn't agree with you is a liar is problematic.
That's not my assertion. If I saw entire communities using the term in conflicting ways, we would be having a completely different discussion. That's not the case though, the flamewars were always being fought in the exact opposite direction by people who perceived copyleft as "not open enough" (and even they usually didn't contest the meaning of the term, just argued about their own stance). In my experience, "Open Source" has always meant what OSI has defined, to the point that I don't even see it as being actually defined by OSI, but rather that OSI has written down the meaning that was already agreed on in practice (it's heavily based on DFSG after all). It wasn't until relatively recently that some projects started to claim being "open source" despite of not actually matching the definition. The benefit for them from doing so is obvious, and when they argue for the definition to become broader this is clearly a classic conflict of interest. I can't agree to that, because this dilutes the meaning of almost two decades of my work. Go use another term, this one's already taken.
> A license by its very nature restricts things.
That's absolutely not true. In many jurisdictions, a license is pretty much the only tool you can use to unrestrict things, the only alternative usually being a Public Domain dedication (which isn't even available at all in some parts of the world). By default, all your rights are reserved and nobody can use your software unless you give them an (explicit or implicit) license. Open Source licenses are meant to unrestrict your code as much as possible, only allowing some specific kinds of restrictions that are meant to protect the unrestricted nature of your code and to give you attribution for your work. That's it, that's the universally agreed meaning that was in use for a very long time now. You can't get to restrict arbitrary things on top of that and still call it open. Sure, you can argue that your particular ideas for restrictions are also meant to protect people's rights, but to make it actually accepted as "Open Source" you would have to convince the community (not just OSI!) to reach consensus that this can, in fact, still be called Open Source. And that's clearly not what's happening, so if you continue to use this term in the wrong way, I can't see it as anything else than being malicious.
I have no problem with talking about Tesla cars; nobody argues that we should be talking about Nikolai Tesla's personal car, or a car down the street that somebody spraypainted the word "Tesla" on.
The closest thing to an authority on what is open source is OSI. Open Source doesn't have a fuzzy definition, except where people would like the definition widened to fit their marketing needs.
Why do we need an authority? When I first heard 'open source' I thought it meant that the source code was open access, I would imagine most people don't put much more thought into it until they see one of these debates on use of the term.
People incorrectly using terms they've heard other people throw around is not a new phenomenon. So your "When I first heard [$TECHNICAL_TERM_FOO] I thought it meant [$BLAH]" is not a compelling argument. The fact that people who are uneducated about something don't understand it very well is not a surprising outcome. It is the expected outcome.
Lots of people think that "public domain" means something other than what it does mean. (Coincidentally, the most common misunderstanding arises from an assumption that it means something very similar to if not exactly the same as what you thought of "open source".) And yet, its actual definition remains in effect.
Whenever this discussion is rolled out again, it's useful to address a few key points. First, open source is defined by the open source definition. This is not contingent on the OSI receiving a trademark for the term. The common usage of the term is defined by the OSD, and much in the same way that selling "cakes" and fulfilling orders with used car tires is dishonest, so too is pushing non-free software under the brand of open source.
And note that open source is defined by the OSD -- not by the OSI. Should the OSI attempt to re-define "open source" without careful consultation with the community or in service of a conflict of interest, the community will withdraw its support for the new definition.
Painting the term as having always been loosely defined is historical revisionism. What we have always seen, and what we are seeing now, is a minority view that seeks to forward an unorthodox definition of open source in the service of their private financial interests. I am not prepared to accept a novel interpretation of open source because you feel that it would be more profitable for your business if I did.
The open source brand has been wildly successful and it's a lucrative target for bad actors to exploit with non-conformant software licenses. Don't be a sucker. These interpretations are not in your best interest.
Thanks for your input here Drew, your actions and writings have been what has inspired me to call out such scenarios (Of non-OSD "Open Source" usage) when I see them.
I have always had to treat different licenses differently. BSD, MIT, LGPL, GPL, AGPL, Apache, etc. They all mean different things which have impact on how and where I might place them.
Subdividing "source available" as implied but not well defined in this article, into its own group does not substantially help me. When someone describes a useful "open source" thing to me, one of my first questions is which license it is under, and that would be the same whether someone said "source available" or "open source".
Perhaps there's something more behind "punch in the gut", but I don't really know what that is. There's a link to a tweet about these projects being boosted under the open source term, but the tweet has little engagement: 20 project logos, 43 likes. Did it actually reach beyond direct customers? Did the term open source have any impact at all here?
Source-available has been called "Shared Source" as far back as the early 2000s. Open Source was specifically intended to include the usual "freedoms" of free software, using a less ambiguous term than "Free".
often depends who you ask, istr stallmans take on "open source" is that the term came into existence explicitly to dilute the free software movement and make space for evil companies to do evil things. if you were to buy this then it may continue from that logic that everything covered in the article is well labelled open source.
i don't really condone stallmans biased and divisive position on this, but it is exemplary of the kind of conversation we're wading into with opinions about what it means historically.
> the term came into existence explicitly to dilute the free software movement and make space for evil companies to do evil things.
It did. That has nothing to do with whether Open Source Software refers to software that uses an OSI approved license. It's the people trying to cloud a technical term with a dictionary meaning in order to be parasitic on FOSS without being FOSS who are being divisive, not people using the terms normally. Stallman isn't doing that.
> if you were to buy this then it may continue from that logic that everything covered in the article is well labelled open source.
I mean, this is just nonsense. Stallman says Open Source is bad, therefore everything bad is Open Source?
Maybe it would be better to promote and rally around a term that’s more specific? The best terms are the ones that naturally lead someone to the right conclusion, and are hard to make incorrect assumptions about. The main problem with trying to gate keep a generic term like “open source” is that it legitimately has multiple meanings, and the obvious and literal interpretation is the very one the author is complaining about. This is going to forever remain a fight against the tide, or a pee into the wind, or any number of other colorful well known phrases. It doesn’t take anything away from OSI to acknowledge their great work and also refer to OSI’s definition of open source specifically, when that is the one you mean.
Article author here. My concern would be that you could use any term but, as soon as that term gains value, it would start to be abused and blurred and we'd be in the same situation. It would be a shame to effectively throw away the value of "Open Source" that has been built up by the OSI and the community around Open Source software. I don't think the term is past being useful, the vast majority of its use i still see it applied as per the OSI definition, but there's always going to be a little maintenance needed.
“OSI open source” would not have the problem you speculate on. I’m sure there are other alternatives that could also work to solve this problem.
At this point, the counter problem is growing: the legitimate alternative meanings of the term “open source” are getting more used as time goes, not less, and speaking personally I don’t think maintenance or blog posts is going to fix it. Creative Commons licenses got pretty good at this by adding more words to each variant of their license titles. What about that? There’s an inherent problem with hoping to keep a term as short and generic as “open source”, when the issues you care about are separate from whether you’re allowed to look at the source.
the reason that creative commons isn't confusing and hasn't been co-opted isn't because it makes so much more sense. the reason that it's maintained it's meaning and can't be co-opted is because it is trademarked. if you make something that isn't creative commons but say that it is you can get sued. if you make something that isn't open source and say that it is you cant.
That’s a great reason to pick a different term, since “open source” can’t be trademarked. Something with “OSI” in the name probably could…
BTW, I don’t think you can reasonably argue that “attribution non-commercial share-alike” is not more specific and not a better summary of the license conditions than the generic term “open source”. Trademark is a good point, however I think you’re wrong to say that better license titles is not part of why CC licenses are less confusing to a broad range of people.
this might make sense if you were talking about free software. free software is very unclear and it makes people believe things that are not true about the software. free software doesn't need to be 0 cost.
this is not the case with open source. for close to 2 decades it was very obvious what open source meant and there was no confusion about it. it wasn't until the source available movement started to take off and they started to cash in on open source's reputation and identity for the sake of promoting themselves and cheapening the meaning that it became confusing.
the problem is that since this is a cultural movement and not a piece of intellectual property like a trademark there is no way to prevent people from contorting open source into whatever best suits their self serving interests.
It makes sense with “open source” too, and the author already acknowledged that explicitly. I don’t think this point is even up for debate. The obvious & literal interpretation of the term is, is the source available to inspect? That is precisely what the author said is not the definition of “open source”, he wants the definition to come with more than that, specifically that “open source” needs to come with limitations on what you’re allowed to do with it once acquired; the argument is that it must not be public domain.
I think it’s hyperbolic of you to suggest that when people use “open source” to mean “source that is open”, that implies they’re being selfish or contorting anything. You’re failing to acknowledge the very real problem that “open source” taken literally doesn’t mean what some people want it to mean.
> That is precisely what the author said is not the definition of “open source”, he wants the definition to come with more than that, specifically that “open source” needs to come with limitations on what you’re allowed to do with it once acquired; the argument is that it must not be public domain.
Uhm, no? That's the opposite of what the article actually says. Public Domain code is Open Source without any doubt, period. There is however a set of acceptable limitations that you can place on your code and still call it Open Source - mostly related to attribution or keeping derived works under the same terms. If you place other limitations on what others are allowed to do with your code once acquired, it's not Open Source regardless of whether they can see the source code or not. Which makes perfect sense to me - I can easily see the source code of Windows 2000, but that doesn't make it open.
This is incorrect, you are mistaken about public domain and about what the article says. The very first criteria in OSI’s definition of “Open Source” is:
1. Free Redistribution
The license shall not restrict any party from selling or giving away the software as a component of an aggregate software distribution containing programs from several different sources. The license shall not require a royalty or other fee for such sale.
“Rationale: By constraining the license to require free redistribution, we eliminate the temptation for licensors to throw away many long-term gains to make short-term gains. If we didn't do this, there would be lots of pressure for cooperators to defect.” https://opensource.org/osd-annotated
Public domain has no such restriction. Public domain always allows people to do literally anything they want, sell & redistribute the public domain code in any form they choose, make it proprietary, redistribute the code under any other license at all, and generally do anything they want for any reason including throwing away long-term gains for short-term gains.
It's you who is massively mistaken. What you quoted is one of the requirements needed to consider a license to be an Open Source license. To paraphrase it, it says "an Open Source license can't restrict any party from selling or redistributing the software (...), otherwise it's not an Open Source license". In other words, a license that doesn't restrict selling and redistribution (and also fulfills all the other requirements) can be considered Open Source.
You even quoted it yourself: "The license shall not restrict". Public Domain code does not restrict these things (nor anything else), therefore it's perfectly compatible with Open Source licenses and Public Domain code can obviously be considered Open Source [EDIT: at least assuming that it's actually effectively Public Domain, which can be hard to determine - see commoner's reply].
Sure, you can also take some Public Domain code and stamp a non-free license on it (just like you can do with code under some [but not all] of Open Source licenses, like MIT) - which I guess may be what caused your confusion?
Public domain code is not automatically considered open source, but for a different reason:
> Public Domain software may come with the rights delivered by those “four freedoms”, but you can’t be sure. It will depend where the software was written, where you are located, who the author is and where the people you are sharing the software with are located. A deployer or developer will need to at very least ask for advice before proceeding, and most likely will need to secure the services of a legal professional with experience in copyright law in each affected country. Even asking the author is unlikely to be conclusive. That’s why public domain software may be free software but is not certain to be.
> A solution would be to create a form of words to be used by the author to dedicate something to the public domain. It could simply disclaim ownership for the jurisdictions where that is possible, and then grant a copyright license that has the same practical effect as a public domain dedication for jurisdictions where ownership of copyright can never be disclaimed. Such a formulation has been published by the Creative Commons. They call it “CC0” and it is widely used and well respected.
However, you are right that the Open Source Definition's "Free Redistribution" clause doesn't require an open source license to be a copyleft license. The "long-term gains" rationale is an explanation for the motivation of the "Free Redistribution" clause, and is not a restriction on what open source licenses can say.
There are many open source licenses that allow "people to do literally anything they want, sell & redistribute [] code in any form they choose, make it proprietary, redistribute the code under any other license at all, and generally do anything they want for any reason including throwing away long-term gains for short-term gains", including 0BSD and MIT-0:
Ha, right, thanks for clarifying that. Public Domain is a legal mess on its own - and in fact, CC0 may not be a perfect license to use if you want to dedicate your software to Public Domain because it's explicit that it doesn't grant any patent rights (which is why CC0 code has been recently disallowed in Fedora [0]). 0BSD [1] and MIT-0 [2] are good alternatives with same practical effect, even though they don't try to use the actual institution of Public Domain where it's available.
How so? Who’s confused about what public domain means? Are you thinking of specific cases or precedent?
Assigning something to the public domain is equivalent to waiving all copyrights and license restrictions. Unlike OSI’s definition of open source, public domain comes with no agenda and no protections of any kind for the author, and no restrictions of any kind for the consumer/redistributor.
I don’t know if you’re misinterpreting my opinion on open source, but I’m in favor of Open Source restricitions, I happen to agree with OSI’s reasoning. I’m just also in favor of open language, where words are allowed to have multiple reasonable meanings, where it’s okay and a good thing if it takes time to explain subtle points, and where we don’t expect people to have to know that innocent sounding words are somebody’s slogan, and also know the complete history and special meanings behind the slogan before using it.
> How so? Who’s confused about what public domain means?
Laws.
You see, what various countries regard as Public Domain differs a lot. In Germany, for example, there's no way for someone to dedicate something to Public Domain at all. You simply can't waive your copyrights, period. Public Domain exists there, but it only consists of works which had their copyright expired or which have never had any copyrights assigned. If all you say is "I waive my copyright and dedicate this to Public Domain", then such declaration is simply invalid in Germany and may have no legal effect. This is why people use licenses such as CC0 to effectively emulate the effect of dedicating works to Public Domain even where there's no way provided by the law to do it.
There are more subtle effects involved when it comes to considering Public Domain status even between countries that happen to allow dedicating your works to it. It's very complicated. I still stand with all my comments about Public Domain being perfectly valid Open Source, but it has to be added that this assumes a CC0-like license (and patent grant if necessary) to make Public Domain dedication actually valid and non-ambiguous worldwide.
So, then public domain is only “open source” in theory but not in practice? You’re right, I’m wrong. OSI doesn’t recommend against public domain for the reasons I claimed, and I’m making mistakes in this discussion. Thanks for clarifying.
It can be Open Source in practice too, as long as you make sure that your Public Domain dedication is actually effective around the world. The best way to do that could be something like CC0, but CC0 itself is problematic because of its explicit no-patent-grant clause, so it's not recommended to use on software. Instead, licenses like 0BSD or MIT-0 are probably your best bet, practically emulating Public Domain and being officially recognized Open Source licenses.
The problem is that there's no such thing as "public domain" that is consistent in jurisdictions around the world. If you want to make something "public domain," the best way is to slap an OSI-approved license on it.
> However, you are right that the Open Source Definition's "Free Redistribution" clause doesn't require an open source license to be a copyleft license.
This is misleading in a way that contributes to misunderstanding IMO. Copyleft means specifically preserving the same license when redistributing. OSI’s license might not be copyleft per se, but it does in fact limit what the redistribution license can do, and requires the source to stay “open”. Public domain does not require the redistribution license to remain open. OSI is very clear about this, contradicting @seba_dos1’s claim that public domain is OSI open source compatible. https://opensource.org/faq#permissive
> OSI’s license might not be copyleft per se, but it does in fact limit what the redistribution license can do, and requires the source to stay “open”.
"OSI's license"? Which one? There's no requirement for the license to require the source to "stay open" for it to be called Open Source, and you have already linked to the page that confirms that.
> Public domain does not require the redistribution license to remain open.
Neither do some Open Source licenses, like MIT, which I already mentioned above. Which is even clearly stated by the page you just linked:
> Most copyleft licenses are Open Source, but not all Open Source licenses are copyleft. When an Open Source license is not copyleft, that means software released under that license can be used as part of programs distributed under other licenses, including proprietary (non-open-source) licenses. For example, the BSD license is a non-copyleft Open Source license. Such licenses are usually called either "non-copyleft" or "permissive" open source licenses
> OSI is very clear about this
It sure is, I don't know why you choose to not read what they write. Heck, you've even got links to 0BSD and MIT-0 licenses earlier, which are recognized as Open Source licenses and which you can both read in their entirety in 3 minutes and that would be enough for you to realize that you're wrong.
> OSI’s license might not be copyleft per se, but it does in fact limit what the redistribution license can do, and requires the source to stay “open”.
Please stop spreading misinformation.
It is not a condition of OSI approval for a license to require that downstream projects be reciprocally open source. You can absolutely take MIT licensed code and put it in Microsoft Windows or some closed source app, for example.
If you're unclear or have questions about something, then it's fine to say so and ask questions in the appropriate forum. What you're doing, though—which involves saying something and then waiting for someone to challenge or correct you—is harmful and contributes to even more outlandish and untrue beliefs about open source than what already exist.
To reiterate: the clauses you're quoting do not mean what you are interpreting them to mean, and claiming otherwise, especially so confidently, is actively harmful.
You’re highlighting the reason the author wrote his article. You’re still misunderstanding the OSI license.
The OSI license is a limitation on redistribution, and in case you missed it, it clearly states “the license shall not require a royalty or other fee”. That clause is fundamentally incompatible with Public Domain. The purpose of that clause was explained in the “rationale” statement; it’s an attempt to preserve and protect the open-sourceness of open source. There is no such protection with public domain. This is exactly what @ssddandrown was talking about, it’s the whole point of the article.
> You even quoted it yourself: "The license shall not restrict".
I did quote it, I think you got confused about which license this is talking about. It’s referring to what the redistributed code’s new license is allowed to do. Public domain code can legally be redistributed under a license that does restrict what others can do with it. OSI Open Source code does not permit redistribution with a restrictive license.
> The OSI license is a limitation on redistribution
There's no "the OSI license". There's a set of requirements OSI defined to classify software licenses as either Open Source or not.
> it clearly states “the license shall not require a royalty or other fee”
Exactly. To be considered Open Source, the license can not require any fee (among other things). Public Domain code (let's assume CC0 here to avoid any confusion on its legal status) does not require any fee, therefore it can be considered Open Source.
> it’s an attempt to preserve and protect the open-sourceness of open source
Yes, that's why the licenses that require the code to stay free are permitted to be called Open Source. It's very clearly not a requirement.
> OSI Open Source code does not permit redistribution with a restrictive license.
Of course it does. Please read the OSI guidelines once again, from the beginning till the end. Throw away your previously held misconceptions.
It’s the redistribution license that cannot request a fee. Public domain allows the redistributor to charge a fee. You’re getting confused about the author’s license vs the redistribution license.
> let's assume CC0 here to avoid any confusion on its legal status) does not require any fee, therefore it can be considered Open Source.
...because of the explicit no-patent-grant clause. Yes, I'm perfectly aware, otherwise I wouldn't have mentioned that earlier myself ;)
> Public domain allows the redistributor to charge a fee.
So do Open Source licenses. You can sell software on GPL, MIT etc. with absolutely no issues. You just need to allow others to redistribute it further without any additional fees (but if they wish, they can charge a fee too).
It's in fact one of the requirements for it to stay Open Source. You are literally guaranteed to have a right to charge a fee with Open Source software. Some people would like to change that and to consider licenses that add additional restrictions like disallowing commercial usage as Open Source too, which is why the article we're commenting under has been written in the first place.
I do see some differences between "free software" and "open source" here, Specifically that "free software" is much more common and something that could be expected to be written naturally, without understanding this space. "open source" is not so common as part of natural language, and its main use has been around the realm of software and licensing.
IMO “free software” is misunderstood in exactly the same way “open source” is misunderstood, at least in the way you mean. Both terms are not good at getting the important point across in the title, because the words people want associated with the title aren’t actually in the title. This will forever be a problem as long as the title requires explanation (which it unquestionably does) and as long as software is a growing field and new people keep joining. I don’t know why the idea to say what you mean would get push back, we can’t expect people to interpret the phrase “open source” to mean more than what those two words imply. We can, and should, IMO, come up with a phrase that has words that imply the important bits. Following the CC model, maybe something like “derivation allowed” and “no discrimination”.
Both terms are fine, long used and well-established. Free Software and Open Source software have complicated enough meanings that they need a sentence, not a phrase to explain; the phrase is just a label. And any software practitioner who doesn't know what people mean by Free or Open Source software isn't very good.
If I start calling my OS "Windows" people might get upset even though windows predate MS by millennia.
You are asking a large group of people to start using a different term, because a minority want a term that has had specific meaning for decades to have a more broad definition.
Oh I think you have it completely backwards. The number of people who understand “open source” to mean OSI’s specific version is a tiny, tiny fraction of the number of people who assume “open source” means nothing more than what those two words clearly express: source available. This includes all new programmers (who outnumber us experienced programmers at all times), all lawmakers and law enforcement, all journalism, all business people, and the general public.
I’m only offering ideas on how to solve an existing problem of unclear words and unreasonable expectations, not asking anyone to do anything. We can continue to use “open source”, and suffer the consequences and continue to complain. I think the problem the author called out will continue to get worse. Plus, as pointed out elsewhere in this thread, OSI co-opted this term from others who had already been using it differently before them.
It would be way better if the term was clearly a name, rather than something that tries to gate-keep the meaning of general words. That would make it so you assume you need to look it up, and not assume what it means. It would be better if “open source” didn’t have other, easier to assume and understand meanings that came both before and after OSI’s ideas. If we don’t want to use a name, then it would be better if the phrase used a word or two to capture the ideas in OSI’s version. Using “open source” to mean something with restriction is a desire to have your cake and eat it too. This is in essence a marketing slogan that wants to rest multiple license limitations on the positive-sounding word “open”, without having to admit there are parts of the license that limit some kinds of openness and freedom in certain ways (under the possibly true belief that such limitations are the maximum balance of openness for everyone). Both term “open source” and “free software” suffer from this problem, hoping to establish a slogan out of two words that most clearly imply a meaning that is unfortunately antithetical to the movement. This is just a choice, and we could choose to be clear, or choose to fight the tide.
> “open source” means nothing more than what those two words clearly express: source available
Perhaps it is my long experience with OSS, but to me "open" expresses freedom to use and modify not just to peruse. Some source available licenses are in more of a gray area here. But I've also seen companies claim that they are open source just because their code is on GitHub, even though their license doesn't allow you to do anything with that source code.
> all lawmakers and law enforcement, all journalism, all business people, and the general public.
Most of these people have no idea what source code is, so open source means nothing to them, unless someone explains it to them.
> It would be way better if the term was clearly a name, rather than something that tries to gate-keep the meaning of general words
I don't disagree with you, but the term "open source" is already widely used (moreso I think than "free software" which IMO is even worse given the ambiguity of the word free in English). And I would argue that it is a term of art in the software industry. It isn't the only case where general words are used for specific meanings in our industry, by far. Coming up with a new term would require a lot of effort to teach everyone already familiar with open source a new term, rewriting a lot of literature about OSS, and probably some rebranding. And then, even if the new term does catch on, as others have mentioned, companies will probably still use the new term to incorrectly refer to things that don't meet the definition, for the same reasons they do now.
Would it be better to have a more specific name, maybe something that could be trademarked? Probably. Is it worth changing the term at this point? Probably not.
As someone who works on an open-source fork of the open-source software mentioned in the post, I am now exclusively referring to the open-source version of it as Apache (APLv2) -licensed. Maybe that's how we fix this confusion?
As long as there is money to be made by blurring the line, people will continue to do it.
But there is no consensus on what is "open source". Is AGPL open source? Despite the FSF and OSI approving AGPL, many would argue that it is a EULA.
The inevitable result is the erosion of the generic branding "open source" and the rise of specific licenses. "MIT" or "Apache" or "BSD" evoke one set of reactions, while "GPL" or "MPL" evoke another.
It's not "consensus" in the sense of "unanimity" (is anything ever, when it comes to widely used terms?) but there's a "consensus" in the sense of widespread agreement to use the OSI's definition.
> Is AGPL open source?
Yes. Why wouldn't it be? It meets all the standard requirements (https://opensource.org/osd) and the OSI has approved it.
> If I were to run a modified AGPL'd SSH server on my own hardware, the question of whether I'm violating the AGPL depends on whether I'm allowing others to access it remotely. If the AGPL can be violated without copyright infringement, it's clearly in a different legal category than the GPL.
That "[AGPL] meets all the standard requirements" and "OSI has approved it" is really an indictment of OSI judgment, not a sound argument with basis in copyright law.
Open source is all about the license. The code can all be available for anybody to look at but there might be a restrictive license preventing you from compiling it. A senior exec at my former company once said "Open source doesn't mean it's free".
"Open source doesn't mean it's free" is typically used to refer to implementation and support costs compared to commercial solutions, not to licensing shenanigans within self-styled open source projects.
Parent’s point is a good one: the specific terms of OSI’s definition of open source does hinge on licensing. Furthermore, “open source” is not mutually exclusive with “commercial” software like you imply. The US government specifically refers to OSI-style open source software as “commercial” precisely because of the licensing that comes with it. https://www.acquisition.gov/far/part-2#FAR_2_101
Must confess, I get a bit irritated when people try to explain "open source", but then very carefully steer away from "free software" in the explanation. It is like trying to explain what makes water flow, but trying very hard to avoid using "fluid" or its cognates.
My irritation aside, though, I can't decide whether the author is passively ignorant of the rest of the iceberg, or actively chooses to ignore it.
Its also noteworthy that the definition of "Open Source" is not something that any organization or entity has ownership over. OSI wants the definition to mean specific things. The FSF wants it to mean similar, yet different things (or that they never wanted "open source" to pollute the vernacular at all, and continue to use the "Free" software terminology.)
But the definition itself is fluid, contextual, and subject to change over time.
Yeah I think the FSF sees significant differences between Free Software and Open Source, in terms of the essential liberties that are not present in Open Source (different things missing in different understandings, but the FSF focuses on the fact that they are missing).
Although I’m no fanatical Stallmanite (Stallmanist? :), I do see the categorical distinctions clear as day; can’t understand why Open Source people play the willful ignorance card.
Free Software is good for people primarily, while Open Source is good for corporations primarily.
If people ignore FS and promote OS, what does that tell us about their intentions?
"Open source" was a service mark of OSI, an attempt to defend the term against people who use it inconsistently with its definition. I don't know if this is still the case, but OSI still advocates for use of "open source" consistent with the definition below.
Both open source and free software are defined as software that allows you to:
* use the software for any purpose
* examine the source
* distribute the software under the same license terms as you received it
* modify the software in any way
* distribute the software in modified form under the same license terms as you received the original
Stallman's quibble with "open source" is that open source messaging focuses on the business benefits of this kind of software rather than the inherent virtues of free software (and the inherent vices of non-free software). Note also that even to Stallman and the FSF, even BSD- or MIT-licensed works are "free software"; software licenses that forbid redistribution under a different license are called copyleft licenses under the FSF.
Article author here, I admit this is fairly specifically targeted at the term "Open Source" as that's specifically the term I believe to see misapplied most often. I wanted to write an article that I could share to provide some context when raising issues regarding licensing. While I could have gone a lot deeper into also describing "free software" and the history of the terms, I wanted something rather direct that won't serve more confusion (Since often this could be the first introduction to the meanings of "Open Source" and "free software").
Maybe I could add a "Further Reading" section at the end though with links to such topics.
Yes, "free software" is a competing term used by Free Software Foundation, but they don't certify licenses in the same way. They have a page commenting on different licenses [1] but it's less useful than what OSI does, in terms of defining a term clearly and attempting to be neutral about what licenses it includes.
Perhaps this is because their goal is to promote the GNU licenses, not to be objective about it.
OSI was created in part in reaction to ambiguity and confusion around "free software."
I'll be honest, the use of 'free' in your comment is vague on the same level of 'open source', ie. do you mean "free (freedom) software" or "free (as in beer) software".
For me F/LOSS is a perfect example of how anarchist philosophy let’s perfect be the enemy of good; practical solutions become as bad as no solution.
Most end users are just happy there’s a ton of free, quality code to run and study. Admittedly that’s only the case because of the FSF, but I think they should just accept that “open source” is a perfectly good term to describe software that is free and you can download the source.
Everything else is licensing nuances e.g, politics
The idea of Free Software, at its most extreme, exists as a useful endpoint of an "Overton Window" for software. Which is to say, all software benefits from its existence even if they don't e.g. sign on to the GPL. You're forced to consider it as a possibility.
Without it, we'd probably think of "Open Source" as the most "open" software can be -- which, as OP proves, is quite wrong and nebulous.
No I have it right, because most people stop listening when you talk about more abstract “freedom” when what they’ve got is literally (the good) and abstractly (the implementation) free. I get it’s an Overton window and admitted they moved it. However they’ve (FSF) failed to adapt to the current times and branch from there.
I think you're missing the possibility (and unfortunately, I would say the inevitability) of backsliding. I don't think we need anything adapting as near as much as we need people to "hold the line" or "stop loss."
For me, the perfect concrete case of this was here reading about youtube-dl being removed for Github. It became clear to me that a not-insignificant number of professionals in this space absolutely equated "deleted from Github" as "literally gone forever such that no one could have it again."
People will forget about the idea of free software and taking a copy of the code home with you as an important idea to be preserved, unless the FSF et al remain.
The minds behind the FSF were very much influenced by anarchist philosophy. The implementation uses the state because anarchism isn’t practical. As I said, they’ve been convinced the movement has failed since it’s inception and that we live in the dystopia they envisioned.
Despite their mission having been an overwhelming success.
Free software is associated with the belief that unfree software is unethical. If someone is making an explanation, but is not trying to tell everyone who makes propitiatory software is evil they should not mention free software.
I will try -- free is not the right word in English, it has ambiguity and also, unproductive aspects of meaning.
Richard Stallman and GPL discussions certainly do mention ethics as part of the considerations. I agree that no person is completely disjoint from ethical considerations, with intelligence. But as this thread explores, there are multiple points of view, unclear and untested aspects legally within multiple legal jurisdictions.
But you make a large leap when you say "they should not mention" .. It is a dramatic statement, not a logical one, it is more in common with persuasive, political speech than logic there.
As noted by others, a useful simplification is to focus on one particular LICENSE at a time. Is this license ... x y z. As Google did long ago and as Github started to do and as the OSI does and as FSF does... like that.. not debate on the entire concept, but instead focus on the implementation, as software does.
>free is not the right word in English, it has ambiguity and also, unproductive aspects of meaning.
I was not talking about free software as in the adjective free applied to software. But free software as in the term from the free software movement where its members call code that follows their ideals of what is ethical free software.
>It is a dramatic statement, not a logical one
The logical reason is that free software is less inclusive because it considers most SWE unethical for working on proprietary software. If the term open source works just as well and is more inclusive it should be used.
There's nothing magical or divinely inspired about the OSI open source definition; it's an elaboration on the DSFG. The DSFG has the hoops it does to facilitate the mental gymnastics of reconciling the appealing concept of free software with the ugly reality that the GPL was a poorly-implemented singleton that formed the bedrock of their enterprise.
Personally speaking it irks me a little when different terms are made ownership by someone. OSI is doing great work but is it considered an authority when it comes to defining what open source is? In other words, can open source exist without OSI? I certainly think so.
For example, what does it mean when Sentry says their licensing scheme is not acceptable by OSI [1]? Does it change the meaning of open source?
It's really confusing.
[0] https://github.com/streetwriters/notesnook/discussions/747#d...
[1] https://open.sentry.io/licensing/