The title will imply to some that these two things are related. You can sort of relate any two things by making a comparison, but here it is incidental. I want to talk about both, but comparison isn't the primary goal.
"I have done a lot of work, migrating my equipment and files and getting familiar enough with BSD."When I think of BSD as an alternative to GNU/Linux, I don't present it as a solution but as part of a solution. BSD is not as free as it could be.
I have done a lot of work, migrating my equipment and files and getting familiar enough with BSD. On some level I had a bit of familiarity already due to using the command line on a Mac, when I was doing some volunteer work that was entirely unrelated to Free Software or even Open Source. They had Macs, Macs have command line tools from BSD -- I am not utterly surprised when I have to add a path -- such as "." when I use find, for example. It still does get me occasionally:
# find | grep [...]
usage: find [-dHhLXx] [-f path] path ... [expression]
# find . | grep [...]
[Okay.]
"Relying on a minimal distribution like Tiny Core has made migration much easier, and that's a feature."At first I thought you needed to EDIT the disklabel every time you wanted to access an ext2fs partition. I was pretty much ready to give up on the human species at that point. But you don't have to, and I would rather change the partitions than edit the disklabel, but you are really going to HATE when you have an ISO image written to a thumb drive and a partition created after that.
Relying on a minimal distribution like Tiny Core has made migration much easier, and that's a feature. Debian may have at one point been a Lexus as far as distributions go, but that is harder to replace than a Nissan Leaf -- at least I thought that metaphor would work, until I looked up the list prices and found the Lexus CT is actually comparable to the Leaf. Go figure.
Only readers who are old enough (or old fashioned enough) will get any reference I make to the Intel 486 sx.
"It is, to be sure, incredibly refreshing to be able to comment on the state of GNU from the outside."Either way, in just over two weeks I have moved myself into an interesting situation: if I wanted, I could be Linux-free (as in not running the Linux kernel on anything I use) within 24 hours. That's not a goal yet, I've been trying to have it as a real option (I now do) and BSD is what I'm doing most of my computing with now. ZERO of my workflow is actually dependent on my GNU/Linux machine. I could move the files off it and run BSD on it instead, it would probably take an hour or two. Installation can be done in about 15 minutes.
This is incredibly refreshing! Linus Torvalds could throw himself off a bridge, GKH could continue selling the kernel off to Microsoft (it's not like he's the mastermind of this, he's really just a guy they approve of -- but that's sort of dubious enough of an honour itself) and the GNU Project could finish migrating to Microknauth GitHub -- by the way, did anybody ever ask which direction GNUstep was a step in?
It is, to be sure, incredibly refreshing to be able to comment on the state of GNU from the outside. I am not typing this on GNU/Linux, but this screen / setup looks just like it did a few weeks ago -- dwm, xterm, tk-based editor -- I am very sadly running Firefox (ugh) instead of IceCat. I miss IceCat, but the Hyperbola team does have its own browser project which in theory (in all seriousness) is better.
"Of course everything GNU (BSD has similar problems) is GitHub-based, from Automake (Perl) to linux-libre (moving to Perl and Python away from awk and Bash) to GNU libc, to zlib1g."I would love to be using that right now -- I don't love the Web anyway (the whole thing is bloatware) but running Firefox makes me HATE using the Internet. I'd rather have constipation than Firefox, but although I've made a couple of sacrifices -- overall, migrating to an entire OS that gives me a little hope for Free Software is better than most applications. I tried NetSurf from ports, but it is completely unusable -- it just sits there, mostly unresponsive, mostly incompatible with anything. If the only thing I wanted to do online was read Techrights, I still couldn't do it with the BSD port of NetSurf.
Firefox brings in dbus, of course. And it refuses to run unless /etc/machine-id is valid. UGH. This isn't an application I would wish on anybody. But the browser issue on GNU/Linux isn't a lot better, it's a little better -- IceCat is a small bandage for a serious wound. It's built on Rust and Jasmine (and HarfBuzz) from Microknauth GitHub, and that's not likely to improve.
Here's stuff you won't have to compile to try out, if you want to run real GNU stuff on BSD: GCC, Wget and Nano and Bash are all GNU tools, available from ports in OpenBSD. Of course ksh is standard, they prefer permissive tools. You won't change OpenBSD, but they won't try to change you either; you can make your own fully-free BSD with GNU Bash and Wget and everything in the GNU Project that you can compile, but I have removed less, lesskey, zless and tmux because they are GitHub-based: http://cvsweb.openbsd.org/src/usr.bin/tmux/alerts.c
"There is logic when it comes to boycotting GitHub, though a lot of it comes down to avoiding it whenever possible."Of course everything GNU (BSD has similar problems) is GitHub-based, from Automake (Perl) to linux-libre (moving to Perl and Python away from awk and Bash) to GNU libc, to zlib1g. GNU Wget actually has code from Fakebook's GitHub (zstd compression library). Hooray. Curl on the other hand, is based entirely on GitHub itself, so that's not really an alternative.
I do have a sort of work-in-progress criteria for what I boycott when it comes to GitHub, since the only way to really avoid it entirely is to create a completely new operating system. With BSD, that is more possible than it is with GNU/Linux -- nobody has demonstrated a will or an ability to fork Linux, while several people (Hyperbola being the one FSDG-respecting example) have forked BSD with surprisingly small teams of developers.
There is logic when it comes to boycotting GitHub, though a lot of it comes down to avoiding it whenever possible. I'd really like to do better than that, but that's where we are at the moment. Without a fork of Perl and a fork of Python, the GNU Project will never be GitHub-free. Nor will GCC or Clang. GitHub is a terrible monopoly. Speaking of, here is a project Google doesn't want you to know exists: https://reverseeagle.org/
I actually couldn't find it with Google. I tried about five ways, some of which should have worked -- I had to use the not-very-privacy-respecting DuckDuckGo (found it on the first try) to get the URL. I had no idea they had their own .org domain.
"The nicest thing about using BSD though, is experiencing a group of people doing EXACTLY what they claim be doing."I'm not suggesting that YOU can't find Reverse Eagle with Google, only that five tries wasn't enough. This project: https://codeberg.org/ReverseEagle/DeGoogle-FOSS is probably the reason, and it's one of the cooler things they do. Props and thanks to Derek Taylor for featuring it on his video series. This page: https://developers.reverseeagle.org/replace/github/ is the one I really wanted to share with you. But there is a lot more going on at Reverse Eagle, and I hope more people will notice them.
The nicest thing about using BSD though, is experiencing a group of people doing EXACTLY what they claim be doing. Sure, OpenBSD has non-free firmware. And I will summarise its lead developers position on it -- and I know that Hyperbola will address this issue.
Unlike FreeBSD, the lead developer of OpenBSD (Theo de Raadt) does not allow any non-free software (binary-only or NDA-stifled) in the main development tree. He makes an exception that Hyperbola won't, for what we Free Software types call "non-free firmware".
"ROM firmware (read-only) means you can't change it without replacing the chip."His position is that firmware does not run on the CPU at all, it runs on the device itself. The alternative in many instances, to non-free firmware -- is ROM firmware. Once it is read-only and cannot be written to (perhaps due to a fuse making a setting permanent in the chip, which is a technology that already exists https://puri.sm/learn/intel-me/ in consumer hardware) the FSF takes no issue with it. This is a very strange position rms has, and from a position of figuring out what to boycott -- I think I sort of get it. But to most people it sounds pretty ridiculous.
ROM firmware (read-only) means you can't change it without replacing the chip. Fuse-set firmware (now read-only) means practically the same thing. Only when the firmware is still writable is it practical to complain about whether or not you can change the firmware. Right?
It's not my intention to misrepresent the position rms has on this -- it's been years since I've read about it.
The position of Raadt (I recently read an email he wrote about it) was this:
Wait -- so you're saying that proprietary firmware on re-writable firmware is LESS FREE than proprietary firmware on a ROM chip you can't change at all?
"You actually lose the ability to fix something with software only."This position blurs the distinction between a purely practical mindset and a freedom-respecting philosophical position. In the short run, ROM firmware is much more useless to someone who wants free firmware than rewriteable firmware with a non-free blob on it. Also (according to some in the email thread) "blob" has a slightly different meaning in the security world, and OpenBSD is written by people for whom "blob" is a different word than the one people from the FSF use. It's simply an opportunity for misunderstanding. (Note also that "Open" as in "Open Source" was allegedly coined in 1998, but OpenBSD was called OpenBSD a couple years prior).
In the short run, based on de Raadt's explanation -- the Orthodox position on firmware for the Free Software makes roughly ZERO sense. In the short run, it actually presents a setback: from firmware the user doesn't control that you ideally want to replace, to firmware the user doesn't control that you CANNOT replace using only software (because it is no longer writable). What's the advantage again? None -- zero. You actually lose the ability to fix something with software only.
But from the standpoint of a boycott, boycotting ROM firmware is a waste of time -- you can't change ROM at all, so there's no gain in demanding it. But boycotting non-free firmware is (to replay the official position on drivers instead of firmware) the only way to get manufacturers and/or driver authors to care about supporting Free Software. It's slow-going and was sabotaged by "Linux" and Open Source, but progress is made from time to time.
So only in the long run does the completely counter-intuitive idea that looks a lot like "ROM is better than rewritable and non-free" actually make any sense. It makes sense to an activist in the long haul for real progress, but not to someone who knows (correctly in fact) that in the short run it won't achieve anything. In the short run it's actually worse, because you can't fix it!
"So only in the long run does the completely counter-intuitive idea that looks a lot like "ROM is better than rewritable and non-free" actually make any sense."The positions of rms are not religious, as some Microsoft "Evangelists" (their own term) would have you believe, but they are orthodox. And with the attacks from Open Source, I will point out my own defense of orthodoxy:
Orthodoxy (even without violence -- orthodoxy With violence is correctly referred to as extremism) is a pretty effective way to preserve most of a culture for a very long time, as society around it changes. Ultra-orthodoxy on the other hand, frequently becomes isolated and cut off from society.
What bothers me about ultra-orthodoxy is that (in common with cults) it requires essentially magical thinking to justify even simple daily activities that the ultra-orthodox have to practice. This can be very tricky to define. For example: every Shabbos, Orthodox Jews are forbidden from kindling a fire from sundown to the next sundown. This is interpreted by rabbinic authorities as including the operation of light switches, but since the prohibition only applies to Jewish people, anybody else (non-Jewish) is free to operate lights for them.
"Activism does mean putting aside what the world wants (sometimes) for what it ideally should be like."It's difficult to define exactly where the line is or should be, but I actually put that custom on the Orthodox side of things, rather than the Ultra-orthodox. If you're an Orthodox Jew, this sort of thing is very basic. If you're a Free Software advocate, you already know that software under an NDA isn't freedom-respecting. The rules are complex, but they don't require sophistry or mental gymnastics to justify them -- they can be explained and justified deterministically.
Ultra-orthodoxy takes things another leap in some direction. It is more serious than orthodoxy, and justification becomes more contentious and authoritative. There will be traditions that the Orthodox are literally barely capable of understanding or relating to. Not everyone feels that ultra-orthodoxy is oppressive, (I used to have an acquaintance who was happy with it) but this is no safe measure of whether it is oppressive or not.
"Free Software is idealist, and if it ceases to be idealist then it ceases to be about freedom."What's most interesting (and relevant) about this is that orthodoxy survives (and slowly evolves with the world) without losing its Orthodox status, while ultra-orthodoxy does flips and lives in isolation for the sake of avoiding any change whatsoever. And my point here is not simply that we should "just relax" and compromise for its own sake, or just to make things easier on ourselves. Activism does mean putting aside what the world wants (sometimes) for what it ideally should be like. Free Software is idealist, and if it ceases to be idealist then it ceases to be about freedom.
But the threat of ultra-orthodoxy is customs that we can only justify with authority, sophistry and magical and non-deterministic reasoning. I'm not against you thinking magically, though I am more or less against magical thinking being IMPOSED on people -- I'm against theocracy for similar reasons.
There are actually very few things about Free Software this will apply to -- most of the logic that causes us to do what we do is pretty straightforward, and most of the responses Free Software has to problems are easy to implement without sophistry or philosophical contortion.
"If a project wants FSDG status, that imposes certain restrictions on what can be discussed on the project's own forums."One corner we could be painted into though, is the horribly-named FSDG, or Free Software Distribution Guidelines. Despite being familiar with both acronyms, I spent more than 45 minutes today thinking I was talking with someone about the Debian Free Software Guidelines, because if you move the "D" in FSDG twice to the left, it becomes DFSG -- the precursor to the Open Source Definition (Debian still uses the DFSG, it has for many years). But 99% of my complaint isn't about the name.
The FSDG goes farther than the Free Software Definition, but that isn't all it does. It extends to matters regarding documentation (mostly for our benefit) and things you might think of as outside of software itself. Not that OpenBSD (yes, we are making comparisons even if they're not the main point of the article) treats documentation as fully outside the creation software -- a feature without documentation is considered a bug in the OpenBSD world. There are some differences between these worlds, but OpenBSD is more freedom-respecting than FreeBSD (which does not forbid software under NDAs, and which develops its package manager on GitHub).
My original problem with the FSDG was the effect it had on speech itself. This doesn't loop back around immediately to forbidding an entire CATEGORY of 4-freedoms-respecting actually-free software, but it does get there (and that is the main inspiration for this article).
If a project wants FSDG status, that imposes certain restrictions on what can be discussed on the project's own forums. I've always been told I make too much of this (great if that's actually true) but I think the practice stands on its own for demonstrating its effects. My original complaint was that there are instances which go beyond "free advertising" for non-free software (I can certainly understand why that would be considered spam on a Free Software forum) but there are legitimate discussions of non-free software, particularly for people who are interested in creating, locating or promoting free alternatives.
The FSDG is in my opinion, stifling in this regard -- but with just the right forum leaders, and just the right amount of interaction from the FSF (they are not strict enforcers about this -- I think they mostly impose the requirement and then leave it up to the project) you could get by talking about software in a reasonable fashion. I don't like it though, because freedom doesn't love rules that are poorly conceived and sloppily implemented (except when they fail, of course).
But there is a "big picture" concern that in my opinion, puts the FSDG squarely in the category of ultra-orthodoxy I've been talking about.
"This is another really refreshing thing about running BSD, particularly OpenBSD -- I strongly feel that it is the most UNIX-like OS (in terms of the Fifth Freedom) that currently exists."We want all software to be free software. That's the goal of the free software movement. I also want control over MY computing. That's one of my reasons for using free software, and it's commonly noted as a goal and benefit -- for the user to have control over their computing.
Of course I acknowledge that when we say "freedom" you have to be specific for it to be a meaningful conversation; as usual, I am referring to the 4 freedoms in the FSD. I would say the 4 freedoms plus the freedom to NOT run (or to boycott) the software. I've been promoting Peter Boughton's Fifth Freedom ever since he wrote one that fit what I was looking for.
This is another really refreshing thing about running BSD, particularly OpenBSD -- I strongly feel that it is the most UNIX-like OS (in terms of the Fifth Freedom) that currently exists. I hope the Hyperbola devs notice, because they can keep that going as they make a fully free version. In some ways they do seem to notice -- Hyperbola has a good track record for removing dubious components from GNU/Linux.
The thing is, GNU was possible in part because the Fifth Freedom already existed. It was much easier to replace component after component of UNIX to create GNU, by virtue of the fact that UNIX was so modular already. Don't like it? Don't keep it. Want it gone? It's easy to remove.
"There are too many efforts to try to dictate or "nudge" what software I have running or installed, and I'm not okay with that."Since pedants will start telling you how dependencies work because it's an easy straw man, thanks -- I know dependencies, I've written scripts to remix software distributions in my own language and even that (the language) has dependencies, I think most people get that.
But the encroachment (the takeover) of dependencies is the issue here, and it's been talked about enough that we don't need to make the article even longer by explaining it.
I want to be able to boycott software -- I want to make it easier to boycott software. I want developers to (within reason, and not necessarily a heck of a lot more than they ALREADY did about 10 or 12 years ago) respect the fact that I want to remove software from my computer that I don't need or want.
There are too many efforts to try to dictate or "nudge" what software I have running or installed, and I'm not okay with that. I am not asking for something new, either -- I was content with the level of modularity that existed just half a decade or so ago. Things have gotten ridiculous. We might have to go back a little farther than 5 years if we really want to fix it, but this is not a theoretical level of control the user has -- it is truer to say this is an established level of user control that we have now lost.
And when I delete a bunch of files, I want to be able to share that with other people who have similar goals. Maybe only a few will care, but that's not the point. It's free (as in FSD) software that may do something like... take a non-DFSG GNU/Linux distribution and automatically turn it into one that fits all the wild requirements of the DFSG.
But that software will never ITSELF be DFSG-compliant.
So that means it has to be separate from ANY DFSG-compliant distro.
And it means (to the letter, at least) that you can't even talk about that software.
The first rule of Free Software Distribution Club,
is you DO NOT talk about Free Software Distribution Club.
"Open Source always wants to change the rules, so they can get more non-free garbage into everything you do."This is the most likely solution, because that's how the FSF tends to work anyway.
2. Change nothing -- if you want to liberate 100 free software projects at once, you can simply go to the same manual trouble Alex Oliva goes to for linux-libre: TIMES 100.
3. Make the rules saner / improve the process
I LIKE THIS ONE! PLEASE SIR, CAN WE HAVE SOME MORE?
Open Source always wants to change the rules, so they can get more non-free garbage into everything you do. I get it. We have to be CAREFUL. This is like changing code that's part of a mission-critical system; you have to sometimes, but you want to avoid it and you want to be extremely careful when you do at all.
But there are precedents as well. For most Free Software, the GPL is designed to not allow linking by non-free software. In some instances where it made enough sense, the FSF has created the LGPL instead. It rarely recommends its use, but the important thing is they went out of their way to allow exceptions when it made enough sense to do so.
The GPL3 is in a few ways that make VERY good sense, stricter than the GPL2 in what it allows to be taken away from the user (less) but it also changes a few rules from GPL2 that are harsher for the user. Check the GPL3 FAQ for details. The point of this comparison is that not only does the FSF sometimes make exceptions when it's reasonable (as with the LGPL) but they also sometimes tweak rules to make them easier, not only to make them stricter.
The FSF has the authority to make a beneficial change to the FSDG. The benefit of the change is already (in my opinion) in the spirit of the FSDG. Although it may at first seem impossible to change the FSDG along these lines without defeating the purpose -- I can hear the parrots sqwaking already that the entire FSDG would collapse "like a flan in cupboard" if you made it so you could reference non-free software. But the FSF is smarter than that, and frankly lots of people are up to making this workable. It's far from impossible, it's not even a magnificent feat if they pull it off.
"The next time someone comes along with an idea similar to linux-libre, Guix or distro-libre, I want them to benefit."None of this implies that the FSF has enough sense to make this work -- if we are talking about problems the FSF "WONTFIX" there are bigger ones, to be sure. I've written off the FSF as a worthwhile organisation. But I consider this to be as much about rms as the FSF, and I'd actually keep 99% of the DFSG (it serves a real purpose) rather than throw it out and make a reasonable exception for scripts the MAIN PURPOSE of which is to remove non-free software -- to do the very work of FSDG-compliance itself, but to make it easier to do at a greater scale.
I do not think paranoia and isolationism is the better option here -- although I do not (AT ALL) trust the Guix devs who would benefit from this, it is not for the Guix devs (who I don't like, support or endorse -- I would sooner accuse, warn of and write off) that I make this argument. On the contrary -- the fact that they are finding the same issue as a problem makes this argument all the more relevant at the moment, though I hardly consider Guix a good example of the benefit. But the issue keeps coming up, and I think this is a real weakness of the FSDG.
The next time someone comes along with an idea similar to linux-libre, Guix or distro-libre, I want them to benefit. The fact that I no longer support Guix is truly a side issue, and entirely so.
Once a script has a primary purpose of DSFG compliance, it should be allowed within the DSFG-compliant distro. And once the script ABANDONS that primary purpose, it should no longer qualify as DSFG-compliant. I believe this is (or is already most of the way to) a safe exception. We can do this. It mainly needs the blessing of the Chief Gnusiance (I honestly don't give a damn about what the FSF thinks).
Will he make this a priority? I doubt it. And to be certain, this change has so much less meaning without his attention. Unfortunately, we were very close to rms having a true successor but the FSF blew that, in its corrupt and mutinous state. Obviously we can all make this change in policy ourselves (as a matter of enforcement) though a legitimate, authoritative change to the policy itself (from a legitimate authority who might make a small reform to the FSDG) would be better, if unlikely.
If I thought it were impossible, I wouldn't go to the trouble of talking about it. But I think the odds are small of this improving. The FSF does make exceptions and improve policies, but I think this is pretty rare.
Ultimately I want the FSF to be conservative in some regards -- I actually think we ALL do, or the FSF would actually fall apart.
I mean in this regard (it clearly has in other ways already).
But even if the FSF really has the job of being orthodox and reasonably strict, there is always going to be a line (somewhere) between "reasonably strict" and "self-defeating and basically nuts". Some people will always act like they're the same thing, too. I think Free Software can be reasonably strict (how about "reasonably principled" -- does that sound more favourable? It's something I think we should continue to care about) without staying in corners we have painted ourselves into -- where we are forbidding from creating software that has a clear and defensible goal of helping people fight against non-free software, but must MENTION it to do so.
I have already defended orthodoxy as a way of preserving culture. But I have also defended evolution as a way of improving it, without losing orthodox status.
The rest is up to... well, frankly everybody.
And let's take into account the possibility that nothing changes in policy (the letter) but that practice alone changes... Or in other words that the "FSDG never REALLY forbade this in the first place..."
Okay, then at least will be one (more?) example of the FSF foregoing literalist silliness for a sane exemption to a very strict rule of its own making.
Some rules actually require a fair amount of strictness to work at all. I think the FSDG (I have to type that about three times to avoid the DFSG habit, you know -- every time) is an example of that. I'm not suggesting we make it NOT strict. Just that we take away the craziest part of it, and that we do so in a way that does NOT cause it to collapse.
Again, if I thought it wasn't possible...
Also, could we please change the damned name? It can still be the Debian Free Software (Damnit, it even unpacks as DFSG!)
It can still be the Free Software Design Guidelines historically, though it would make 2020 a slightly nicer year if it was the year we decided to rename them to the "Libre Distro Guidelines" or LDG. SO. MUCH. NICER.
My poor fingers thank you in advance. (No, not that one. That finger is reserved for "thanking" the backstabbers who signed the Guix petition).
Also rms, if you deliberately named the FSDG to be nearly identical to DFSG on purpose to mess with them, that was just bastardly:
Free Software D...n Guidelines
D...n Free Software Guidelines
I doubt you did it intentionally. "GNU/Linux" has a very good reason for it. "GNU's Not Unix" is clever and funny. That's more your style than being nasty. Either way, I really do hate the name!
Long live rms, and (At the very least!) Tolerable Hacking. ⬆
Licence: Creative Commons CC0 1.0 (public domain)