Systemd's Growth Over 2022 (phoronix.com) 236
Phoronix checks systemd's Git activity in 2022 (and compares it to previous years):
If measuring a open-source project's progress by the commity activity per year, while not the most practical indicator, systemd had a very good year. In 2022 there were 6,271 commits which is under 2021's all-time-high of 6,787 commits. But this year's activity count effectively ties 2018 for second place with the most commits in a given calendar year.
This year saw 201k lines of new code added to systemd and 110k lines removed, or just under one hundred thousand lines added in total to systemd in 2022....
Systemd continues to grow and is closing out 2022 at around 1,715,111 lines within its Git repository.
Also interesting: "[W]hen it comes to the most commits overall to systemd over its history, Lennart Poettering easily wins the race and there is no competition. As a reminder, this year Lennart joined Microsoft as one of the surprises for 2022."
This year saw 201k lines of new code added to systemd and 110k lines removed, or just under one hundred thousand lines added in total to systemd in 2022....
Systemd continues to grow and is closing out 2022 at around 1,715,111 lines within its Git repository.
Also interesting: "[W]hen it comes to the most commits overall to systemd over its history, Lennart Poettering easily wins the race and there is no competition. As a reminder, this year Lennart joined Microsoft as one of the surprises for 2022."
Unusual growth? (Score:5, Funny)
If a doctor spotted something on your leg that was growing like this, everyone would know exactly what to do about it.
Re: (Score:3)
yes, true
Re: (Score:3)
Yeah, when I saw the headline, my immediate thought was -
'I believe the correct term is "metastasis"."
Re: (Score:2)
Give you a recipe for some SysV ointment to get rid of it?
Re: (Score:2)
Growing so fast I needed a new belt.
Click systemd story (Score:2)
Get 503. Lol.
Too big (Score:5, Insightful)
Re: (Score:3)
Of course not. Its size is one of its design problem: it's too big, and doing too many things.
No. The "problem" here is people looking at the main systemd repo and declaring it to be too big and doing too many things. In reality it's a system made up of many modules. Some grew. some didn't. Each module does what it needs.
You're just upset at the name of the package, and likely the programmer involved.
In other news the Linux kernel is 8million lines. It needs to be thrown out because number of lines is the only important metric amirite?
Re: (Score:3)
Take out one module, and the whole system collapses. So, no, systemd is not modular at all. The Linux kernel contains many modules which can be left out completely, and it will still cheerfully boot.
Try, for example, deleting journald (this means: actually removing it from the system's disk) and replacing it with something else, like syslog-ng.
Systemd will crash and take down the entire system with it.
Re: (Score:3)
Re: (Score:2)
now, ask yourself, how come that ALL mayor distro switched to systemd ?
Why would you ask yourself that? Just look it up. Their reasoning has been quite public. The reason given by almost all of them is that systemd solves problems they had, it makes maintenance for a distribution easier, it makes handling application starting easier. There's literally only upsides to a distro maintainer switching away from sysvinit. ... So many upsides that some distributions actually started their own project to replace it, projects with a similar initial core feature set to systemd.
and WHO has such magical power in the IT world ?
Anyone wh
Re: (Score:2)
now ask yourself, why is McDonald's the most popular restaurant in the world?
Re: (Score:2)
I am sorry, but your argument isn't logical.
You are implying "Argumentum ad populum" or appeal to popularity. Which is a logical fallacy that just because the general population likes it, it doesn't necessarily mean it is good.
Many of the distributions are branches of other distributions, often from Debian (Witch included Ubuntu, and then all the spinoffs of Ubuntu), Red Hat (Fedora, Suse), or SlackWare. And if the spin off becomes popular, they will often then bring features back into the core other one.
Re:Too big (Score:5, Interesting)
What would have been even nicer would be text logs, not binary logs that require journalctl to read, and where tail, less, and such were options. Something managing permissions of init scripts, including security contexts, resources, and all that could be done in a ton fewer lines of code as well.
The problem with such a mass of code is just by statistics alone, there are show-stopper bugs just waiting to be found. Has systemd ever been audited recently, or at all by a solid security team?
Something that is almost as privileged as a kernel needs to have as few lines as possible. Ideally, systemd should be broken up into a number of smaller programs that can be easily audited and certified. Otherwise, having a large blob of code, if someone managed to get systemd context, they basically have full kernel access.
It would be nice to see something smaller that is functionally identical to systemd take its place, but that ship has long since sailed with regard to mainstream Linux distributions.
Re: (Score:2)
How about s6?
It's what systemd was meant to be *initially*.
Re:Too big (Score:4, Informative)
What would have been even nicer would be text logs, not binary logs that require journalctl to read, and where tail, less, and such were options.
Systemd provides you literally that option. But I'm guessing you never bothered to RTFM.
The problem with such a mass of code is just by statistics alone, there are show-stopper bugs just waiting to be found.
So what? Systemd including all of its entirely optional modules which are independent of the init system itself is tiny compared to much of the software you're running. Systemd is full of bugs and they have been overwhelmingly irrelevant, the last security issue of significance scoring 6.2 and was 2 or 3 years ago (depending on which timezone you're in today).
Something that is almost as privileged as a kernel needs to have as few lines as possible.
It has few lines. Very few lines. TFS is talking about the systemd project not the component of it that runs with high privilege. Most of the code written by the systemd developers is not running highly critical system level privileges.
It would be nice to see something smaller that is functionally identical to systemd take its place, but that ship has long since sailed with regard to mainstream Linux distributions.
Functionally identical but smaller implies that you have identified inefficient coding practices. Have you? Or are you wishing for a magical unicorn?
Re: (Score:3)
One can build a search index over text logs just like with binary logs. It only requires consistency, not a binary format, in the original logs.
Re: (Score:2)
Then you have indexing as a separate step that can be out of sync with the actual log.
Yeah, you can do things differently, but really the whole binary log issue is a non-issue. It's a very simple format, and it's well documented. And there's an option in the config file to produce text logs if you want them, by the way.
Re:Too big (Score:5, Insightful)
No. You're wrong. On every count. Think about what you write before you click "Submit", dude.
You can create a search index on text logs as they are being written. You can write binary logs that need a separate indexing step. And the whole complaint about binary logs is that they require custom tools that are different than what people use for their other system administration and data processing work. It doesn't matter how simple or complex the format is, because it violates the Unix philosophy of simple tools that compose nicely -- which is systemd's core mistake that gets repeated over and over within its design.
Re: (Score:2)
Transactions on filesystems weren't a thing last time I checked, any index separate from the file is likely to be out of sync sometimes.
Which you mean? As far as I know all the old stuff still works.
Re:Too big (Score:5, Insightful)
Start doing what I told you already. One benefit of text logs is that it doesn't matter if the index gets wonky, or you decide you don't want it, or something ate the file -- you can easily rebuild it (in whole or in part). Your imaginary sync issues are imaginary!
All the standard Unix tools that work on standard text require preprocessing the logs with a custom tool that has no purpose except to make the logs readable. Just make the logs readable in the first place and everyone's life gets simpler.
Re: (Score:3)
So would text logs, if systemd did them that way. Here's what a log entry looks like:
Re: (Score:3)
Yes, we get that you do not understand that you are pitching solutions for things that 99% of people simply do not want as "good ideas". For example: If someone wants to ship logs, inject hostnames at that point.
Unfortunately, this monstrosity was jammed down the throats of a bunch of distros by a Microsoft mole and his collaborators, so it is really hard to pick a good solution.
Re: (Score:2)
Date format has been its own issue. RFC 3339 is stable and provides regexp compatible outputs, especially if the logs are consistently published with GMT rather than other timezones.
Re: (Score:2)
"What happened on Friday between 2 and 4 AM is a hard question to answer with text logs."
LOL Hardly!
Re: (Score:2)
Depending on log rotation, you have to locate the right file, eg, .log, or .log.1, or .log.gz, then you may or not have to decompress it, then you have to craft a regular expression to match a time range.
If you're interesting in more than one service, you'll have to look at multiple logs side by side or have one huge log that duplicates everything.
Or you could use journalctl -S '2022-12-30 2:00:00' -U '2022-12-30 4:00:00' and not care about any of that.
Re: (Score:3)
SysV scripts were fine. You could read them and understand them, they made sense, and were easily customizable.
Re: (Score:2)
They were single threaded, and prone to the failure of a single init script causing the failure to complete the others. They were also prone to crashes of the relevant daemon requiring manual restarts. When the daemon was vital, such as sshd for remote access, the problem was especially notable.
Re: (Score:3)
And if had genuinely stopped there, I think it is highly unlikely that it would have caused the divisions in the commujnity that it did.
Inside job (Score:5, Insightful)
I'm pretty sure that Lennart has worked for Microsoft the entire time...
Re: Inside job (Score:4)
Yep.
Embrace, Extend, Extinguish has always been Microsoft's plan.
Re: Inside job (Score:4)
Next step: A Microsoft signed UKI.
Next, next step: systemd-boot only runs in a VM under Windows.
Re: (Score:3)
It has been the plan of Redhat, GNU, and many other players. M.s. is simply the only one who named it so.
It is what players do regardless of it being a plan to begin with. Standards only cover a bare minimum. The GHC has always embraced and extended the Haskell standard, and extinguished most competitors due to simply offering more.
Re: (Score:3)
Does Microsoft really have a shot of "extinguishing" Linux in the year of our lord 2023 when Linux in it's various forms runs on like 70% or some shit of the worlds computing devices?
Just because Linux will never compete in the general desktop space does not mean MS will ever be able to take the server crown from Linux. It's too efficent, too adaptable and too entrenched.
Re: (Score:2)
MS really doesn't have to bother doing much with Linux. If they can get people running Linux in Azure, they get their money. In fact, MS makes a ton of money on every single Android device sold due to patents, so Linux is a cash cow for them.
The desktop isn't a major revenue stream anymore, so they don't really care about Linux on the desktop, or even in the server room. They make their money on backend apps, and cloud support. For example, Exchange is likely to all but disappear to be replaced by M365.
Re: (Score:2)
I think this is spot on. This ain't the Ballmer/Gates Microsoft from the late 90's and 2000's
We don't have to think MS is "good" today but they are different and their position on Linux is an acknowledgement of that and of Linux's success.
Microsoft can make money with the first 2 "E"s, they don't need the 3rd, it's actually detrimental to their model now.
Re: (Score:2)
ahhhhhh .... conspiracy theory !!!!
just kidding
you are right
Re: (Score:2)
You can't be sure. He might just really think that's the proper way to do things. That I disagree, doesn't really mean he actually has malign purposes.
Re: (Score:2)
I'm pretty sure that Lennart has worked for Microsoft the entire time...
By producing a product that people (distribution maintainers) were asking for? That doesn't sound like a sane conspiracy to me.
Re: (Score:2)
Ney year entertainment (Score:2)
Nice, a New Years entertainment thread for me while I am doing system support tonight.
If you want to avoid systemd, there are distros out there. I am a Slackware user, *not* because it does not have systemd but because on the whole I like Slackware better then others.
Re: (Score:2)
Growth (Score:2)
Nice way of saying. (Score:5, Insightful)
Poettering (Score:3)
No real shock he took a job at Microsoft. Eventually the not invented here crowd is going to dislike the many plain text files in /etc and come up with a new solution. Maybe put them all in a flat database file. Oh no you won't be able to edit it yourself of course. You'll need a graphical interface. This interface must be run from a docker container because it's impossible to build outside of the author's dev environment.
Re: (Score:2)
Re: (Score:2)
To me systemd looks like an artifact of Linux growth. New users and programmers came to Linux from Windows and found things unfamiliar. My opinion is that systemd makes Linux look more like Windows, more familiar, and therefore "better". (I'm largely ignorant on this, but in terms of function and form, how far is systemd from "svchost.exe"?)
What I dislike about systemd is its appropriation of already operational function and the attitude behind it. As I was reading news tidbits on Phoronix over the past
Re: (Score:2, Interesting)
No, systemd exists because lots of Unix is frankly obsolete crap.
The old way of doing things worked back when a computer was a thing that sat in a server room, resources were scarce, the world was static, and the usage model was that the sysadmin owned the system, and the user was allowed to use it.
So for instance stuff like /etc/resolv.conf being a text file that root had to modify by hand made perfect sense -- right up to when laptops were invented, and a computer suddenly had to expect to dynamically rec
Re:Poettering (Score:5, Interesting)
No, systemd exists because lots of Unix is frankly obsolete crap.
What's funny is that you know little enough about computing to think that this is true, but still think you have something to teach us about Unix. Doing everything in flat files was an advance, in fact a strength, over more structured file formats. You can still read and write them with APIs, but you can also do it with shell scripts. Those init scripts people always complain about are core system functionality, not some hacky addition.
Now a program had to overwrite your config files, so sucks to be you if that deletes some valuable comment, or it causes you to lose some manual configuration. It's effectively the worst of both worlds.
We moved to config dirs with flat files in them to augment the flat files, which made use of another core Unix feature, the hierarchical filesystem. A lot of them weren't back then.
Systemd in good part represents realizing that the old way of doing things just doesn't really cut it anymore.
Systemd is basically just a lame-assed layer on top of the old things, it's still just doing things the old way, but with more obfuscation on top of it.
Re: (Score:3)
Times change, get used to it. It was an advance until it wasn't anymore.
No, init scripts are a horrible hack. I've worked with enough of them, that I'll be very
Re: (Score:3)
My workstation in the office is sitting in the same place for 20 years (if we ignore HW upgrades). The same is true for my home machine. So why is a thing designed for mobile devices shoven down my throat?
If you use a notebook and take it to a different place 10 times a week, you travel, you move between a showroom, your desk and the conference room, you need to connect in the train and airport ... be my guest a
Re: (Score:2)
Because your use case isn't considered the primary one anymore. Your use case was solved decades ago.
Modern distributions sell themselves to users who use laptops, and VPNs, and virtualization setups where machines come and go. They also want more users, which means removing all possible roadblocks a new
Re: (Score:3)
No, systemd exists because lots of Unix is frankly obsolete crap.
"Those who don't understand Unix are condemned to reinvent it, poorly" applies to systemd.
Re: (Score:3)
It already exists; it's called dconf, and it's horrible to work with and most GNOME applications depend on it and store their configuration in that binary config store. Sharing configuration to another machine or versioning it is no longer a trivial file system operation but an arcanely complex command and discoverability is nonexistent as the entire thing is of course designed on the assumption of a graphical user interface because being slow and inefficient is what Freedesktop considers “modern
Worst thing that has happened to Linux ever (Score:5, Interesting)
Re: (Score:3)
It is trying to move Linux back to old windoze ways.
Pretty much. And a lot of idiots are cheering it onwards.
Re: (Score:2)
There are lots and lots of reasons for systemd existing, mostly solving various wrinkles in the unix model, which was built for large boxes that sat forever in the same place, and which is unsuitable for a laptop used by a normal person.
And why would you want a full blown ntpd on most systems? 99% of people want a simple, foolproof way of synchronizing time. In fact due to SSL accurate time is now a requirement and no longer optional.
Re: (Score:2, Insightful)
It's called "resistance to change".
People have learned things a certain way and are kicking and screaming while being dragged against their will towards something new.
Now, I am not knowledgeable enough to lean either way and bring well-documented arguments about whether X is better than Y, just analyzing reactions here and elsewhere, and most if not all negative comments are akin to pearl clutching: "oh, there's warnings in this and that conf file".
Re: (Score:2)
There's a simple project out there called sntpd, all it does is fetch the time, and set the system clock accordingly, from a ntp server. It's not 1.7 million lines of spaghetti code with root privileges, so it's also not a security disaster waiting to happen.
Re: (Score:2)
First there is not much of a reason for it in the first place. There are well established, workable methods
Anonymous horse carriage driver about automobiles, cca 1905
Re: (Score:2, Insightful)
If that's true, then why was it adopted so quickly by so many major distributions? Either IBM d
Re:Worst thing that has happened to Linux ever (Score:4)
If that's true, then why was it adopted so quickly by so many major distributions?
GNOME
Re: (Score:2)
"The way we always did it" - the lowest IQ form of dismissing an argument.
Hint: Your "established" way of doing it spawned close to 20 projects to address its flaws and deficiencies, some of which were adopted by major distributions before systemd came along (e.g. Upstart), and some of which still are in use today (e.g. OpenRC).
The single dumbest thing about this entire debate is people like you suggesting it's bad for Linux for more choice to exist in the market. You don't like systemd? Use one of the dist
Ugh. (Score:5, Interesting)
I can't say I know anyone who likes systemd. Everything systemd does could have been done by a few small programs that the init scripts call to support desktop environments. Everything about systemd is the wrong way.
Re:Ugh. (Score:4)
Indeed. My take is that people that "like" systemd have never really worked with it. I took a look and stayed away. There is just too many things done the wrong way, starting with the architecture.
Re: (Score:3)
use devuan.org
Re:Ugh. (Score:5, Interesting)
Re: (Score:2)
Just pure garbage syntax - a script is so much more elegant.
I wouldn't describe an average of 100+ lines of script being replaced by a 3 line unit file as "elegant". It's quite ironic that people complain about systemd's complexity and at the same time complain about unit files compared to long scripts.
Don't like it, don't use it.
Dependency management, log aggregation (Score:2)
Re: (Score:2)
It is a sign of incompetence. Reliable and secure code is lean and mean. Systemd is complex and bloated.
Re: Dependency management, log aggregation (Score:2)
Re: (Score:2)
It is what actual engineers get taught. Unfortunately we still do not have many of those in the software space. Hence atrocities like systemd and a lot of morons that cheer for it. Calling it "marketing BS" just marks you as utterly clueless. I guess you have never heard of the KISS principle either. Or of the Second System Effect of which systemd is a stellar example.
Re: (Score:2)
It is a sign of incompetence. Reliable and secure code is lean and mean. Systemd is complex and bloated.
Reliable and secure code also has a ridiculous amount of error checking.
Not to mention how many of those lines are comments? Examples? Unit tests? Random configs or whatever you call this [github.com]?
I'm not sure line count is that useful as a metric.
Years ago I heard a story how at an open source conference in Australia someone pointed out how autoconf had 3x as much code to deal with Linux as it did to deal with OpenBSD. A prominent (I forget the name) OpenBSD hacker claimed this was evidence that Linux was complex a
Re: (Score:3)
It is a sign of incompetence. Reliable and secure code is lean and mean. Systemd is complex and bloated.
Reliable and secure code also has a ridiculous amount of error checking.
Only when done by amateurs. Or rather bloody amateurs. Regular amateurs may have heard of KISS. Of course that "ridiculous amount of error checking" does lead neither to security nor to reliability. And that insight is not even from the sofware age. It is far older. There is no replacement for simplicity in engineering and anybody trying to get reliability (or security) by increasing complexity significantly will fail.
Re:Dependency management, log aggregation (Score:4, Informative)
Yeah, that's the kind of crap code I'm talking about. Look at action_dissect(). It's a 200 line function. Why do they have a 200 line function? No good reason, it's not particularly organized.
Each function should do one thing and one thing well. Using shorter functions forces you to think more about how to organize things, which the systemd people really need to do.
Re: (Score:2)
I like that systemd has dependency management,
Good.
log aggregation,
Not so good. Logs in other than text format are useless if you are trying to repair a system without the requisite binary log tools.
a cron scheduler
Already have one of those. It couldn't be simpler to use. Anyone who can't figure it out needs to put the computer back in the box and return it to the store from which it was purchased.
and users can manage their own services.
What sort of services? Network facing? On any competently administrated system, this should attract the attention (and rage) of the admins responsible for system security. You may just have
Lost opportunity (Score:2)
As a reminder, this year Lennart joined Microsoft as one of the surprises for 2022."
And sadly, he didn't take systemd with him.
That is a bad sign (Score:2)
In a central OS component the absolute last thing you want is lots of changes and growth of the codebase. Apparently, Poettering is determined to drag systemd down to the level of MS windows were things are never reliable and change all the time.
Fortunately, I have still managed to avoid this crap and I have yet to find any problems with SysVinit and OpenRC. All the "arguments" given by the useful idiots cheering systemd onwards are simply bogus.
I feel reassured (Score:2)
A couple of years back I would have predicted that all the ranting and raving over systemd would have been over by 2023. But apparently not, and the same old same old complaints in forums show up over and over and over again.
It's bloated. It's taking over everything. It's buggy. Logs not in text. I hate systemctl. It isn't the way I do things!
Yeah, right. Those /etc/rc* skills you got were hard won decades ago and you ain't giving them up for no reason nohow. True bootstrappy individualism should
Major systemd bugs do not get fixed (Score:3, Interesting)
Log line dropping behavior has been reported for:
and suggested workarounds like removing the above journalctl options and changing
to not output any log lines is not heeded. So basic features are currently unreliable and cause developers/administrators considerable confusion when debugging/administering unit files.
See https://github.com/systemd/systemd/issues/2913 [github.com]
Can anyone explain to me what systemd is for? (Score:2)
Maybe then I will appreciate it. Like I wrote a daemon and a shell script that handles its start and stop, where stop writes a special command to its socket and waits for process to exit. How will systemd improve on what I have?
Or I want to run a command to index my photos to thumbnails once a day. What benefit do I get compared to cron?
Or logging - do I get splunk like search tools for everything logged in the system and a simple way to configure what I want to keep and for how long?
If I don't get anything
In case you feel like running away from systemd... (Score:3)
There are various distributions which do not have systemd, I'll list some of them for you guys:
* Gentoo Linux (well... ok, it's possible to install systemd, but you can also perfectly live without) -> openrc (huge and friendly community)
* Artix -> openrc, runit, s6: take your pick (also: take your pick what desktop you want), Arch based.
* Void Linux -> runit, really stable too. As if that's not enough it also has a small memory footprint.
* PCLinuxOS -> openrc(?), desktop (kde) really small footprint (built for old machines, desktop)
* Devuan -> fork of Debian, sysvinit
* Tiny Core Linux -> ok, your potato from 20 years ago (that somehow survived) should still be able to run this... with some performance.
* Venom Linux -> no systemd smell, ever.
Re: systemd = garbage (Score:2)
init forever
Re:Aaaaaannnnnddd 6 posts so far... (Score:4)
Re: (Score:2)
Let's face it. "constructive" and "positive" debate around systemd hasn't existed. There's more ignorance around this than there was covid misinformation the past 2 years.
Heck this entire article is flamebait as it implies that systemd is this single mega project with 1.7 million lines of code that keeps growing, instead of ... reality ... a bunch of modules most of them optional and independent having the systemd- name in them. This would be like AMD releasing a new graphics card, adding a driver kernel mo
Re: (Score:2)
All I care about is that someone sits down and does an audit and certification process on the systemd code. Supposedly it has happened, maybe ten years ago around the Red Hat 7 days, but I've not seen anything recent.
Re: (Score:2, Interesting)
All I know is my system shouldn't have the audacity to tell me what files I'm allowed to edit. Open up resolv.conf and it says not to edit this file directly. Excuse me? When did this happen? I'm an adult and can specify my own DNS servers, thank you. Now without googling, tell me how you change an ip address in Ubuntu? Is it, network manager, connman, netplan, or something else this week? Hell even setting an address with "ip" takes extra steps compared to ifconfig.
Re: (Score:2)
That's hardly specific to systemd. Simply general Freedesktop philosophy. ## DO NOT EDIT THIS FILE everywhere.
## DO NOT EDIT THIS FILE is a blight that existed long ere systemd was even conceived.
Re: (Score:2)
So you're saying that systemd had plenty of opportunities to learn from errors of the past, and didn't?
I guess that covers about 90% of the criticisms of it, in some form or another....
Re: (Score:2)
Systemd had many opportunities to simply not bother adding unwelcome, unnecessary, and even dangerous capabilities.
Re: (Score:3)
Wait until you try to edit grub.cfg! This 'we edited the file so you should not' type of message is probably the first comment in half of
Re: (Score:2)
It happened pretty much when DHCP and laptops were invented.
See, most networks have their own DHCP servers, which means that whenever you hook up to a LAN or WiFi somewhere, your resolv.conf gets overwritten. This at this point is absolutely ancient, if there's anything new about it is that they bothered to add the warning.
Re: (Score:2)
Re: (Score:2)
It is just the same old appeal-to-emotion approach that systemd was pushed with originally. They use it because they do not have any sound technological arguments for it and still do not.
Re: (Score:2)
They actually have a reason. Mr. Pottering believes in an "immutable" architecturre, one with no configurations whatsoever edited in "/etc/". The result is turning /etc/resolv.conf into a symlink into the systemd integrated network configuration tools. This has broken a _lot_ of tools.
Re: (Score:3)
This kind of elitist attitude is exactly why most people wouldn't touch Linux with a 10-foot pole.
Re: (Score:3)
Soo, you want "Design by Moron" in technological systems that can do real damage?
No, I think he just would like to be able to be able to get help with a problem without being insulted by some condescending asshole. I've been a Linux user for 25 years, and it's better than it used to be, but there are still a lot of ill-tempered twats out there.
Re: (Score:3)
Exactly this.
When I was younger (that's more than 25 years ago), I had my first run-ins with Linux, which was horribly difficult (for me, at the time) to install and configure. I joined several Linux mailing lists and I was horrified by the cesspool of toxicity found there.
Every couple years since, I made one more attempt and got out as quickly as possible, for the same reason.
About 8 years ago I have decided that publicly available knowledge would allow me to learn Linux without having to interact with the
Re: (Score:2)
No, I'd just like people like you to make a bit of effort to be just a dat less that complete arseholes and bring clear, detailed and logical arguments, rather than "ad hominem" attacks, to prove their point.
Re: (Score:3)
Because that article from 2014 is still relevant somehow...?
Let me find that newspaper article from 1836 about trains being shit, or that article from 1904 about cars being crap...