muckrights-sans-merde
bonum fabula frat
### gnu-is-truly-dead
*originally posted:* sep 2021
the gnu project did not go out with a bang-- it simply fell apart until it stood for nothing.
actually thats not entirely true, gnu most certainly stands for something historical-- it will continue to inspire for generations.
but as a status report for the entire collection of projects known as "gnu", gnu could not be much more dead than it already is.
personally, i blame wget. wget wasnt the most important gnu project, it isnt the most obvious indicator of gnus status one could rely on, so i will explain that reasoning in a moment.
gnu was already falling apart before i blamed wget.
ive been paying attention to the status of gnu for a while now, trying to keep watch over how free it is (compared to what it once was) and also watching its drift towards and takeover by github.
the takeover is not complete. and that honestly hardly matters. what matters is that the the drift away from autonomy and freedom and towards corporate control will outrun the pace of reform.
in that most important regard, gnu is dead. it will not recover, and when they finally turn the lights off it will not be much different than this.
perhaps it would be more fair (or accurate) to say that gnu is already "undead". thats more hopeful than the message im trying to convey here, though its true in a way. gnu will without question live on in a variety of ways. it is the child of richard stallman, and i dont believe it will be his last. in this sense as well, gnu will certainly live on.
but as dead as it is now, gnu is propped up like the titular (albeit "former") character in weekend at bernies. people who have helped to kill gnu are treating it as though it isnt dead. its sort of the perfect crime really, since you cant be blamed for killing something if people dont realise its dead.
whether you believe me or not, im not too hung up on blaming the maintainers. some of those who failed actually tried, though i wont say im too impressed with the effort. im sure there is a great amount of technical work, but that is no indication of a political defence or a stand for any principle. takeovers are a lot of work too, they are very heavy work against everything that gnu once stood for.
the fact that gnu stands for so little at this point (other than in a historical sense) is what really does it in. and the people who deserve the most blame are those who worked to take it over-- probably less the maintainers, some of whom certainly share in the blame. oh-- i dont count the gnu.fools among maintainers since they technically left the project, i consider them traitors and saboteurs who defraud the public as well as the project. but the responsibility for this tragedy also goes beyond the work of the gnu.fools and their little (but not insignificant) coup. the big picture is still larger than any of its individual parts.
it is time for a revolution for certain, in the world of free software. gnu is dead in the sense that right now is just as good a time for that revolution as any other time, but in reality it will almost certainly take longer for a project like gnu to gain any sort of foothold, because right now there simply is no project like gnu-- including gnu. of all the projects i am aware of, the closest thing is gnew.
if you are looking for a free platform to modify into a free software collection like gnu, i still recommend you start with bsd. however, it is too early to say whether the next truly gnu-like platform will be based on a mixture of gnu and bsd, as i originally predicted, or whether it will be built from even more disparate (and newer) components. i see advantages to both approaches, in terms of practicality and ideology. i believe gnew will be more "original" than just a mix of openbsd and gnu sources, but whether the revolution goes that way or not i believe it will happen.
the key to all this is that people who actually expect gnu to live up to its promises will not settle for what they actually get-- and the way that gnu is run now, not only by stallman (i continue to have no objection to stallman running his very own project which he himself founded, nor have i called for stallman to be overthrown from gnu leadership) but by all other gnu maintainers, is INSUFFICIENT to the goals of gnu.
the false (or half-true) promise that organisation-centric reformers like roy will offer us is that eventually, enough people will call for improvements to gnu (and the fsf) and these things will BEGIN to be fixed.
and thats useless. in theory its just what we need, in practice it is clearly useless. no reformer is going to give us anything other than mediocre improvements against the continuing backdrop of tragedy after tragedy. this has been the course for years-- long enough that it is starting to carve out a hefty percentage of the entire time gnu itself has existed. when we talk about the amount of time that gnu has dwindled in what it stands for, we approach fifths, fourths and thirds. anybody who believes gnu will be reformed is fooling themselves. they are certainly welcome to try.
i used to debate (pointlessly) with alex oliva on the significance of systemd, to which he insisted that i focused too much on the (most prominent and well known) example. i dont know why he insisted the most prominent and well-known example of a problem was not worth my time, but i certainly foresaw the problem getting bigger than systemd (even as it continues to grow and take over gnu/linux altogether).
gnew speaks (very importantly, i think) of "nebulous programming interfaces". this is as good a term as any, and it applies not only to systemd but also to rust, gnome and ultimately the linux kernel itself.
if we broaden this even further, to encompass the problem in the broadest political terms (that remain technological at least) we could call it "the deliberate and strategic unmaintainability of software".
in other words, destabilising free software projects.
when the halloween documents made it clear (with very little pussyfooting around) that this was already a goal roughly two decades ago, people shouldnt balk too much at the idea that the goal has been pursued with some success. free software has been destabilised, corporations have done no better to free software than stephen elop did to nokia, and the effect / consequences are similar: hostile takeover by ibm, microsoft and others.
one shouldnt project their desire for stable, working software onto large corporations (like red hat, even before the takeover) who benefit from fixing and maintaining bloated shit that addicts people with features. features are not the enemy, they are the substance that the enemy uses to lead people away from autonomy into corporate dependence. this was always known as "extend" and it isnt a new problem. when people are constantly (over)extending software we rely on, they are putting it at risk. how often does it get unextended? no, really. just as with the best route to privacy being not giving up information in the first place, the best route to stability is to hold onto it for as long as possible. software that goes from stable to destabilised usually stays that way.
overextending software is too reliable a weapon, and thats why non-free software is built around it as a practice. they arent so stupid they never noticed, they are at least clever enough to notice and exploit it. if we want to win, we need to be clever enough to notice them exploiting this and NOT DO THE SAME to ourselves. this needs to become part of freedom 101. (ive campaigned to make the freedom to NOT use the software the fifth software freedom).
some applications at least, we can probably afford to see take foolish directions and destroy themselves. with gnu at least, we had stable components that (if properly maintained) would not become stupid in exactly the same way, if only due to a more or less complete lack of interest in extending them to death. i am aware of counterpoints to this argument, dating back at least as far as the unix haters handbook, but i am unswayed by the idea that this was always the problem it is today. (i am more sympathetic to the idea that it was always a factor of some sort, but on a scale that is as categorically different as cross-breeding plants vs. direct resequencing of their genetic molecules, which comes with additional hazards and complications such as introducing allergens into crops at a rate never possible before).
if we have stable development tools and a stable operating system, we can watch some applications (such as audacity) come and go without worrying how we will perform basic tasks in the future-- in other words, some applications will break but the kernel will be relatively stable, as will the most important features of userspace.
we no longer have this, not even to the degree we once did. relative stability was a byproduct of gnu, if it was ever a distinct goal then its not anymore.
the relationship between (relative) stability and autonomy and freedom is not difficult for any political analyst, historian or expert in business or military strategy to explain-- destabilising free software is a prelude to takeover, just as it is for non-free software. the primary difference is that free software sometimes makes the claim that this is impossible, because the source code is both available and free. but this ignores the fact that the source code (and collaboration in maintaining it) is exactly whats being systematically destabilised and taken over. it is a retort to the problem that simply denies the problem exists. the license and source code by themselves do not guarantee anything-- collaboration itself is a weak point, and weak points without adequate defences are often exactly where freedom falls apart.
so people who insist that its impossible for free software to be taken over in the first place-- this is a pointless debate with people who dont know what theyre talking about and who ignore what has already happened. unfortunately that probably includes most gnu maintainers. it may even include stallman (though i still give him more credit than this-- hes certainly no idiot).
but there is a difference between being vulnerable and being taken over. as i said, ive spent a long time watching the drift of gnu towards github and other things, and i consider gnu as having reached the point of being unmaintainable (whether deliberately and strategically or otherwise).
if you wonder who the next chief gnuisance is going to be, i can tell you that it wont improve under them, as the gnu project at this point would need to change so much to even be fixable by a chief gnuisance that it would not be gnu anymore. and even that would be better than simply letting it die. people need to start thinking about what parts of gnu are the most important-- im averse to things that require gtk or gnome (unless gnu forks gtk, which of course will not happen) and i am similarly averse to github dependencies in the gnu project.
the reason to single out gnu is a simple one-- gnu (unless you insist on emacs, which i consider a quibble but more power to you) is the original flagship of free software. it sets the standard for all other projects, so if it compromises the rest of free software suffers.
alternatively, gnu is NOT as important and vital as that and some other project sets the standard-- but i have not found any such project anywhere. the closest things i know of are openbsd, gnew and hyperbola, none of which are complete replacements for gnu yet (i dont think openbsd is an adequate replacement for gnu itself, only that its the best place to start if you need a working and reasonably stable platform to begin with). these are still the things i look to for hope, though i somewhat doubt there will ever be a SOLE free software flagship like gnu again. (more likely there will be a few key projects brought to you by different groups of people).
as for what gnu still stands for, gnu will never give the user control of their computing. in theory it will, because source code and free licenses. in practice, it will hand more and more control over to corporations that have a mission to reduce the control you have over your computing. and the people who fall for that ploy are frankly, a little bit stupid to do so. i mean the people who claim to "stand for your freedom" are either suckers, or utterly full of shit themselves. to say the least, they are ineffective and unrealistic.
i switched to bsd for a variety of reasons-- to protest the state of gnu/linux, to get away from the linux kernel as quickly as possible, and to reduce github dependencies. i think boycotting github is of the utmost importance, and any standards free software once had about avoiding non-free software while they rely on github are a joke.
its still very obviously key to avoid non-free software-- for example, i applaud and encourage efforts to remove non-free firmware from openbsd, as openbsd bows the least to non-free software of all the bsd flavours (which is another reason its ideal for making a fully-free bsd flavour) and although i recently got a motherboard that required a non-free driver for the network interface, i have also destroyed that board. (i could have simply run openbsd instead of netbsd on it and only the network interface would be unsupported, but as it happens i destroyed it anyway).
moving away from non-free software is probably half the point. with non-free software, you dont have control of your computing.
but with corporate overthrow of free software, you dont have control of free software. in theory, you can take the source and do whatever you like. in practice, this broadly requires more people-- and thats where things get tricky. not all of these people have any real standards at all-- gnu is not holding them to one either.
now that i use bsd, its easier to keep track of what i actually need from gnu (and what needs to be salvaged the most urgently)-- openbsd (and not only openbsd) has a policy against copylefted software. if you use bsd and still need a copylefted program from gnu, this is definitely a sign that its important (despite this policy, openbsd has never managed to purge gnu programs entirely).
wget is one of the most indispensible programs in the gnu project. until i discovered that ksh is actually far better than bash in terms of ease of use and stability, i considered bash important as well.
obviously i could use curl, but curl is a worse example of a project in almost every way-- i like that it supports gopher but i still think it does too much. its developed on github (where the main repo is located) and shows no sign of ever extricating itself. i avoid it, though i used it (even wrote programs around it) in the past.
if you want to get pages or even binaries from the web from the command line, youre probably going to use either wget or curl. thus in an ideal world, or ideal other than curl being a sellout, wget ought to be some kind of bastion of free software.
here is what it is instead, and i consider this basically the last fucking straw (there are too many) as far as pretending gnu is actually still free software.
first, the developers i consider key to wget are darshit shaw (the lead) and tim ruehsen. i dont blame these two for the destruction of the gnu project, i blame them for the destruction of wget. the original (and inactive) author is hrvoje niksic, who will hopefully forgive my use of the ascii-friendly spelling (which i got from the gnu savannah page).
one appears to be the wget project lead.
both use github. both use github in connection with wget. both are involved with wget2 (a separate project from github).
wget has a dependency (libpsl) which ruehsen maintains on github. i consider this at least, an act of treachery.
this article started as curiosity about wget2, which represents the future development of (but does not, at least yet, replace) wget.
it is now possible to compare wget and wget2, and see the trend that wget in general is moving towards. this spells bad things for gnu in a few ways.
its not so much that wget2 is worse than wget (though it probably is worse) as it confirms that wget is going to stay on a trend of being a corporate hostage. it moves from mailing list collaboration to gitlab for official purposes, and it is mirrored on the gnu website. the gitlab page directly promotes non-free software by providing a link specifically for using it with visual studio code-- again, this is a gnu project; this is on the same page where you download the latest official sources of gnu wget2, and it promotes non-free software on the very same page.
i think its worse than you have to use gitlab to be a first-class contributor to the project. you are also welcome to use github for this (per the invitation of ruehsen, an alleged gnu maintainer).
one of the things that i find really scandalous about this is ive been told, by gnu maintainers who are even stallman supporters (defenders-- having a position i strongly agree with and really consider vital) that gnu is NOT drifting towards github (it clearly is) only SOME people have CHOSEN to use it... fucking what, as though somehow a repository will magically and of its own volition, MOVE ITSELF there from savannah?? theres a painful level of bullshit rhetoric that free software has fallen into, as free software is overthrown by the much more corporate (and more dishonest, and less free) open source.
where are gnus defenders? theyre abandoning gnu as gnus maintainers drag free software itself towards the toilet. the reason i dont blame them entirely is that it is indeed part of a larger picture, albeit one that seems to have made gnu maintainers relatively naive and dishonest with themselves (most of them at least, are more honest than the gnu.fools traitors, so please disregard anyone who may claim that im lumping them into the same category).
the people who want gnu to be free are realising it isnt going to happen-- theyre moving on (reluctantly i might add, and with great disappointment) towards things that are more maintainable, less bloated, less controlled by corporations, and easier to defend.
gnu wget2 does not reduce github dependencies. it is (for now) a smaller download but contains 12 times as many files (i dont know what the fuzzin stuff is, but in terms of FILE COUNT it is now 97% of the project-- im aware that lines of code is almost always a better metric than file count, but this is still peculiar in my opinion).
and thats 16,000 files, by the way. it is obscene that you need this many files (even the 1,300 that wget includes in its source) to create a binary that downloads one file from the internet.
of course i say this as a python user which is certainly hypocritical on the surface of things, but its an interpreter and the thousands of files gnu wget2 needs do not include the compiler-- which is the only way to make the comparison fair. not to mention that i think the file count in python is obscene as well, and i would love to strip it down more. python also does a whole lot more than wget, being a programming language.
gnu keeps getting less modular, and dragging more and more bullshit into everything it needs to be gnu.
it used to be able to afford the luxury of doing so. at this point it is a threat to the sustainability of the project, following a similar trend that is destroying free software in general and replacing it with github, open source and even non-free software.
gnu is becoming slave software. but it is already nearly (mostly) useless to people who want a free and maintainable platform for their computing. if you want to create such a platform, you are far better off starting with openbsd-- but openbsd is NOT the point of this article. the point of this article is just how far gnu has fallen.
it will always be the original project of its kind and for its purpose. we absolutely will need others, and gnu will (in many ways) provide an essential template for such projects.
whats changing the most is how many really terrible ideas are making it into the gnu project, for anyone hoping to control their freedom. if gnu is to be a historical example of how to put the user in control of their computing, then people are better off looking at older versions of the gnu project than they would do to look at what its become.
thats never going to change at this point, and gnu will never significantly improve or contribute more to freedom than it already has.
theres no question that its the most important free software project of all time. theres nothing on the horizon either, that is going to quickly replace what gnu provided.
there is simply a great historical contribution, a contemporary farce, and a future that is uncertain-- except for the certainty that gnu is dead.
if they ever do bring it back, it will no longer be gnu. it may be as good, someday it may be as vital, but it will absolutely be something else.
i am in favour of preserving and salvaging the gnu project. and yet it reached its peak years ago, so what precisely do you think we should preserve?
free software may get better from here, after a while. gnu, most unfortunately, will not. the reformers will bank on "regaining interest", but that misses the point. nobody interested in saving gnu who tries to do so, will manage to avoid the step where they throw their hands up and say "this will require something else"-- a revolutionary change, not just building on an unstable foundation.
gnu WAS the foundation of all free software projects. it no longer is, it no longer can be, and most of all it no longer will be. its dead. perhaps it will be cloned, but it will never be revived.
this, is a late gnu.
but we can be certain that for years, we will be told "naw, its just resting!"
alright, if its resting, then lets wake it up: "HELLO, GNUUUUUUUUUUUU!"
look, my lad, ive had just about enough of this. and thats what youll be hearing from those who know where gnu is headed-- it will be stuffed and placed in a museum, and people will say it looks incredibly lifelike.
the truth is that gnu was a reimplementation of an existing platform, for the sake of liberating it. and in accordance with the nature of freedom, that truth is as recursive as the name itself.
saving gnu will never happen through incremental changes. it will be a complete overhaul at the very LEAST-- and at least as revolutionary as the creation of gnu itself.
not even because its easier to do that way, but because it is necessary.
but first, we will have to put sunglasses and a hawaiian shirt on it, and parade gnu around like its still with us. and it will be just as ridiculous as that stupid movie. speaking truly, it already is.
=> https://muckrights-sans-merde.neocities.org