Not at all, really. Forking is fine and building a business off of it is fine (I don’t personally see the value in it but apparently Y Combinator saw fit to invest in this so what do I know). Where they fucked up was replacing the existing free software license with some “AI” generated mumbo jumbo, because they were “too busy building” to “bother with legal.”

You didn’t have to “bother” with creating a license, because there already was one. No one in free software should be rolling their own custom license (GPT generation aside) because there exist perfectly good ones already.

I’ll be “that guy”:

F-Droid is a software repository, not an app store. The distinction is subtle but important. A software repository offers a community-curated collection of software packages whereas an app store is just a marketplace for software developers to offer products to end-users. A software repository serves the interests of its community first, whereas an app store is merely a means for developers to sell products to end-users.

There are those who believe that F-Droid’s role as a “middle man” vetting and building packages from source instead of blindly shipping builds provided by upstream makes it a security risk, because you’re trusting F-Droid in addition to (some say instead of) the upstream developer. Perhaps telling is that none of these critics can offer an alternative solution.

Before anyone mentions Obtainium and Accrescent, these are not alternatives to F-Droid, they solve completely different problems.

This is why the marketing around flatpak bothers me. It’s touted as some kind of “universal Linux package manager” but Linux is just a kernel - all the stuff that apps depend on comes with the distro. So, in order for flatpaks to be “distro independent” they basically have to supply all the stuff that normally comes from the distro - effectively building a second distro on top of your existing one.

Nix and Guix are the same but at least I think they’re more up front that they are effectively distros that can run on top of your existing distro or as a standalone operating system directly on top of Linux.

FUTO changing the definition of open source to suit their business model is like that time US Congress decided that pizza was a vegetable because it has tomato sauce.

FUTO’s EULA may superficially resemble a true free software license (and may be good enough for you, personally) but it fundamentally undermines core tenets of the free software movement in order to preserve their business interests. All pseudo-FOSS licenses (whether of the “ethical” or the “business” variety) do this, because they prioritize the interests of the rightsholder above those of the community and the user. If important free software projects like Linux and Firefox were released under this license the free software world as we know it would not be possible.

As proprietary licenses go, it’s certainly far from the worst.

Free software is a matter of liberty, not price. It is perfectly legal and ethical to sell free software. Keep in mind if you’re using third party code (whether it’s libraries or external contributions to your application) you must abide by the terms of whatever license it is under, this is whether it’s paid or gratis.

It’s even perfectly legal to fork an existing free software project and sell it on the play store, although whether that is ethical or not is up for debate - depending on what efforts you put into your fork before selling it, an orthodox Stallmanist might have no problem with it but the original developer(s) of that code may perceive this as “theft.” Keep in mind you must abide by the terms of whatever license the project is under, so if it is a copyleft license like the GNU GPL you must either provide corresponding source code or an offer for such.

Assuming they own the copyright (which I believe they do, since they were able to relicense it to begin with) they can absolutely offer it under a dual licensing arrangement even if the licenses are incompatible. It would only be an issue if other peoples’ AGPLv3 licensed code was in there, but as it is not the only copyright they would theoretically be violating is their own, which is literally not possible.

Dual licensing under a free software license and proprietary EULA is a common business model, especially when the free software license is a strong copyleft like the AGPL, since the proprietary licensors do not have to abide by certain conditions that free license users have to.

When one asks if something is free software (a.k.a. FOSS) the concern isn’t so much trust but rather can one view, modify, and share the program. Sandboxes solve a different problem.

In the case of a javascript bundle, in order for a user to exercise the Four Freedoms they must at minimum be provided with corresponding source code for each component in the bundle, and preferably some way in the browser for the user to inspect and modify it. In other words, it must be treated like any other compiled binary program. A lock file with specific versions probably isn’t necessary (and server configuration and source code definitely isn’t).

You are right in that this would require cooperation from the service provider to provide this metadata, and most definitely would not do this. Therefore, such an extension as OP suggests would have the effect of blocking the vast majority of javascript on the web today. LibreJS tries to some extent but I don’t know how well it can handle bundled javascript files.

No opinion on OSMand but Magic Earth is proprietary which is a deal breaker for me.

Aseprite is proprietary, but so is this. There is LibreSprite though.

AFAIK FUTO/Rossmann, to their credit, stopped trying to openwash themselves. They created their own phrase “source-first” to describe their proprietary EULA.

That doesn’t change the fact that it’s proprietary and not worth promoting here.

If Linux is just the kernel then Android and Ubuntu are equally Linux.

It is and they are. These are demonstrable facts.

I have no problem with referring to the family of Linux based operating systems collectively as Linux (with GNU/Linux being a subfamily of such), however, I firmly believe that the mythical concept of “real Linux” where some Linuxes aren’t really Linux is what creates the confusion. I would rather use other terms, like POSIX, Unix(like), and FreeDesktop to refer to so-called “real Linux” (with the caveat that they also include BSDs and the like - but I include these as part of the free desktop operating system spectrum, as most so-called Linux apps also run here. I don’t place special importance on the kernel because it is technically the furthest thing away from the user experience).

(Android being Linux isn’t a mere technicality - it means you can get a full terminal environment with a package manager and “Linux apps” and even run a full desktop environment if you really want)

Stallman’s attempt to rename Linux

There was never any “attempt to rename Linux.” Stallman simply wants to clarify which part of the operating system is “Linux” (the kernel) and which part(s) are not (many of which are his work, which Linux fans insist on also calling “Linux” even though the GNU project predates it by almost a decade).

Any “confusion” on this point is the result of Linux fans spreading mistruths (I assume only sometimes intentionally). Unfortunately at this point the myths are so firmly ingrained we have myths about the myths (like “Stallman wants to rename Linux”) and in my mind Stallman is definitely fighting a losing battle nowadays. Still, a falsehood being widely accepted does not make it true.

Linux is an operating system kernel and Windows is an entire operating system. You can’t really compare them

The most obvious difference going from Debian stable to GNU Guix is that Guix is a rolling release distro, not stable (in the Debian sense) at all.

Package management is also very different as it’s fundamentally a source based distro, although sometimes the build servers can provide prebuilt packages if they’re available. Also, Guix has the concept of “profiles” which group sets of installed packages; typically, there is a system profile as well as a profile for each user, but users can also create their own separate profiles. This means that a user can install packages to their own profile without needing root permissions.

Profile updates are done in an atomic manner, such that changing the set of installed packages (installing, updating, or removing a package) actually creates a new generation of the profile, and it’s possible to roll back to a previous generation if something breaks. This is true of the system as well as the user profile(s), of course. A profile generation can also be exported as a manifest, which can then be imported to create a profile generation on another system, allowing package management to be done in a declarative manner.

Finally, Guix has a commitment to ship only free software, and uses linux-libre as its kernel. Debian has a clear separation between free and non-free components but does ship non-free software, including firmware blobs, and I believe as of recently the installer provides them by default. There are unofficial Guix channels (=repositories) that provide these things.

I will continue to maintain that it is bitterly ironic for a product which is 95% based on free software to be so hostile to software freedom. They feel so entitled to take but don’t want to give back, and they justify it by saying that that others will do the same thing they did if they do make it fully free.

A proprietary browser is a non-starter for me, especially when there are many free alternatives, even Chromium based ones. I’d take Ungoogled-Chromium on desktop or Cromite on mobile, heck I’d take Brave even.

Currently I run GNU Guix on my desktop, laptop, and servers. I like the dedication to software freedom and the way package management works. Before that I used Debian until 2019, Trisquel until 2014, and Ubuntu until around 2010. Debian and Trisquel are fine and I don’t have anything against them, I just like the Guix package manager more. I’ve used Xfce with all of these (and before then, GNOME 2). I set it up the way I like it and it never changes.

I typically run LineageOS on my mobile devices, without microG or any proprietary apps. As I’ve said before my preferred OS would be some variant of GNU/Linux, preferably Guix as well, but LineageOS works well enough.

I run OpenWRT on my router, and had a previous router than ran LibreCMC (a variant of OpenWRT using Linux-libre).

Windows games are made for Windows so I prefer to use Windows for them. I don’t particularly want to turn GNU/Linux into Windows, I think it deserves better than that.

For me LineageOS is a good baseline. I don’t have anything against “privacy” OS’s but they’re not really for me. I just use F-Droid to get apps and don’t care about compatibility with proprietary stuff so neither microG nor the GrapheneOS sandboxed Play services are of interest to me. I don’t use GrapheneOS because I don’t have or want a Pixel phone.

LineageOS significantly increases the lifespan of devices it supports and that’s important to me. Planned obsolescence is cancer.

My ideal mobile OS would be something like Mobian (or even better, a GNU Guix based distribution) but it should be noted that AOSP is also a Linux based operating system and thus anything derived from that is a Linux mobile OS.

It’s worth noting the reason why Molly isn’t available on F-Droid proper: it’s entirely F-Droid’s decision. F-Droid policy requires that developers support (or do not oppose) inclusion in F-Droid, which Signal’s developers haven’t. Signal’s developers are also hostile towards forks and do not want them interoperating with Signal. Thus, F-Droid does not allow Signal or any fork of Signal.

https://gitlab.com/fdroid/rfp/-/issues/2297

I am not really a fan of this decision, I don’t think it’s logical that an upstream can forbid a fork from being published anywhere - but that’s F-Droid’s call to make.