Fast forward fifteen years and I am finding that Richard Stallman was right. Despite the commercial success of open source, I have found that usefulness does not always lead to a healthy society. (And I am aware that, zealotry and fanaticism have their own issues).
It’s not that I am a purist. Nix is still more practical. I have also heard that Nix and Guix can run on the same machine. However, I think Guix and Free Software conserve principles that I think are important for our society and civilization.
Ideals are the compass, shows you the final destination, where you are going.
Pragmatism is the next step towards ideals. It prevents you from being paralized when the ideals are very far away.
Without ideals it would be as if you don't have a GPS on your car and roam around randomly. Without pragmatism it would be as if your car didn't have an engine.
Yes, RMS is living a completely impractical and undoable life for the rest of the people, no matter what he says.
Yes, RMS is right and he must exist to show us in which way we should work towards
I think in almost all cases the leftists are right, but too often they have have no way to realise their ideals.
The liberal platform is the only left-wing platform that can currently feasibly win an election in America. Maybe that will one day change, but I’ve seen too many on the left throw the baby out with the bathwater because they want their ideal society right now, and nothing else will suffice.
At the same time, I’m grateful they continue to show us which way we should work towards.
But if GP is right about their attitude, it saddens me.
Can't there be a middle ground? Like, promoting free software and still helping people getting things done in the world we live right now?
Rules about gently directing folks to nonguix ( https://www.google.com/search?q=nonguix ) resources are there, and people do try to be non-confrontational about it. It's a pretty businesslike vibe though.
i think ideally user groups like System Crafters might be a better introduction for those not steeped in GNU culture: https://systemcrafters.cc/
On the #guix IRC channel you'll often get a private message to direct you to other channels where it's fine to discuss proprietary software and how to make it work on Guix.
Does it feel good to you when you ask a question and the reply you get is "this isn't for you"?
Requiring a private reply on an IRC server to unblock a comically common use case is exactly the problem I am trying to highlight.
You don't have to belief anything to use Guix or participate in the community. It's an odd sense of entitlement to demand that others accept, accommodate, and welcome your ideology in dealings with you.
I also find it strange that you ascribe "this isn't for you" (a comment made here by strangers) to the project. No wonder then that you interpret entirely harmless community guidelines as if someone told you "to go fuck [your]self".
Yes, and I believe that this attitude harms community building, education and adoption efforts. You specifically use the words "not welcome". Other, more welcoming projects typically don't use such language outside of a Code of Conduct policy document. I'm pointing out that I've been made to feel unwelcome and you're saying that is entirely harmless.
Do you see the problem?
You're (inadvertently?) twisting the phrase "not welcome", which I apply to behavior, to make it apply to people. If we can't agree on how to apply words we cannot use them to approach understanding.
The problem here is the problem of incentives: you can state your desired end goal (prosperous communism, or free software utopia), but there is no mechanism to get there from here, and even if you somehow do get there, no mechanism to preserve that state without all encompassing coercion.
Moreover, the lack of incentives results in lack of prosperity: if there is little incentive to produce a lot (like in capitalism, where your profit grows with volume), people produce just enough to get by, plus some for individual entertainment. I’m sure Guix is good enough for the author’s needs! The problem is to get people to do things that nobody really wants to do. Who will build free software Salesforce? In capitalism, we do that by paying people money to do things they wouldn’t want to do for free. In fact, same is true in free and open source software: I can really only name a handful of projects that are widely used and of high quality that have not been largely funded by commercial world: OpenSSH, some other BSD stuff, what else?
Since commercial businesses want to use these things alongside their proprietary stuff, they necessarily cannot constitute part of pure (in free software sense) distributions. Thus, if you aim for purity, you either end up with shoddy quality, or depend on the largesse of the non-pure activities. This is analogous to modern welfare state, which can only function by skimming the value created by advanced commercial economies, and only to an extent that doesn’t actually choke them altogether.
In short, incentives matter.
Is this perhaps though a result of most projects allowing donations, and some companies (e.g. Google) having a lot of capital to throw around and funding them more-or-less just because? Would Linux not still exist in a similar form to today?
Also, I would add things like the GNU project, and notably VLC media player, alongside OpenSSH and the BSD stuff.
I agree with your final point, though. I can't actually think of a company whose product only in-house FOSS software. Red Hat, Suse, Purism, etc are building off community code, not something primarily developed internally. And companies like MongoDB make all of their money off of non-free cloud products.
One that I can think of that was once like this, the Subsonic media player, actually went closed source due to too many people re-hosting the code and not paying the maintainer.
Majority of work on Linux is done by people who are paid by their employers to do exactly that. Just look at this:
https://www.google.com/amp/s/news.itsfoss.com/huawei-kernel-...
Google involved in Linux is mainly not through their donations to Linux Foundation, but through code written by Google employees on the clock.
There is idealism guided by reality. This is the basis of Naturalism generally.
And there is idealism centered on what people can created that reality as less important then the social implications of that idealism. This is the basis of the Dialectics including Marxism.
I like FreeBSD and the general pragmatic Linux generally for the idealism founded on reality.
I dislike GNU often for the idealism void of thinking about reality.
I'm using Guix for work, at home, my wife has Guix on her gaming laptop with nvidia card, playing games on steam. Nobody has any problems with enabling rpmfusion repo on Fedora, using aur on Arch, adding backports repo for kde or whatever on ubuntu. But as soon as it's nonguix channel it's suddenly a problem. Huh. It's literally just a channel and ultimately it is still same Guix.
"What's that? NONGUIX? Hell no, I ain't having Guix working on my hardware! I don't want a better experience using Guix! Not in a million years!" And then:
> Guix has truly noble goals (in classic GNU fashion) but they're ultimately misguided.
> Actually useless.
Which are really big claims with nothing to back them up.
> The standard encouraged response to "Hi I'm a new Guix user and I was wondering how I can run steam?" is roughly equivalent to "Go fuck yourself" and I'm just amazed by that.
Look, if you can't figure out how to `guix install flatpak' and then flatpak install com.valvesoftware.Steam, then... well, I've nothing to say actually.
That's definitely better than "go fuck yourself". "If you don't have something nice to say, don't say anything at all" and all that.
Tbh, I've seen the "go fuck yourself" attitude in almost every Free Software community I've spent time around. It's really just a matter of whether it's encouraged, tolerated, or actively rejected by the project maintainers and influencers.
I would hope that someone writing a blog post this positive would be one of the "actively rejecting" folks, but I really don't know anything about Guix (hadn't heard of it until this post).
I will say that I run NixOS on my primary home server, and while there's definitely a learning curve there, the community I've experienced is largely very helpful, friendly, and positive. Now that I've got most of it working the way I want, I don't see myself moving to something else any time soon (it's absolutely rock solid).
Both variants are very similar. The Guix tool comes with a "guix system" command, which lets you build Guix System containers, container images, system images, qemu VM images, qemu VM scripts, etc. The only command that won't work on a "foreign" distro is "guix system reconfigure" --- but that's only a "guix system init /" away ;)
You wouldn't expect a bunch of people giving away a vegetarian cookbook to respond super well to, "This is all fine and great and all, but how can I adapt this recipe to grilled dog?"
Especially not after saying their project was pointless since you couldn't easily use it to shish kebab your cat.
> couldn't easily use it to shish kebab your cat
This seems like an unnecessarily hostile response, kind of inline with what the OP mentioned.
Which is totally sensible!
nonguix is proof that adding proprietary software doesn't break diddly squat.
It's all just gatekeeping. They could still achieve all their goals without being zealous, whether you believe me or not.
Guix makes it trivial to add other channels, such as nonguix, providing first-class packages. You can easily add proprietary software to Guix. It's just not part of Guix, because that's not the point of the project. Your "non-mythical" workflows that depend on proprietary workflows are not in any way impeded.
You are projecting zealotry and gatekeeping (both terms with strong negative connotation) onto the project. And I must say that your attitude does nothing to advance your point, however deeply buried in hostile rhetoric it may be.
As to mythical workflows without proprietary software: I'm not using proprietary software. Not on my laptops, my home network, my servers at work. Not because I'm some sort of "zealot", but because I simply have no need for it.
It's really quite odd to observe how worked up you get in response to seeing a community that is content with not using proprietary software.
Figuring out how to use proprietary software with Guix is a miserable experience. As a new, curious user, I was told: I won't receive help, I can't discuss it, I shouldn't use proprietary software, that Guix isn't for me.
This is not a welcoming response. It's not useful. It's hostile. It's zealotry.
Instead, a welcoming response would be to say yeah, that's unfortunate but understandable, maybe help educate why free software is important, and suggest a path forward, such as pointing the user to unofficial channels. By comparison, Nix excels here -- There is no hostility.
It's like being told by a mathematician that I shouldn't use math to calculate my taxes because taxes are unethical and illegal. Yeah that's nice, but I still have to pay them. I just want to calculate my taxes.
I think the very article we're commenting on here shows that it is very definitely not actually useless, it may merely have a different true use than is readily apparent.
Guix the package manager, while still not usable at work for me (lack of MacOS support - just 4 of my eng org's ~25 folks use Linux), is full of interesting ideas and features that work well in Linux settings, and can do useful things. It could make interesting dev shells, for example, much like nix-shell (or Flakes).
GuixSD without `nonguix`? Yeah, I unfortunately share your "it's dead in the water" opinion here. I'm not aware of much, if any, modern hardware that can fully boot to a usable-framerate GUI with networking (and that is also safe to use for such) without at least linux-firmware, which I believe is only available in nonguix. If there's exceptions to this, I'd genuinely love to hear them in a thread here.
"dead in the water" looks different.
In case you're not aware, you can easily create a Guix pack[1] in the Docker image format that can be run on MacOS.
[1] https://guix.gnu.org/manual/en/html_node/Invoking-guix-pack....
Why is ‘become mainstream’ some aspirational goal? What would that even mean?
Making a viable operating system is valuable without having it be the the one everyone uses.
The Guix zealotry will be their undoing, and frankly this is where Nix shines. I can use Nix at work, at home, with family, with vendors, with binary blobs, on macOS. The community tries very hard to help.
With Guix, sure, there's stuff like nonguix, but take a gander at this from the nonguix readme:
> Please do NOT promote this repository on any official Guix communication channels, such as their mailing lists or IRC channel, even in response to support requests! This is to show respect for the Guix project’s strict policy against recommending nonfree software, and to avoid any unnecessary hostility.
Guix has truly noble goals (in classic GNU fashion) but they're ultimately misguided. Yes sure it's great that it's pure and reproducible and whatever else but if I can't get stuff done with it it's useless. Actually useless.
The standard encouraged response to "Hi I'm a new Guix user and I was wondering how I can run steam?" is roughly equivalent to "Go fuck yourself" and I'm just amazed by that.
If Guix ever becomes mainstream, it will be in spite of their efforts.
In the meantime, I'm betting on Nix.