Hacker News new | comments | ask | show | jobs | submit login
How I realised “Free Software” is a better term than “Open Source” (naughtycomputer.uk)
110 points by xylon on Apr 30, 2017 | hide | past | web | favorite | 118 comments



>> It’s not really about Freedom at all, because >99% of computer users are non-programmers.

> Free Software gives us as a society the freedom to control our computers. And it gives us as individuals the freedom learn how to program them if we want to.

Most people are not car mechanics yet they benefit from the rules that require 3 party products and repair shops to be allowed to operate.

And the ability to rapidly build upon the work of others reduces the barriers for new, interesting products to be built, which benefits everyone, even non-programmers.


Yes but you system is unable to create monopolies which is what business owners want.


Only big fish businesses benefit from monopolizing markets, everybody else suffers from it.


Those are the ones who pay politicians to change laws for their benefit.


When I talk about free software I make sure to mention the word 'infrastructure' because that's what it is. Just like roads or electricity it is a predicate for building a business (life) on top of.

Look at the biggest companies in the world (google, apple, facebook , amazon, microsoft ..) they own platforms i.e. infrastructure! but since these platforms aren't free you hear constant complaints from those who try to build a life on top of them (app store ranking not fair? Search not fair? OS/messenger/web browser/social network is walled garden? Etc.)

But even breathe the suggestion that taxes should support building reliable infrastructure (i.e. already existing free software projects) and you'll get a "government shouldn't interfere with functioning business" response.. no matter what structural risk that business carries for society and even though the "interfering" is just offering a viable alternative.


I would be happy if taxes couldn't be used to pay for nonfree software. If all government servers and government office PCs and school computers had to run free software, there surely would be money to support its development.


FWIW, the goal of the still-working-to-fully-launch funding system Snowdrift.coop is to do the best we could on a voluntary basis in the absence of tax-funded free/libre/open work. Like all public goods, these things are the sort that make sense to fund through taxation, but getting political power to make that happen is even harder than a half-way effort to increase voluntary mutually-assured donations…


I'd love it if SD.C allowed users to contribute extra money each period that was banked as an endowment-like-fund and drawn on only to cover unexpected shortages.

They're doing neat things, and I'd love to see it up and running.


Government can not be "alternative" to private business, for the simple reason they are fundamentally unequal. Government can use coercion to achieve its goals, private business can not (unless, of course, they capture a part of government, thus producing the worst of both worlds). Government can force people to consume certain products and avoid consuming others. Government can use coercion to extract money to cover economically deficient strategies. Government is largely immune from the consequences of their mistakes, even disastrous ones - if a private business does a mistake that harms people, they can be fined, or even put out of business, but if a government agency does it, it usually ends up with somebody getting early retirement with full benefits, and that's it. There can be no fair competition if one of the competitors vastly overpowers the others and can destroy any of them on a whim. That's why government's activity is called "interfering", not "providing alternative" - because of fundamental inequality of positions and capabilities.


<eyeroll> Standard libertarian/ancap nonsense that confuses interference with regulation, and actively maintains a blindspot to the role of government in fundamentally supporting every single business within the economy by providing a monetary system, markets, legal systems, law enforcement systems, transportation systems, as well as a myriad of subsidies and direct loans.

Please don't go on about 'privately produced law' or how private roads will replace public roads. I used to be an ancap and looked into PPL quite deeply back in the 90s - the field has made little progress since then that I can see, and appears to serve mainly to assuage the conscious of ethical libertarians, providing flimsy-yet-vaguely-plausible reason that something will fill the void when their desired collapse of government occurs.

>"Government is largely immune from the consequences of their mistakes, even disastrous ones - if a private business does a mistake that harms people, they can be fined, or even put out of business, but if a government agency does it, it usually ends up with somebody getting early retirement with full benefits, and that's it."

Are you freaking kidding me? Private business doesn't make myriad of mistakes and keep on going, even being rewarded? Did you miss the 2008 financial scandal? BP managed to poison an entire vast ecosystem and shoreline populated by millions, they're still going strong. Have you ever heard of the term 'golden parachute'?


As someone who has always seen ancap stuff as a mix of bullshit and good intentions (and it's easy to correctly identify a flaw in the status quo while having only non- or counter-productive ideas about what to do instead), I'd like to know more from a former ancap…

Do you have any ELI5 about stuff like (A) whether there's anything interesting in ancap stuff or if it's just repeating the same simplistic dogma all the way down and (B) how to get ancap folks to recognize the validity of agnosticism (i.e. somehow get them to take the first step from blindly dogmatic to being able to express their ideology as interesting ideas they like but acknowledge that they can't actually know omnisciently that it aligns with reality…)


> <eyeroll> Standard libertarian/ancap nonsense that confuses interference with regulation

<eyeroll> standard statist nonsense which thinks by changing labels the essence of things changes. See how easy it is to argue a position when you think argument means finding a name to call your opponent?

> Please don't go on about 'privately produced law' or how private roads will replace public roads

Please don't attach your comment on discussion with voices in your head to mine. I never said anything about private law or roads.

> Private business doesn't make myriad of mistakes and keep on going, even being rewarded

Sometimes they do. Sometimes they don't. Government always do, that's the point. Excepting violent revolution, of course, but you don't want to go there, do you? That's the difference - business very well may escape the consequences of mistakes, just as murderer can escape justice. That happens and nobody would pretend otherwise. But for the government consequences are not even on the menu - like for an executor in state prison, there's not even a question whether he can be accused of murder - of course not. That's the whole difference.

> Have you ever heard of the term 'golden parachute'?

Yes, I did. And a private company can afford only finite number of those. Unless, of course, it is too big to fail - in which case, government comes in, which can force me, the taxpayer, provide the gold for those. I don't like private sharehodlers being fleeced, but what is even worse is when I'm being fleeced even without being a shareholder. And that's what the government does.


Roads and electricity are paid with taxes, they have nothing to do with being free.

We see how well they are kept in shape in countries where the money doesn't go to them.

I don't want electricity that only runs for a few hours in the day, or roads full with crates created by the weather almost unusable when there is a bit of heavy rain.


Who gets to pick which "reliable" infrastructure program wins the subsidy, and when I become that person how can I craft the definition of "reliable" so that taxpayer money gets sent over to my friends?


A common problem with government funded projects which neither health care or education is immune from. In some cases money goes to someones friend (corruption), and in other cases it goes to people that are sincerely trying to make society a better place. The problem is to minimize the corruption. Oversight programs are one way, certification is an other, and in cases like medical funding there are boards which rules on a case per case and are governed under ethic rules (ie, money can't be given to anyone that any person on the board has a personal relation with).

There is however no perfect way to do it, and neither political party would want to take a extreme side. The right do not wish to stop medical research funding (private pharma can't survive without the government supporting the industry) even if some will be wasted on incorrect definition of "reliable". The left would not want to stop oversight programs and just hope that everything will work. As such we get some form of middle ground.


I've been kicking around a business plan idea for about seven years: A software marketplace for items under a GNU license. The key: I neglect to put that fact in the marketing.

The full license is posted with the software download, via a pop up EULA-style window before the download proceeds complete with an "I Agree" nuisance button.

The market starts with a price discovery model where I suggest a price and allow the person pay whatever they want. After some data gathering, I simply obfuscate the "pick-your-price" mechanism.

Yes, this is a bit of cherry-picking, where I employ some dark patterns. My reason for this thought experiment is the crux of this article and discussion: How do you explain software in the English language for which the rights of the end user are enshrined while leaving a profit-path open for the creator? My reply is: You don't. You focus, instead, on making the process of obtaining the software as friendly and service-oriented as possible.

I wonder how spectacularly this idea will fail?


We have a software marketplace for items under a GNU license. # apt-get install package

I would say you will fail most spectacularly, but it seems to work for RedHat Enterprise Linux


You would have a point if # apt-get required registering a credit card number. Last I checked, it does not. True, apt is a pedantically-speaking economic marketplace: a true free market. It is a free market based on ideas, usage, and popularity. It is not a market based on directly moving funds to the software creators.

GNU encourages software creators to sell their software for as much money as they can make while using their license. What I am playing around with as a thought experiment is a method to create a market that trades the scarcity of a software writer's creativity for money.

And yes, the attempt would probably fail spectacularly. Anyone who wants to attempt to execute on my idea is free to do so. Ideas are cheap. Execution is expensive.


With RHEL you are paying for support, not software.


Both, really. Support is included in the price, but the subscription fees for Ceph, Openshift etc are not exactly gratis.

The SRPMS are typically public so you can compile them yourself, but official binaries are only available through RHEL subscription channels.


RedHat Enterprise Linux is targeted at companies that pay for support and trainings.

I bet RedHat gets nothing from home users.


Any free or open source software license does not need an EULA, and presenting it as such is antithetical to the entire idea of free software.

You by definition do not need to agree to any free software license before you use the software, you can vehemently disagree with it or completely ignore its existence.

The license only exists to give you extra rights you didn't already have under standard copyright law (permission to distribute the program & modify it, under certain conditions).

An EULA is a dubious legal hack used by proprietary software to deprive you of rights you'd have had anyway if not for the EULA, which is the exact opposite of how free software works.

Edit: A lot of people are replying to this saying I'm incorrect. I'm not. A lot of free software licenses such as the GPL v2 spell out explicitly that they have nothing to do with anything except copying/distribution or modification:

    > Activities other than copying, distribution
    > and modification are not covered by this License;
    > they are outside its scope.[...]
The GPL v3 says the same thing in different words, and as FigBug points out downthread the general FSF license FAQ makes this clear[1].

The Open Source definition also asserts this by proxy. Licenses aren't allowed to impose any restrictions on users that they wouldn't have anyway if they had no license to the software.

Really, if you use free software you can just completely ignore all the licenses and you don't need to understand or have read any of them. This is one of the beautiful things about free software that separates it from EULA-using proprietary software.

Of course you need to understand the licenses if you want to distribute the software in any way, but that entirely falls under the umbrella of activity you didn't have permission for without the license, it would be copyright infringement by default.

This is also why free software organizations like the FSF would as a last resort sue someone for copyright infringement, not violating the GPL (but of course adhering to the GPL would bring them back into compliance).

This is different from proprietary software companies which might sue you over violations of their own made up rules, which you supposedly agreed to via the EULA.

The EULA is a hack to try to apply contract law to your use of the program, whereas free software & open source licenses purely piggy-back on copyright law.

All those licenses are a way to say "here's stuff we allow you to do, which you didn't have legal permission to do anyway in the absence of this license, given certain conditions" (or no conditions for e.g. the WFTPL).

1. https://www.gnu.org/licenses/gpl-faq.en.html#NoDistributionR...


Both EULAs and FOSS licenses are license agreements. Presenting them in the same way is not disingenuous. The length of the GNU GPL rivals that of some EULAs.

If you do not need to agree to a license before you use a piece of software, then EULAs would be invalid. You do need to agree to the license before you use the software. That the license grants permission to run the software for any purpose, usually with a disclaimer of liability, is irrelevant.

If you do not comply with (for instance), the terms of the GNU GPLv2, then you no longer have a license to the software. That means that you must cease distributing and using that piece of software or be at risk of a lawsuit from the copyright holder.


If you do not need to agree to a license before you use a piece of software, then EULAs would be invalid.

As they should be according to 17 USC 117: https://www.law.cornell.edu/uscode/text/17/117

The standard dodge is that it refers to the "owner of a copy", while EULAs claim that you do not own a copy and are only "licensing" it. Which is BS in my opinion, but regardless the GPL specifically doesn't make that claim, so you don't need to accept anything to run GPL software.


As others in this thread have said already, the GPL is written such that it does not need to be accepted in order to use the software.

> If you do not need to agree to a license before you use a piece of software, then EULAs would be invalid.

I believe the GPL was written the way it was, in order to remove any concern that it might not be enforceable. If someday the courts decide EULAs aren't enforceable, the GPL would still stand.

Of course, the GPL also relies on standard copyright law forbidding modification and distribution without express permission, so if the courts ever change that then the GPL is in trouble since it's not an EULA. But that seems pretty unlikely.


"The license only exists to give you extra rights you didn't already have under standard copyright law (permission distribute the program, under certain conditions)." Wildly false - not just copyleft, which I like to grab your work (fair enough) - but a thorough patent grab to take both your hardware and software patent rights from you.


This is not the case. As per section 9 of the GPLv3:

>You are not required to accept this License in order to receive or run a copy of the Program. Ancillary propagation of a covered work occurring solely as a consequence of using peer-to-peer transmission to receive a copy likewise does not require acceptance. However, nothing other than this License grants you permission to propagate or modify any covered work. These actions infringe copyright if you do not accept this License. Therefore, by modifying or propagating a covered work, you indicate your acceptance of this License to do so.

GPLv3 does require a patent grant if you convey the software, but as GP states, you wouldn't have the right to do that otherwise anyway.

EDIT: And as per the FAQ[0]: https://www.gnu.org/licenses/gpl-faq.en.html#NoDistributionR...

[0] https://www.gnu.org/licenses/gpl-faq.en.html#NoDistributionR...


Or if you contribute to it. So you're right, plain use is without imposition.


> A lot of free software licenses such as the GPL v2 spell out explicitly that they have nothing to do with anything except copying/distribution or modification

Check out AGPL. It exists specifically to remove the "loophole" of using the software without distributing it.

> Really, if you use free software you can just completely ignore all the licenses

For some very narrow definitions of use, maybe. But not if your usage includes building upon it, which is what OSS world excels at.


The AGPL closes no such loophole. It says you must make the source available if you distribute the output of the program itself over a network, e.g. in the case of an open source web application. This is something the GPL itself goes out of its way to allow (unless your output is the program itself).

So it falls squarely under distributing the program, which you didn't have permission to do without the license itself.

> But not if your usage includes building upon it[...]

Building upon it & distributing the result, but yes, this is what OSS is great for, but unrelated to my point that you as a user can completely ignore the license & don't have to know about it. You can just follow standard copyright law.


> It says you must make the source available if you distribute the output of the program itself over a network

Are you missing the difference between distributing the program and distributing its result? The latter lands firmly into "usage" category.

> So it falls squarely under distributing the program

No it doesn't. This claim makes zero sense - it's like saying if I send somebody a Word document it's the same as if I sent them a copy of Word. That's baloney.

> you as a user can completely ignore the license & don't have to know about it.

Looks like you somehow ignored the whole comment. Again, only if "user" means "never modifying or building on it", which in OSS world is a very restrictive condition.

> You can just follow standard copyright law.

Did you just claim "following standard copyright law" and "ignoring the license" is the same and both things are easy and obvious? There are people that get literally thousands to millions of dollars income by helping people to "follow standard copyright law". It's not something that you can follow just by going on with your business - it's insanely complicated and dense set of laws (actually, a lot of sets since we're in a global world now and each country has their own) and dismissing having to comply with it as "just follow the standard" is plain ignorant. Of course, if you're a random Joe or Joann, the chance that somebody would come after you is low (though see Prenda Law case), but if you have any money riding on it, be assured it's very far from "just ignore everything and follow the standard" - you don't even know what the standard law is unless you're a lawyer, and unless you also spent time studying caselaw and such (or asked a lawyer), you have very little chance of getting it right.


    > Are you missing the difference between
    > distributing the program and distributing
    > its result? The latter lands firmly into
    > "usage" category. [...] it's like saying if
    > I send somebody a Word document it's the
    > same as if I sent them a copy of
    > Word. That's baloney.
The AGPL is meant to cover the cases where the "output" is what the program itself does. E.g. free software web applications like bug trackers, wikis etc.

So the analogy is not sending someone a Word document being like distributing Word, neither the APGL or the GPL cover that. It's setting up a paid-for VNC service where users can access your copy of Word.

The GPL explicitly permits this sort of thing, because it's quite lax about defining network interaction, the AGPL is more strict.

    > Again, only if "user" means "never modifying
    > or building on it", which in OSS world is a
    > very restrictive condition.
Copyright law has no purview over whether you modify or build on something, and neither can any free software license. I can modify and build on any software I have that I haven't signed a contract to obtain, copyright law only comes into play when I want to distribute the software in any way.

If and when I've done that whether I've modified or built on it has little bearing on whether I'm committing copyright infringement.

    > Did you just claim "following standard
    > copyright law" and "ignoring the license" is
    > the same and both things are easy and
    > obvious?
I really can't see how you've read this thread and somehow become this confused. All I'm doing here is correcting WBrentWilliams's misconception that free software licenses are somehow something you need to get end-users to agree to before they use freely licensed software.

That this isn't how things work is something the GPL itself explicitly states. The same applies to any free software license, the GPL just happens to be a rather verbose license which goes out of its way to explain the basics of copyright law to you.


> Copyright law has no purview over whether you modify or build on something, and neither can any free software license. I can modify and build on any software I have that I haven't signed a contract to obtain, copyright law only comes into play when I want to distribute the software in any way.

False; copyright law makes it an exclusive right of the copyright holder to make copies and derivative works (the former is where the name "copyright"—the right to copy—comes from); distributing them may increase the likelihood of damages or other penalties beyond whatever minimum statutory damages exist for such a violation, but it is not the root violation that brings copyright law into play.


It says you must make the source available if you distribute the output of the program itself over a network, e.g. in the case of an open source web application.

Don't you confuse "distribute" and "let users interact with" now?

IIRC AGPL doesn't make a distinction about whether i send the bytes of the program over the network (distribute the program) or let a user send input and receive output to/from it (interact with the program).


>The license only exists to give you extra rights you didn't already have under standard copyright law (permission distribute the program, under certain conditions).

The license are the terms you must agree with or else you are breaking copyright law, if you ignore it and do things covered in the license you can be sued.


As a friendly pile-on, yes, you do have to agree to the terms of the GNU license or cease using the software. That's in black-and-white in the text of the agreement, along with making the source available and releasing changes to the source code under the same or later license.

My hypothetical model is choose not to do GNU's marketing. Frankly, they do a good job of it without my help. However, I cannot hide the fact that the software is GNU licensed. The EULA-style pop-up approach is a bit tongue-in-cheek, but presenting the license before download is vitally important.


> As a friendly pile-on, yes, you do have to agree to the terms of the GNU license or cease using the software.

That is incontrovertibly false and has always been false.

Per the GPL v.2, clause 5 (https://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html):

"5. You are not required to accept this License, since you have not signed it. However, nothing else grants you permission to modify or distribute the Program or its derivative works. These actions are prohibited by law if you do not accept this License. Therefore, by modifying or distributing the Program (or any work based on the Program), you indicate your acceptance of this License to do so, and all its terms and conditions for copying, distributing or modifying the Program or works based on it."

A end-user is free to run GPL software without accepting the terms of the license. However, they may not distribute it (modified or not) without accepting the terms of the license.


Not really. According to the GNU FAQ if you only use GPL software and don't distribute it then there are no conditions that apply.

https://www.gnu.org/licenses/gpl-faq.en.html#NoDistributionR...


I'm personally not a fan of the term free software. It makes me think of trashy freemium software filled with ads and tricks to get you to install browser extensions. Using the term open source creates a nice distinction between ad funded software and truly free software.


In contrast, for me "open source" literally means opened sources, i.e. one can read the source code. "Libre software" or FLOSS may be better alternatives for those who care.


"Libre software" and FLOSS immediately sound politicized to my ear.

As in, such projects carry baggage, in the form of conflict and infighting among developers, and one can expect irritating pedantry when discussing use and preferences.

Usually, you'll encounter something akin to fanatical fans, expressing highly opinionated views about minor details. Sometimes it's because lawyers have been involved. It's not necessarily toxic, but often reflects splintered cellular activity, among groups seeking to distinguish themselves from one another for some reason.


>"Libre software" and FLOSS immediately sound politicized to my ear.

They are, because pursuing your rights is a political move.

"Free software is a political movement; open source is a development model." — Richard Stallman [1]

[1] https://en.wikipedia.org/wiki/Richard_Stallman#cite_note-95

>Usually, you'll encounter something akin to fanatical fans <...>

As in any fight for freedom there are all kinds of people: those who do not understand, those who pretend they do not understand, those who do not know why they are fighting, and so on. No one of them defines the goal of the fight though.


The early Linux movement that later became part of the Open Source movement was certainly also political. Stallman was actually one of the heroes of that movement, which he apparently didn't realize. The difference was strategy. The FSF wanted to ideologically educate and convert people while most people in the Linux movement were focused on the number of users. The more users the more influence. Facts on the ground instead of ideology. History has proven them right.


you can't be neutral on a moving train.


Except some Open Source software is not, in fact, Free. Then there are sneaky variasions such as "Open Core", that indeed looks Open Source if you squint just right, but is pretty far removed from the original Free Software philosophy.

Thankfully, Most languages don't have this problem, since they have 2 distinct adjective for "gratis" and "freedom".


All the terms we have suck for a different reason:

* free - b/c freemium or worthless

* open source - a term big companies who don't care about it like to use to project a positive image of themselves

* libre - imo the best term atm. However, it just doesn't sound alluring.

Idk, how about Freedomware ? ;)


How are you even supposed to pronounce "libre"? :D

I think the "big companies who don't care about it" thing is exaggerated. Who cares about what terms they use?


This.

Free Software sounds cheap or trashy. It requires explanation. "Free as in beer, and Free as ..."

Open Source is a marketing term to avoid that stigma.


There is quite a difference between Free Software, which has the user rights as its main objective, and Open Source, which has the developers rights as its main objective.

https://www.gnu.org/philosophy/open-source-misses-the-point....


This is why I've always preferred FOSS, Free and Open Source Software. It's a bit more specific, alas however, FOSS seems to also be some sort of science program in the US, so it's ambiguous.


At snowdrift.coop we believe both freedom and openness are important values, neither of which encompasses the other. So, we use the acronym FLO, for free/libre/open, which has the nice side effect of sounding like the English word "flow".

https://wiki.snowdrift.coop/about/free-libre-open

---

The author misses the most important reason, in my opinion, why "free software" is also a bad term: people think they already know what it means, and thus don't really pay attention when you try to explain it. As such, I've stopped using it completely during my advocacy.

Instead, I use "FLO software", "software freedom", or "unrestricted software" - the latter being good for when I don't have time to explain fully but also don't want to give the wrong impression.


I realize that Stallman is a polarizing character, but I think the best rationale for Free Software (the concept, perhaps not the term) being better than Open Source is articulated by Stallman himself: https://www.gnu.org/philosophy/open-source-misses-the-point....

I liken democratic institutions such as free elections and due process to the differences between Open Source and Free Software. We don't do these things primarily because of their instrumental utility or because they necessarily yield the "best" results (e.g. Brexit, Trump, climate policy, etc.). We could imagine a political framework with a benevolent computer working as dictator that might do a better job, for some definition of better.

Instead, we have free elections and due process because we believe that they are, morally and ethically speaking, the right thing to do. If you believe as I do, that the human capacity for reasoning is greatly magnified by computers, then that's a recognition that human cognition and computers are in some way linked. Computers are the factors of production. De-democratizing those factors of production—even if Microsoft/Apple/Google/Facebook make a great product—is de-democratizing human thought to some degree.


I foresee that in about 10 years time when everything besides the Linux kernel has been replaced by business friendly licenses, we will be back to the days of shareware and public domain.

Personally I don't care, as I mostly use commercial software, but I bet those FOSS devs that helped change the landscape and won't be able to access whatever code comes in their devices might think otherwise.


> I mostly use commercial software,

You probably meant "proprietary software". Please say so.

Much free software is also commercial. A famous example is RedHat. A less famous, but much more common example is custom software: written for a fee, it is definitely commercial. And it's Free too much of the time: customers often have access to the source code, and the right to do whatever they want with it.

"Free" and "commercial" are not mutually exclusive.


Proprietary software, there. Happy?

> Free" and "commercial" are not mutually exclusive.

Red-Hat earns money in enterprise consulting and support, they withdrew from the desktop market because in what concerns the desktop, games and mobile markets, they are indeed mutually exclusive.

Which happen to be the markets I care about.


"Much free software is also commercial."

Almost all commercial software is proprietary. So much that free is barely a blip on the radar. It's reasonable default much like a person can say they ate a lobster without specifying it was red rather than the ultra-rare blue or gold varieties. Everyone will assume the default of red unless specified otherwise.


Perhaps you and the post you are replying to are talking past each other. There are at least[1] five worlds of software development: Shrinkwrap/Web/Mobile, Internal, Embedded, Games, Throwaway. All are largely commercial - throwaway software perhaps less so than the others. Embedded software is often non-free due to restrictions on loading a modified version of the software onto an embedded system. Games are rarely free. Shrinkwrap/Web/Mobile software is also rarely free but very frequently it is built on top of free software components - either permissively licensed ones or via the Web such that the vendor does not actually "distribute" the software. Internal and throwaway software is almost always free - the users and developers are either one and the same, or both are part of the same organization.

I would not go so far as to say that almost all commercial software is proprietary, but it is true that a lot of it is.

[1] https://www.joelonsoftware.com/2002/05/06/five-worlds/ It is posible that Spolsky's post is non-exhaustive


You're basically agreeing with me outside the internal/throwaway categorization. Your post adds some good details and categories. It's almost all proprietary. Some are supporting a bit more openness or modification allowed than before but always a cut-off point where I loose rights to benefit their wallets. There's a tiny percentage of companies working with truly open and free stuff pervasively but a good chance average public will never have heard of them with notable exception being Firefox. And that one's revenue for development and maintenance depends on several, proprietary platforms that themselves make money selling off users via surveillance. Bright, big picture, eh? :)

"Internal and throwaway software is almost always free - the users and developers are either one and the same, or both are part of the same organization."

This one is proprietary, shared source by default. In practice, it may have the advantages of FOSS for users within the company. It depends on the politics of the company, what 3rd party stuff it integrated with what licensing terms, and so on. Medium-to-large companies like IBM also treat stuff in one unit as owned by that unit even charging other units for using it often with a discount or certain allotment for free usage. Quite the opposite of FOSS. Now, if hobbyists or companies share their internal or throwaway code, they have to put a specific license on it. They might throw it on Sourceforge or Github but I still have no right to it without a license. This is why we encourage an open-source default on submissions to code sharing sites w/ user able to change it to another if they choose. So, status of throwaways and internal software still varies. And that's without getting to the fact that patent suits can hit you if a license doesn't factor in patents. It's why I don't fully trust BSD licenses as ensuring anyone can use it when it is something covered by patents.


Software doesn't have to be freely available to be free.

Software is free when the user is free. This also counts if there is only one user.


> Almost all commercial software is proprietary.

Not quite.

Almost all software people use is proprietary. Almost all software developers write is Free.

This is not contradictory. People mostly use massively distributed software, made by a relatively tiny number of people. Developers generally work on custom software, that has zero or one customer.

If we care about how much money can be gained with Free Software, one should look at the production, and see that custom software is a fine way to make a living. If on the other hand we care about freeing people from proprietary software, you want to look at the consumption, and see that there is much too much proprietary software around.


"This is not contradictory. People mostly use massively distributed software, made by a relatively tiny number of people. Developers generally work on custom software, that has zero or one customer."

It's not free until it's unencumbered by copyright or patents. I don't trust that in the least in the U.S.. Maybe in some other places. Here, anything you produce and use might get you in trouble. That doesn't feel Free. ;)


If it's only encumbered by the copyright and patents of the user (which is the case much of the time, because all rights are often waived), then it's free.

Surely you wouldn't sue yourself over your own claims?


"If it's only encumbered by the copyright and patents of the user (which is the case much of the time, because all rights are often waived), then it's free."

You can't guarantee that in the us. We're not even legally allowed to show someone a work with a copyright claim. We can't even reproduce a patented thing whether we knew it existed or not. Copyright and patent trolls increasingly sue businesses to the tune of billions of dollars. The risk is there so long as the system works the way it does.


It makes a huge difference when you're trying to sell blue and gold lobsters to people.


That's why the vendors of luxury goods state the difference. The rest don't have to, though.


My OS (macOS), laptop, router, gaming console, watch, camera, computer, phone, tablet, TV, set-top box, GPS unit and more run commercial software. They all include "free software" somewhere in the stack. Whether it's the Linux kernel, uclibc, busybox, libpng/libjpeg/libtiff or Webkit, it's there.

It's much more prevalent than people think, partly because it is rarely discussed or acknowledged.

If we look at web applications, you can be assured that the vast majority of them use free software somewhere in the stack. It's usually the case that the majority of dependencies in such products are "free".


"If we look at web applications, you can be assured that the vast majority of them use free software somewhere in the stack."

There's a difference between them being proprietary software and using some free components somewhere. You about to suggest Windows ecosystem is open or free because they used some open-source code in Windows somewhere? Or IBM's mainframes are open because they can run Linux VM's? Or the Facebook platform itself is open since they use and contribute so much FOSS?

They're all still proprietary with the motivations, methods, and limitations that come with it. They can use whatever available components they want to build proprietary software. It's still the dominant model, though. That's why I'm constantly running into systems I have no source for, can't modify, and can't extend to deal with their horrible UX. They're all closed to me even if some OSS or FOSS is inside.


Your OS has a tiny layer of FOSS, which was partly part of NeXTSTEP and what Apple needs for UNIX compatibility.

Everything that makes OS X special is proprietary.


Exactly. The core that was open-source, but not FOSS per FSF, is a system almost no Mac user will directly use or develop apps for. The experience Mac users get is almost totally proprietary by Apple. It's not only theirs in copyright: they patent every aspect they can to sue anyone doing something similar. Samsung will happily attest to that. Mac OS is about as non-FOSS as it gets. Right up there with Windows on proprietary components, lock-in, and patent suits by owner.

At least they're doing a nice, open-source project with LLVM. I'll give them credit where it's due.


Lots of software have been licensed under Apache, BSD and MIT licenses for two decades...why do we need another decade for this doom to arrive and what are the early signs we can see now?


Because LLVM wasn't a thing, nor the ongoing replacement of GPL stack like what Google is trying with Fuchsia.

How are you enjoying the PS 4 features that were added to FreeBSD and LLVM?

Already updated it with your set of changes?


Companies don't contribute code to FOSS projects because a copyleft license forced them to. (If all they need is to comply with the GPL, they just throw an undocumented unsupported code dump onto their website!!!)

Companies contribute code because the upstream project will maintain the code for them!

Sony PS4 is a horrible example to cherry-pick. They probably didn't do any serious modifications to anything, they mostly built their stuff on top. A much better example: Netflix upstreaming their reimplementation of sendfile() to FreeBSD. Because it's been upstreamed, they don't have to maintain it for compatibility with other upstream changes anymore, people who make those changes will fix the sendfile() code.


The contributions of companies back to *BSD versus GNU/Linux is the living proof of what would happen if Linus had not chosen a copyleft license.

We would still be using Solaris, HP-UX, Aix, Tru64, DG-UX to this day.


You should care! To wit, the Proprietary software you use is better for having to compete with Free software.

Free software's 'users are the customers' mindset competes with the 'users are the product' trope that proprietary software can get quagmired in.

Even if the Free software version is a smaller community, the mere fact that a user-centric alternative exists can help prevent this.


I'm looking forward to the followup where they realize "Free and Open" is a better term than "Free" or "Open."


What?

As terms, none is "better" than the other, those are different things. A simple google search will reveal the definitions.


> A month ago I wrote an article called "How I realised “Open Source” is a better term than “Free Software”". Now I change my mind completely and explain why everything I said was actually wrong.

https://www.youtube.com/watch?v=pACePi441ds


This is a bit funny. His argument is that the "Free Software" explanation is longer and thus gives you more opportunity to explain it properly.

However originally the opposite is the case: The FSF explains free software with the four freedoms [1], the OSIs Open Source Definition has 10 points [2]. If you dig into them you'll realize that they more or less mean the same. Which is also usually the point I tend to make in these debates: How you call it isn't so important, important is what it means.

[1] https://en.wikipedia.org/wiki/The_Free_Software_Definition

[2] https://opensource.org/osd-annotated


The main point seems to be that the term "free software" is misleading, so you get to describe its benefits when someone asks you what it is.


Or, more likely, the term is so effective at misleading that 99% of people hearing a reference to it assume it means software they don't have to pay for, and don't ask what it means at all. cf Open Source, which is unlikely to be misunderstood and likely to be the basis for a discussion on why access to source code is important with anyone that isn't familiar with the concept.

The whole article reads like parody tbh. Also like the argument that true software freedom involves not being free to use proprietary software, because politics, and the bit about treating non-free software users as heretics also being important for "a society that stands up for what is good"


My thoughts exactly. I get what they're trying to do with the free software term but they're ignoring the psychology of the masses. Once a definition has taken hold, it's best to use it the way it was solidified or create a new one for a new concept. Bringing up "free," correcting the user that it's about political freedom, politicizing software, and so on will bring up many tangents like how people will get paid or should they.

Open-source w/ its differentiators focuses right on the benefits to the user. I follow-up with real-world examples with things such as car development or repair to help them see necessity of the benefits. Also explain how copyright and patent monopolies reinforce the problems with simplest examples I can get. This approach has worked well where people usually agree with most of it regardless of political spectrum. The next thing they ask is about quality of stuff nobody was paid to write. Replies to that seem to boil down to whether they trust me to assess and tell them the truth. Very impacted by their own beliefs on work ethic and markets.


> The whole article reads like parody tbh

The "I've now changed my mind" bit seems suspicious, like he planned it out from the very first article. I suspect he'll publish a third piece in a few weeks saying neither term sounds OK, then suggest his own new term that "just occurred to him". Of course all 3 articles will have been a marketing strategy for the new term from the very beginning.

I personally like "open software" (without the "source") because it sounds like it includes the important ideas from both "free (libre)" and "open source", but without suggesting anything about a free price. It also suggests an "open process" of development, something lacking in many free(libre) and open source projects.


My thoughts exactly. To a lot of people (if I'd guess, I'd say a vast majority indeed) it just means 'free' as opposed to 'paying money for it'. Especially because used in combination with 'software' of which most people know there is also a paid-monney-for-it version. So no they won't ask what it means. If you want to express something means 'free' as in 'freedom' (and other meanings the article gives to it) then a more effective way to communicate that might be coming up with a term better epxressing that than just 'free'. Thinking of it from that angle, maybe 'open' isn't even that bad.


Think about a cell-phone which has a built-in battery you can't replace. Isn't that bit like software you can't modify?

Now should we require the cell-phone manufacturers to make their batteries replaceable on ideological basis? It probably makes it cheaper to produce the cell-phone if the battery is non-replaceable.

Similarly making your software open-source probably has costs associated with it.

Should we insist that every producer of software must make their product open-source? If it is unethical to produce closed-source software then clearly we should. No?


>It probably makes it cheaper to produce the cell-phone if the battery is non-replaceable.

From the manufacturers point of view, it's cheaper and more profitable. It doesn't help the consumer though, I upgraded phones sooner than I otherwise would have because my battery was dying.

>Similarly making your software open-source probably has costs associated with it.

In the current climate, yes. If all software was already free, you could save a lot of money not needing to repeat work that has been done.

We ideally should insist all software is free, but not everyone agrees that proprietary software is unethical and laws aren't always driven by what is ethical.


> Think about a cell-phone which has a built-in battery you can't replace. Isn't that bit like software you can't modify?

No, it misses the essential point: Software freedom is primarily about power, not about being able to modify software.

If by "you", you mean "noone", that's not really what free software is concerned with.

What free software is concerned with is if there is only one party that can swap out the battery, and that party is not you, therefore making you dependent on a monopoly.

Building software that literally noone can modify (aka "unmaintainable code") might be a bad idea, but for reasons largely orthogonal to software freedom.


What Open Source achieved is that we can now build businesses and consimer products quickly by sharing innovation with few strings attached.

Free software has limited overlap with that particular benefit of Open Source.

At this point in time I think of the two things as widely different so the question does not even come up what's the better term.


True so long as you aren't including GPL2 or GPL3 as open source - hellualot of strings there for anyone with a hardware or software patent, which is why so many billions have been spent replacing those licenses with more liberal (MIT, BSD) ones across so much software.


So how well are *BSD doing in the market in regards to GNU/Linux?

When the only thing standing will be the Linux kernel, don't be surprised by the amount of "freedom" you will get.


The rights granted to the Linux kernel are "stranded" - the kernel can't do anything by itself so licensing to only it doesn't matter, so Google paid to redo the GNU libraries under a BSD license to strand the kernel rights. Even so, with Fuschia Google is replacing the Linux kernel, too, with its own BSD (or MIT) licensed kernel. So, BSD is actually taking the whole field, particularly measured by sales.


The license yes, but not the BSD operating systems variants.

How much do you think companies will bother to give back once they alienated GPL software, given the actual contributions?

Nothing of relevance I tell you, just enough to keep the masses happy, AOSP style.

I won't care, as I said I mostly use commercial software anyway, but it will feel good to say "I told you so".


The answer seems to be that Android, before Google acquired it, took up Linux thinking it was a bit better than BSD alternatives without realizing that the license downsides would inevitably mean it's replacement (starting with the GNU libraries.) That naive view of real license costs was tragic for the BSD operating systems variants and society as a whole, wasting immense resources getting back to a BSD license they could, indeed, have just started with.

Sadly, GPL - because it punishes contributions with patent loss, among other things - has a terrible record in curing freeloading. Yup, I agree it's a real problem. Nope, no very good solution in sight.


> How much do you think companies will bother to give back once they alienated GPL software, given the actual contributions?

Plenty of permissively-licensed software projects have strong contributions from commercial users.


Such as?

The AOSP, BSD and LLVM ones surely aren't such projects.

AOSP doesn't get any of the Google Services stack nor OEMs changes.

The BSD hardly get any updates from all those routers and embedded devices using them, to the point they were forced to beg for donnations last year.

clang is being adopted by embedded vendors that had to contribute their changes back to GCC and now don't need to keep doing it.


> Such as?

SQLite and Postgres are the main examples I think of for this.

> The AOSP, BSD and LLVM ones surely aren't such projects.

AOSP is almost entirely ongoing contribution by the primary commercial user, so it would probably fit the description, though I was thinking of—without having stated—specifically third-party commercial users, so it wasn't what I had in mind.

> clang is being adopted by embedded vendors that had to contribute their changes back to GCC and now don't need to keep doing it.

Clang has a lot of corporate contributors, including all of the big three tech companies. Not sure if any of the big contributors have closed downstream derivatives or if they are just open-source users contributing back. Sure, copyleft prevents downstream use that doesn't contribute back, but it doesn't necessarily maximize contributions back.


BSD licensed projects include the Android libraries (and now Fuschia, too.) Huge corporate investment there from Google. BSD Linux, sadly, is playing the part of the poor orphan, as you say.

For my part I agree freeloading is a problem, but clearly GPL2 and 3 hasn't solved it. It certainly isn't a general solution. True, the Linux kernel gets real support from a couple of companies (principally Redhat) but everyone else freeloads. Remove the patent grabs and maybe you'd see more contributions, but maybe a new form of license is needed, or much better: we need the govt to fund this infrastructure; the present situation is like having volunteers build some highway some of the time while the government sits on its hands. Crazy if you ask me. Of course, Apple and Microsoft (and maybe Google) who want to have everyone on their wholly-owned toll roads would disagree strongly with me.


Patents are cancer.

Also, GPLv2 does not say anything about patents.


Google implicit patent grant. Turns out it does. Just doesn't use the word. Re patents, how would you have James Watt raise money for his steam engine? It took long enough with the patent he got.


I always wish they used the term "liberated" rather then "free". "Liberated software" detaches from prices, detaches from licensing, keeps the desired political associations, and translates fairly well internationally.


I understand where you're coming from. To me, "liberated" implies that the software wasn't free in the past. I think that's one of the reasons for the use of "libre" in some circles.

Now I've got images of a commando raid on a data center running through my head. They're throwing open the doors to the cages and racks. Jerry Bruckheimer, call me. Let's talk.


I suppose so, but I don't see that as a disadvantage of the term, in fact it opens up an interesting aspect of it: all software is, in a sense, imprisoned by default. The author must explicitly "liberate" it for it to be considered free.


I found this Richard Stallman talk to be fantastic (found it here on HN) http://audio-video.gnu.org/audio/rms-speech-arsdigita2001.og...

In it he describes that he searched for better terms than "free" as in free speech not free beer and there were 60ish of them, but they all had problems of their own.

the Free software movement did more to help open source OS's happen because it was it's goal, than open source has done to create free OS's (because their focus was only the specific problem they were solving, and not the holistic many parts of an open source OS.)

Open source is less likely to pass the same licenses and freedom it was built on down to the next iteration, for it's not legally required to do so (which is why whether you got freedom with your X software or not was a function of where you got it from. X software was free, but became unfree if you got it compiled from certain vendors.

Open source documentation is also important, as software wit documentation is better/easier to improve software.

When you use non free software, you advertise that you're ok with not helping anyone else that might want you to, because you can't give them a copy of what you have.

When about half of linux is Gnu and Gnu preceded linux, it's pretty unfair to the Gnu community that they get about 0 credit from many publications, events, the general public.

Thus you can give free software to the world, and they'll never even understand what free as in speech and not beer is.

Free software is a political movement, open source software is more politics agnostic. I think credit should be given where it is due, for only accurate assignment of credit can reward the good and punish the bad.

P.S. Stallman is happy to see you make money on Free software. Free as in speech, not beer. Also the best numbers I could find on what percentage of linux is actually Gnu code is about 8% and linux code was about 9% with the amnt of Gnu going down quickly if you were working on an embedded system. I believe those numbers were from debian and 2014, but I'd be happy to hear a better analysis.


Free software is definitely better than Open Source, but only if you imply that free-as-in-freedom is granted. Otherwise you often get something free-as-in-beer only like a mobile phone OS, which you surely don't have to pay for, but restrict your use of it in many ways and doesn't respect your privacy, all aspects having nothing in common with the concept of freedom. GNU GPL grants this freedom, but the term "free" is often abused to imply free-as-in-beer without the end user noticing the difference.


>> So in fact we must use only 100% Free Software to live the message and spread the message.

For work & daily computing/tinkering, I don't mind ditching Windows completely and use FOSS system like Linux (I mean GNU/Linux).

But leaving those AAA games is... kinda hard. Really really hard...


[quoted as in article] >>DRM is great because it stops people pirating movies.

>It's like saying police brutality is good because it deters bad dudes. If we have to do despicable things to force people to pay for things then our society is broken.

lmao


Bottom line nobody outside of the software community cares, and those in the community already know what "free software" is.


I've had TONS of experience with dozens, maybe hundreds of people professionally involved in software who did not understand free software until I talked with them.


Free Software makes people think of "freemium" software.


Most people just think they get it without paying for it. That's the universal definition of free at least in the U.S.. Attempts to redefine it are foolish. When they talk about freedom, they literally say freedom. I'd have gone with Freedom-Respecting Software if I was, say, trying to tie in freedom with software for a U.S. audience. I have no idea what would appeal to people in other countries.


Other countries have separate words for "free" as in speech, and "free" as in beer. Eg. french has libre and gratuit. German has frei and gratis.


I suspected as much. Thanks for tip.


And both, while noble causes, are unrealistic for most software products.


I would assert the contrary, that the types of software projects that benefit from keeping the source secret are the minority. Most software is written to fill some immediate and boring need, not as a commercial product or business logic secret sauce. There's no real downside to sharing that - most likely no one finds it useful, but on the off chance they do you suddenly have help with it.

Admittedly it's a bit of a prisoner's dilemma; no one wants to put in the (certain) effort of packaging up their software for external consumption for the (uncertain) benefit of external help. But the culture is definitely shifting.


Availability bias.

Most software people see is mass-distributed software: browser, OS, office suite, games… And it is much easier to extract money mass-distributed software when it is proprietary.

In reality, most software that is actually made is custom software. Software that only a handful of people see, because it has only 1 customer, if any. Making that software free makes it easier to extract money from (at least in the short term), because source code availability is often sold for a price (the understanding being, giving the source code also means giving up on exclusivity for later maintenance).


> In reality, most software that is actually made is custom software.

Do you have a source for this claim? (not denying, just curious to look at the numbers!)


I recall Richard Stallman saying custom software comprises over 90% of the effort. I myself has to physically meet one person who is currently working on massively distributed software. (I have met some who have worked in games in the past.)

I don't have anything more precise, but it seems to make sense.


A detailed study clearly shows that you are wrong:

https://www.dwheeler.com/oss_fs_why.html


Like, say, run smartphones or webservers?




Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: