Gnome and many of related projects in the hand with Red Hat has been breaking the user experience for years and the problem is everyone else except them.

This article from 2013 explains why in 2021 Gnome only get forks and "haters". they use whole linux world as a testing polygon and if you try run rhel it is and stays quite stable

@jrballesteros05 > users of arch-based distros complaining about upstream breaking something for them
Never change, rolling release camp.

@tdemin @jrballesteros05
>upstreams complaining about users using Debian stable to avoid breaking updates

@tdemin @jrballesteros05
Have you missed all the "hurr durr Debian ships old version of my software for which I get bug reports why can't just everyone curl|sudo bash or docker or flatpak, distros are a hinderance" dramas?

@tdemin @jrballesteros05
I don't remember the exact developers / dramas / links, but I do know I've seen them semi-regularly over the last 5 years.

Also, there were enough of them for Drew to mention it on his blog:

> Some developers may complain that distros ship their software too slowly, but you shouldn’t sweat it. End-user distros ship updates reasonably quickly, and server distros ship updates at a schedule which meets the user’s needs.

@tdemin @jrballesteros05
I really wish I had a bunch of links to upstreams complaining to post here.

@wolf480pl @jrballesteros05 > not having a link to every single drama on your browser bookmarks
(btw i have a folder for that but forget to add stuff to it)

@jrballesteros05 I think after this thread

it is a very reasonable idea to avoid anything connected to gnome whatsoever. I am not talking exactly about decisions made by gnome team, but the energy and the spirit they are operating on

I wish some of the forks would be as popular but more cooperative with the community, if system76 are working on this - very good news for linux community!

@fedor I wish the best for those forks too, but I can say that sadly GNOME kind of owns GTK too, so GNOME's bullshit will affect the library and many forks.

I wish the best for all the alternatives, this is libre/free software.

@fedor @jrballesteros05
it's impressive the levels of "we don't want to support this thing that is standard behaviour, and it's your program that is bad for relying on this standard behaviour and not doing extra work just cause we don't feel like it"

@fedor @jrballesteros05 jesus christ how one can be so dense. And I thought gnome would want every program to look gnomish.
and this:
>Enough spam for today. Locking the issue.
The issue appears not be resolved in any way but the discussion was locked anyway.
@jrballesteros05 I agree that RedRat is pure evil and what they do now is copro-rat-izing the open source. they try to make it all dependent on their solutions. nothing is worse and more dangerous for the community than enterprise monstrosity in open source.
but the problem is not just interfaces (although I must note that Gnome intentionally did this to make everything depend on its awful DE). the problem is the endless "updating itch" as well. there're two kinds of software: the one that works and the one that doesn't. if a program works fine, performs all needed tasks, it's optimized and well tested, it does not need constant "development" on itself. stable good things just work. but nowadays we see the poisonous "scrum" rot that ruins the normal development, in which the first stage is thinking, research and architecture planning and only then writing code, tests and debug come. this perverted way of writing software implies constant fucking up around the code and the code is actually never ready and working. it's always in the process of adding some senseless whistles and bells that make the code more clumsy and error prone, but don't add anything useful to user experience and performance, as a fact. this perpetual useless complication of code leads to bloatware as it is. and we see it everywhere. when the motive for development is not the finished working software piece that delivers solid and stable set of features for user, but constant meaningless rewriting of senseless features for the sake of rewriting itself - this is a serious problem.

@iron_bug I honestly don't like some changes that Red Hay are trying to make and they are getting successful.

I am getting quite paranoic because the only OS which does not distribute spyware crap by default is GNU/Linux and I think the first "legal spyware" will come in the hand of systemd and/or flatpak.

I think it always be better to have different paradigms and not a only one. The one which RH wants to "impose".

@jrballesteros05 @iron_bug thanks for this, I've been thinking I might like something other than systemd, at least for home

@jrballesteros05 yes, exactly: different paradigms. diversification of software, not the diversion with "diversity" of developers crews that they popularize, distracting the attention with senseless controversy and weakening the opensource comm unity with internal strife. meantime, the software itself rapidly loses its diversity and independence. open source turns into coprorate enterprise. it becomes handcuffed to the single solution they impose. and this is really bad and dangerous for the free software. one more freedom (and actually the most important one) is the freedom of developers to choose the way they write their code. this is the core engine of progress in such voluntary development. people write "what they should" at work. and pet projects and open source always were the realm where developers applied their new ideas and an independent perspective in coding and software architecture approaches. but now it imminently turns into coprorate "one-way" dull conveyor that does not implies any creativity. it simpy strangles any initiatives and variations and this kills the developers eager to contribute to free software, it turns open source into just another dull stiff corporate edition that dictates the rules to everyone. and this is utterly wrong.


> it turns open source into just another dull stiff corporate edition that dictates the rules to everyone. and this is utterly wrong.

this accurately describes a majority of popular open source software now. everything has to be "professional", where professional isn't simplicity and functionality, but how many enterprisy things you can integrate. "build pipelines", "testing frameworks", "vulnerability checkers", "containerization", ad nauseam.

if the software wasn't so complicated that no one understands what is happening anymore, one wouldn't need these things. redhat actively tries to make linux as complicated as possible because it's a sure-fire way to increase sales.

the worst thing is that a mainstream lennux distribution with systemd and other abominations is usually less functional than SuSE 9.0 for which i recently found the install CDs in an old box. everything which _really_ is better is following the unix philosophy of doing one thing only, like new audio and video codecs. also, the software which existed back then still works great! FVWM? still working! mplayer? still working! vim? still working! bash? still working? GNU coreutils? still working!

pulseaudio barely works. everytime i interact with systemd, somethings broken. containerized apps are _at best_ barely working and at worst containing malware. that there is server software which is only able to run in a hand-crafted docker container is a disgrace to the whole profession.


@bonifartius @jrballesteros05 I don't have anything against professional software. actually, professionally written code is well designed and optimized and this is especially important for shared resources like network. on his local computer user may run whatever he likes. but when it comes to network software should not generate flood. this is actually driving me mad when I see the logs of Fediverse data exchange logs - it's just terrible. it's non-professional. and it kills the netowrk and speed.
but professionalism does not dictates "one-way" solution. one-way is typical Windoze. and we know well that that means. but professionalism does not kill the diversity in software. it just makes things optimal and fast.

the complexity of the system is the result of development of hardware. really. operating system solves many tasks and they're not simple nowadays. but I cannot say there're something unclear in Linux kernel. I dealt with it quite a lot and wrote drivers and code for it - it's really well-designed. it's huge because of too many drivers it should support. but it keeps the whole thing maximally clear and friendly for developers. I mean system developers. not people from the street. to write any code like this it takes at least special education and a lot of experience. this is not an amateur level software.
considering userspace, and especially coprorative bloatware that seized the realm it all looks much worse. there're no clarity, no optimization, no general ideas behind it. it's just a try to stuff all functionality possible into one single solution. and Uniox way is decomposing of complicated tasks, first of all. and I also prefer to have a certain simple utilities for each task. we cannot design an OS in this way, alas. but we can do it for the user space. users don't actually need all-in-one software that weights gigaytes and devours all the CPU and memory at once. and it's really difficult to make such a code clear and free from errors and vulnerabilities. besides, a vulnerability in coprorative crap like useless-d means vulnerability on most coprorative servers that use it. and this is like a weapon of mass destruction nowadays. so we should keep it simple (K.I.S.S. principle) and decouple all non directly interconnected things possible.

I personally keep strictly out of everything that RedRat and Lenny try to impose on us. I don't have useless-d, pshhh-audio and the rest similar bloatware on my system. I try to get rid of Gnome dependencies. it's not that easy, but I think it's not impossible. we should get rid of all coprorative anal probes to get free from their "solutions" and their way leading to hell.
@bonifartius @jrballesteros05 and as a professional developer I cannot call what Lenny does a "professional software". it's terrible badly designed bloatware and spaghetti clumsy error-prone code. no way I ever use such creations. the problem of coprorative code is it never has competition or rivals. so it becomes bad, fat, full of bugs but nobody can send it to hell because everything is dependent on it. this is why depencency on a single solution and absence of alternatives is very dangerous thing. it leads to shitware, like Windoze and thing.

i wasn't criticizing professional software but the "enterprisy" stuff where buzzwords are more valuable than the code!

SQLite is an example for good professional software. they probably have the biggest set of tests for any software in existence.

while not having as extensive tests, the linux kernel too is pretty amazing. i think i really never encountered a situation where installing a newer kernel broke things for me.

the development process of TCL also seems to be pretty sane, judging from their mailing list.

considering lennys software: pulseaudio kind-of works now after other people took maintenance of it :P

at least there is a growing number of distributions without all this stuff! :)


@patterfloof I honestly don't think that a change of paradigm was a mistake. I think it's good for the ecosystem to try new things and it can brings some innovations. What I honestly annoy me is the "I break things and I don't give a shit" attitude.

Sign in to participate in the conversation

Fosstodon is an English speaking Mastodon instance that is open to anyone who is interested in technology; particularly free & open source software.