> For one thing it hasn't failed, the free software movement has been an amazing success.
It has failed.
Open software hasn't failed. But thats not the same as free software. Sure there is a lot of open and free software but there is soo much more open software which is not good for the free software movement. Weather that is because of it being open but not free or weather it's open and free but tightly controlled in ways which hurt the free software movement in the bigger picture.
You are measuring success in terms of specific license use. Or authorship. And by those measures its easy to say the FSF has failed.
If however you look at software as a whole, and compare today to say the early 80's, then you might note that the landscape is very different and wonder how that happened.
Today software exists unler a multitude of licenses, written by innumerable authors. A huge fraction of all software is either open, or built on open.
While Open Source and Free Software are different idealogically, they ultimately, broadly, confer the same freedoms on users.
That any of this exits _at all_ is thanks to the FSF - they won the war a long time ago. They are the very definition of success. A success that so many take for granted that we now wonder if they are useful anymore.
>While Open Source and Free Software are different idealogically, they ultimately, broadly, confer the same freedoms on users.
i don't know... i could be way off base but MIT and BSD allow devs to FORK the code and turn it into proprietary code throwing all benefits of "open source" into the wind. "enterprise friendly" is often mixed up but the reality is the end user ends up on the short end of the stick..
Hence the word "broadly". MIT etc allow for Open Source code to be used in proprietary software, true.
Whether this is good or bad, depends on your point of view, and is very newanced.
For example I can use OPENSSL (MIT licensed) in a proprietary system, but I would argue that is better for end users than not. Having a few solid libraries in play is arguably better than a multitude of home-grown closed commercial libraries.
OSI is clearly not FSF. FSF persues a future with no proprietary software. That's a necessary stance for someone to have in the world, it sets a goal post. OSI seeks to make code that has maximal utility by being useful in open and closed environments.
MIT does not prevent companies contributing code, it does not hold back a project in any way, it just makes the code more widely spread and more "utilitarian".
Edit: so it does not eliminate "all" the benefits of Open source, it may remove some of them. Alternatively it may be a net gain if the software is actually used more.
Technically, yes, a company can fork OPENSSL, add features to it, and ship it closed. But why would they? Certainly if the did they remove no utility from openssl for everyone else.
Equally SQLite has a public domain license and that does not seem to have hurt it in any way. There are plenty of contributers etc.
> I can use OPENSSL (MIT licensed) in a proprietary system
You can use LGPL libraries in proprietary software.
> MIT ... makes the code more widely spread and more "utilitarian"
Not for the end users and for developers that want to modify existing software. The "utility" is lost.
> a company can fork OPENSSL, add features to it, and ship it closed. But why would they?
It happens every day. Weak licenses allow freeloading.
> Equally SQLite has a public domain license and that does not seem to have hurt it in any way.
That's not the right metric. The people who are being hurt are developers and users who cannot buy a truly open phone/smartwatch/router/tv/car even if such devices are built on OSS.
yeah, a dev worked for no financial incentive other than the feel good of "open source" while doing MIT code and someone comes in, takes that code, slaps their proprietary tag on it and suddenly the same dev cannot get the benefits of "open source" even when he was doing the work for the greater good.
on the same vein, the same dev works on GPL code and he/she knows the virality of GPL makes it that no one can slap proprietary license on any fork of his work and he/she, if they became the customers of any forks, they CAN expect and demand source, completing the feedback loop.
That same dev _chose_ the license they wanted to use.
Some choose GPL because they value freedom over utility. Others choose say MIT because they value something else.
Naturally all open licenses (free and open source) allow anyone to take that code, compile it, and sell it. This is not a bug, this is a feature.
If the one selling the code makes changes to it, or augments it in some way, and assuming they follow the terms of the license, this is all good. The original author is getting exactly what they hoped for since that's why they chose that license in the first place.
In exactly the same way users make a choice about what software to use. They can choose something proprietary like iOS or something more open like Android. They can choose to use Android coupled to Google services, or not.
I get the goal that the FSF has of making users care enough to only use Free software. That is a noble goal, and we need a organisation that sets that goal.
Equally with hardware they can, and should, champion the cause of completely open hardware.
They are successful not just when all hardware is open, they are successful by making it part of the conversation. It took software 40 odd years or so for Open to become mainstream. It'll take some time for the hardware case to reach some form of critical mass.
All correct except the word "virality". There is no such thing. People can use [A][L]GPL software according to the license or not use it.
Software licenses, and any other work under copyright law, cannot "infect" other works. It's not Covid.
If I draw Mickey Mouse on my hand I might be in breach of Disney's copyright and I might be required to wash it away. It does not mean that my hand magically becomes Disney's property.
And as far as I can tell their methods are subpar for today. This in no way means that they didn't work well in the past. Or didn't do a lot of good things in the past. But times change.
And I'm also not saying they are not useful today anymore, I'm saying their methods don't work well anymore and could lead to them becoming increasingly irrelevant and powerless which would be a loss of all consumers and most developers.
This comment is splitting hairs that don’t even exist. I have never once heard the term “open software”. The closest match is Open Source. The OSI definition for open source is exactly the same as the FSF Free Software.
Another adjacent term is “Source Available” for software that isn’t open source but you can get the source, like Unreal engine.
There is also Open Core for software that has an open source part and a proprietary extension. Like GitLab.
I intentionally didn't use a well defined term as it would miss the point.
There is all kind of somehow "open" software which might be OSI open source or even be GPL licensed but is de-facto not free software.
Like I have seen more then one case where GPL or similar was used to archive "open sourcing the software" while "making sure no on complying with law can benefit relevantly from it". E.g. publish a software under GPL which is strongly dependent all through the stack on some proprietary software which is GPL incompatible but has some pseudo decoupling though some interfaces. The company which wrote the software just uses a different license when they use it. Everyone else would have to spend impractical amounts of time to create adapters to make it actually usable under the GPL license. Sometimes not much less time then rewriting it from scratch. I have seen this pattern quite often. Through more often with AGPL then with GPL. Sometime it did feel that such theoretically but not practically free software likes (A)GPLv3 more the actually free software does like it (but I'm pretty sure this is a bubble).
> The OSI definition for open source is exactly the same as the FSF Free Software.
It is not. If it is, you should explain that to every other expert and lawyer involved in litigating the differences, because they're operating under a delusion.
They are mutually inclusive licenses, because OSI licensed stuff can always be relicensed to GPL (or anything else, really), and because OSI says that GPL is also OSI (although I don't understand why.)
But they're different. You have to share your changes in GPL if you distribute them. That's it. That's the difference.
Not all free software is copyleft, and not all copyleft licenses are GPL compatible. But all free software is open source and vice versa. The four freedoms and the open source definition are intentionally equivalent.
The OSI doesn’t define their own licences. They set out a bunch of rules about what open source means and then list which common licences fit those rules. Their rules are basically the same as the FSF essential freedoms list. Which is why GPL software falls under the OSI definition of open source.
It has failed.
Open software hasn't failed. But thats not the same as free software. Sure there is a lot of open and free software but there is soo much more open software which is not good for the free software movement. Weather that is because of it being open but not free or weather it's open and free but tightly controlled in ways which hurt the free software movement in the bigger picture.