What do you think about the BSD family and its ideas?
- Inicie sesión o regístrese para enviar comentarios
Do you think it was beneficial to free software? What interests me is your thoughts concerning two major free software philosophies, share-alike in case of GNU and Linux and free-to-share in case of *BSD. There seems to be perpetual (friendly) rivalry, and after acquainting myself with *BSD, I see this distinctive philosophy that's not really better or worse, but quite different.
Do you think free-to-share allows private companies to exploit free code? But there are such arguments as this - if you take BSD code without changes and re-publish it as proprietary, I can simply tell you, look, there's the same code available for free, download it instead of paying for nothing. Those who release a proprietary product have to demonstrate its significant new features first, and also show that they somehow can't be implemented with freely licensed code in a much more convenient and productive way, which they often fail to do.
So the stigma of BSD being this proprietary-welcoming mole in the world of free software seems undeserved. (Maybe I don't know it well enough?.. :) What are your thoughts?
Theo de Raadt is recipient of the 2004 Free Software Award. OpenBSD is great.
https://www.fsf.org/news/fsaward2004.html
For recognition as founder and project leader of the OpenBSD and OpenSSH projects, Theo de Raadt's work has also led to significant contributions to other BSD distributions and GNU/Linux.
heh, check out this.
https://www.openbsd.org/lyrics.html#43
> heh, check out this. https://www.openbsd.org/lyrics.html#43
Even RMS has advocated the use of permissive licenses (in the case of
Vorbis) or the LGPL (still copyleft, but more permissive than the GPL)
when doing so is strategic. Distributing cross-platform free software
is also strategic, because it facilitates migration to free software.
One might argue that distributing blobs for better hardware support can
be strategic, but that doesn't have anything to do with copyleft. If
OpenBSD think that their approach to licensing is strategic, I wonder
what their goal is. If their goal is to get people to use OpenBSD,
their strategy is clearly not a good one, because nobody uses OpenBSD.
If their goal is to have their software distributed under licenses *far*
more restrictive then the GPL, then their strategy is (unfortunately) a
good one, but makes them hypocrites for whining about copyleft being too
restrictive without condemning even more strongly the proprietary
licenses under which their software is often distributed.
> Do you think it was beneficial to free software?
Overall, I think it probably was, but only because their work has been
beneficial to other free software projects, not because of BSD itself.
There are free versions of BSD, but virtually no one uses them. The vast
majority of BSD users run a proprietary, locked-down version on their
iPhone or Mac.
> Do you think free-to-share allows private companies to exploit free
> code?
Yes, but I don't think "free-to-share" is a good term for permissively
licensed software. Copylefted software is also free to share, just not
free to turn into proprietary software. "Free-to-not-share" would be
more accurate.
> But there are such arguments as this - if you take BSD code without
> changes and re-publish it as proprietary, I can simply tell you, look,
> there's the same code available for free, download it instead of
> paying for nothing. Those who release a proprietary product have to
> demonstrate its significant new features first, and also show that
> they somehow can't be implemented with freely licensed code in a much
> more convenient and productive way, which they often fail to do.
This sounds plausible in theory. Unfortunately, things have played out
differently in reality.
> So the stigma of BSD being this proprietary-welcoming mole in the
> world of free software seems undeserved. (Maybe I don't know it well
> enough?.. :) What are your thoughts?
"Mole" is harsh, but I think "proprietary-welcoming" is fair, especially
when they create permissively-licensed reimplementations of copylefted
software. Users and free software developers don't gain anything from
this. It only benefits proprietary software developers.
This is not to say that copyleft is always enough to prevent free
software from being exploited (GPLv2's tivoization loophole comes to
mind, as does the ongoing corporate takeover of Linux at a political
level) but it's better than nothing. In addition to BSD software, Apple
also includes some GNU software in their operating systems, such as
Bash. However, the versions of GNU software that Apple ships are more
than a decade old. This is because the GNU project has switched from
GPLv2 to GPLv3, and Apple would rather ship outdated software from
before the license change than stop exploiting the tivoization loophole.
This gives the developers of free operating systems, who can use the
latest versions of GNU software, an advantage over Apple. This advantage
would be much larger though, if BSD software also used copyleft
licensing to ensure that their code stays free.
Fair enough.
"Free-to-not-share" would be
more accurate."
Hah, that's spot on, I see your point. But BSD people may say, freedom is freedom, share it or not, that's your choice and your consequences. If code by its nature is infinitely copiable, then why should anyone have a problem with some people using it however they see fit? It will not harm or exhaust original free code in any way.
"Unfortunately, things have played out
differently in reality."
Why do you think it failed in practice?
If code by its nature is infinitely copiable, then why should anyone have a problem with some people using it however they see fit? It will not harm or exhaust original free code in any way.
If the proprietary program has attractive features the permissively-licensed base has not, then that base will lose the audience that values features over freedom.
That said, it is not my reason to prefer copyleft: I simply do not want my work to help subjugate any user, i.e., to end up in proprietary software (regardless of its success).
Then how did BSDs develop over time into quite robust and venerable (and mostly free) operating systems they are today?
"If the proprietary program has attractive features the permissively-licensed base has not, then that base will lose the audience that values features over freedom."
Well, we must presuppose that there exists this mythical audience of obstinate pragmatists who always value features over freedom, despite the fact that all BSD communities are very DIY-oriented, even more so than Linux distro communities, who are much more commercialized nowadays. So I don't find this line of reasoning convincing.
> "If the proprietary program has attractive features the
> permissively-licensed base has not, then that base will lose the audience
> that values features over freedom."
> Well, we must presuppose that there exists this mythical audience of
> obstinate pragmatists who always value features over freedom, despite the
> fact that all BSD communities are very DIY-oriented, even more so than
> Linux distro communities, who are much more commercialized nowadays. So I
> don't find this line of reasoning convincing.
You aren't taking into account that the most popular BSDs, by a long
short, are iOS and macOS.
Then the most popular Linux-based OS is Android :)
It is, indeed. But its kernel (including all drivers, not firmware) is free software, under the GPL. I am aware that some manufacturers violate the license and include proprietary drivers. They could be sued for that. On the contrary, permissive licenses allow proprietary derivatives.
> Then the most popular Linux-based OS is Android :)
Yes, that's probably true, assuming that by "Linux" you refer to the
kernel, and not what most people refer to when they say "Linux" which is
an operating system which uses the Linux kernel and GNU userspace.
(Android uses the permissively-licensed Toybox as its userspace.)
Android vs iOS is a good example of why copyleft is better for software
freedom than pushover licenses. iOS uses only permissively-licensed
software, and turns it into a completely proprietary operating system.
Android uses some copylefted components, and therefore must share their
source code for these components under a free license, resulting in a
partially free, partially proprietary operating system. The free parts
are the base of Replicant (a fully free mobile OS) and LineageOS (free
except for some firmware needed for certain hardware). If Linux were
under a permissive license like BSD, Google could have made their entire
OS non-free like Apple does, in which case Replicant and LineageOS
wouldn't exist.
Yes, I do refer to the Linux kernel. Nonetheless, Android is the most popular OS whose core component (kernel) is copylefted under GPL.
> Yes, I do refer to the Linux kernel. Nonetheless, Android is the most
> popular OS whose core component (kernel) is copylefted under GPL.
Exactly, and accordingly, Android's kernel is free software, whereas
iOS's is proprietary software.
> There are free versions of BSD
Are there? I don't see any listed on the GNU website's free distro list. AFAIK every existing distribution of BSD includes binary blobs or other proprietary components of some kind. If I'm wrong about that, I'd love to know, as I've always thought about giving BSD a try.
> > There are free versions of BSD
>
> Are there? I don't see any listed on the GNU website's free distro
> list. AFAIK every existing distribution of BSD includes binary blobs
> or other proprietary components of some kind.
LibertyBSD removes all firmware blobs from OpenBSD,[1] so it is a free
distro in the sense that it is entirely free software (which is the
sense in which I meant it). That does not necessarily mean that it meets
all of the FSF's criteria for endorsement. The FSF requires much more
than only including free software. Discussion of licensing and firmware
blobs only make up about half the text of the FSDG.[2]
[2] https://www.gnu.org/distros/free-system-distribution-guidelines.html
"Do you think it was beneficial to free software?"
The lawsuit was but this seems to rewrite some history and attribute different philosophies than what was there. Please don't forget that UNIX was originally proprietary. It wasn't started as a thing to make a free operating system or with a "free-to-share" philosophy that seems to be getting imbued here but instead for something to run internally at AT&T. That it later became free software was a side effect over a corporate lawsuit between AT&T and the University where AT&T agreed to settle the case in response to the defenses that Berkeley raised instead of continuing to push the issue.
Moving on, the common thought process at the time was to write operating systems in assembly language which made them unportable. This became a problem that RMS ran in to previously with the Incompatible Timesharing System (ITS), which was written in PDP-10 assembly language. This operating system essentially died when DEC discontinued the PDP-10 computer because it could not be easily moved to another type of computer.
Making a free operating system (which was what RMS set out to do) was going to take a long time and there was no way to know what computers would be available in the future. RMS didn't want to end up in the same situation as before. The only connection with the free software movement was that RMS thought UNIX would be a good design to emulate because of the portable aspect so as to avoid a repeat of what had happened with ITS.
So: The lawsuit over UNIX itself turned out to be beneficial to free software but that was only a side effect of two organizations trying to fight out ownership and not because of any particular "philosophy" regarding sharing.
Frankly, I simply took "share-alike" and "free-to-share" expressions from E. Moglen's website, who described them in this way. Personally I know little about these intricacies, that's why I asked.
Stallman created GNU in response to the "freeware" lawsuits of his day in the 70's and 80's, which were threatening to bog down "free" source code projects eternally in courtroom battles. To my mind, BSD's permissive license model falls on the wrong end of that spectrum. As jxself pointed out, BSD's lawsuit history is very instructive in this regard.
But permissively licensed giants like Apache are alive and well, despite their licenses being similar to BSD.
The key question to ask in any discussion about freedom is this: who is free to do what?
Software freedom is based on the proposition that there ought to be a balance between the freedoms of computer users to use their devices in any way they see fit, and the freedoms of software developers and distributors to create and modify software. The principle could he expressed in the old saying that "your freedom to swing your fist ends where my face begins". Software developers freedoms are important, but their freedom ends just before the point where it starts to take away the user's freedoms.
The "freedom" BSD folks talk about is essentially the freedom of developers to swing their fist wherever they want, even if users' faces get in the way. Even though the practical consequences are very different, defending the "freedom" of developers to make their software proprietary, or to allow others to do so, is logically equivalent to defending the "freedom" to own slaves. It's a "freedom" that cannot be exercised without taking away someone else's freedom.
I wrote a piece explaining the ethics and practicalities of copyleft vs. laissez faire licenses in more detail back in 2012:
https://www.coactivate.org/projects/disintermedia/blog/2012/08/10/what-share-alikecopyleft-means-for-you/
"is logically equivalent to defending the "freedom" to own slaves."
Is it? I think being a slave is a status, whereas using a program with a certain license means you voluntarily accept a contract, you're free to not accept it as well.
Bear in mind that not every BSD follows the same philosophy. No one would put ZFS under OpenBSD because of CDDL licensing reasons, ever. If any, they'd choose HammerFS as the new FS instead of UFS.
Yes, they have firmware blobs, but they consider those as "pluggable" EEPROMs", as if they weren't software.
And well, you won't have the Intel/AMD blobs installed on base, among some wireless and webcam drivers for licensing reasons, too. It's up to you to get those.
They are even against the LLVM license change (> 8.0) because on how the Artistic License can deal to some issues with potential forced patents.
Ironically, they take the potential freedom better than most GNU/Linux distros full of propietary software.
But I'd take OpenBSD as an technology standards defining OS core full of great and libre technologies such as OpenBSD HTTPD, tmux and OpenSSH, and for free standard reasons even RMS recommended a BSD license instead of the GPL. For example, libre video and audio codecs, network protocols and so on.
"Ironically, they take the potential freedom better than most GNU/Linux distros full of propietary software."
maybe because they develop a separate, well-established operating system, and not a 100..th sketchy "Linux distro" put together by **** knows whom. Recently, people have been talking about how MX Linux became popular because of bots. Older and more experienced GNU/Linux users simply don't know what MX Linux is, whereas new users install it due to its vague "popularity". Not sure if it's wholly true, but it's telling. I for my part have no idea what MX Linux is and how come it is "the most popular Linux distro", all of a sudden. If it's so big, then why is its community almost non-existent outside of its forum?
- Inicie sesión o regístrese para enviar comentarios