Linux Sucks presentation interrupted by a higher being
-
-
"complexity" is not a criteria for judging software unless all-else-is-equal, and it's not in this case.
Most programmers love complexity. Fuck them.
Did "complex" just get a new meaning?
-
@blakeyrat said:
"complexity" is not a criteria for judging software unless all-else-is-equal, and it's not in this case.
Most programmers love complexity. Fuck them.
Did "complex" just get a new meaning?
Yes, the instant a certain someone needed it to have a new meaning.
how else can one move the goalposts so effectively?
-
-
I believe so; pretty sure this started in Sidebar.
Thing is, category changes never used to reset the read-state of posts…
-
category changes never used to reset the read-state of posts…
No repro. That is, this category change didn't reset any read-state for me.
-
Did "complex" just get a new meaning?
No? I don't see how the two statements are contradictory.
I don't like complexity because as I said above it makes the code harder to work with, etc. But I also am wise enough to realize that complexity doesn't necessarily make for a bad product. Lots of good products have super complex code behind them. Lots of really shitty products have simple code.
-
Same reason they're all using a CLI from 1975.
I still don't understand this whole "graphical is better" idea. It reminds me of this recent article: Dual Helix, i.e. when programming, we still pretty much write textual commands for the computer and don't click the source code together with a drag-and-drop-gui, simply because the textual interface is more suitable.Sure, a command-line interface might not be suitable for casual computer users who are not into programming - but that is not the target audience of this forum, is it?
-
Sure, a command-line interface might not be suitable for casual computer users who are not into programming - but that is not the target audience of this forum, is it?
Quick, someone tell Microsoft they wasted millions of dollars on Visual Studio!
-
Fine, I'll fork some Linux X terminal program and rename it Visual Command Line, so that we can settle this issue for once and for all. Ok?
-
Only when you've also added something akin to Blend
-
What's more likely is that all the Linux old-timers hate it not because it's worse, but because it's different. They hate change.
We hate change so much we have a gazillion different distro, and a million different GUI we can run on top of it.
Even worst, we can run multiple different GUI on the same computer, just so we can hate it more.
@blakeyrat said:Same reason they're all using a CLI from 1975. And the only reason people bring up "complexity" is because it's pretty embarrassing to admit that your only reason for hating it is you hate change.
From what I understand, the problem with SystemD is that, instead of trying to be the best init system there is, it's trying to do everything.I'm still waiting for the announcement where they integrate a GUI in it.
Then, it will only need a kernel to be complete.
-
From what I understand, the problem with SystemD is that, instead of trying to be the best init system there is, it's trying to do everything.
Which is bad because...?
-
@TimeBandit said:
From what I understand, the problem with SystemD is that, instead of trying to be the best init system there is, it's trying to do everything.
Which is bad because...?
Why would it be good?
-
The claim is that SystemD trying to do "everything" is a problem. I'm asking for the reasoning behind that claim. I'm not going to just take it on faith.
If you don't have an answer, just fuck off.
-
And you're ignoring the part where he says they're compromising quality for complexity?
-
If he said that, it wasn't in this thread.
-
There's probably a
Vimplugin for that.
-
If he said that, it wasn't in this thread.
Emphasis added:
@TimeBandit said:
From what I understand, the problem with SystemD is that, instead of trying to be the best init system there is, it's trying to do everything.
Which is bad because...?
-
That could not be more different than what you typed.
Boomzilla, I don't know what's up with you today. It's like you got replaced by a really poorly-written Boomzillbot or something. What the fuck man.
-
That could not be more different than what you typed.
What are you reading?
Boomzilla, I don't know what's up with you today. It's like you got replaced by a really poorly-written Boomzillbot or something. What the fuck man.
Dude. It's like you turned into the different animal here.
-
This year, the presentation was pretty boring... until the part linked below.
It sounds like the video I watched with Jobs in it saying the same stuff about Macintosh.
He's basically saying, "Throw shit at the wall works for us because there are more of us".
That does the image of open source no favours whatsoever.
Yep.
You're stupid if you think there's a problem with software that's falling way behind, because we have a potato, and we can potato farm.
Well, so can everyone else. You don't think any of these other companies have R&D?
The problem with this image is that it doesn't seem to be learning from why they are 'supposedly' failing. That's ignorant. I don't know if they really are failures, but how would you ever know if you ignore why you have a problem and keep throwing random shit on a potato as your philosophy.
Complexity is not a valid criteria unless literally all else is equal.
You're not going to change my mind with silly ideas about complexity, either.
Why, it's basically the same statement but said in reverse.
- Go for the simplest solution that solves the problem.
- Complexity isn't a valid argument for change, if it is the simplest solution that solves the problem.
These a coexistent. You can say one without implying the other. Once you achieve the first, the second is automatically true. Therefore saying "it's complex" alone isn't a valid argument unless you say, "It is more complex than necessary for the problem".
Complexity has a cost.
Software has a cost.
You're making the false assumption that it can always be made simpler. Complexity more than required to solve the problem is bad. If you said, make a calculator program, and I gave it to you, and you said, "it has too many buttons", you have to tell me which buttons aren't necessary. You can't just say, "I bet you can do this with less buttons". Sure, you can make it to where short pressing button A is a "+" operation, and long pressing button A is a "-" operation, but is that less complex? It sounds more complex to me. When you get to the point that simplifying it further creates more problems, you can no longer argue that complexity is a problem.
That's not much of a cost.
This is where I disagree.
You have a point in that complexity itself is not always wrong. However, it is the bane of our existence.
You know the old food pyramid, with sugar at the top. Use sparingly they said. That's complexity. You have to have complexity to have the energy you need to solve problems, but you need to be sure it's as small as possible.
but it's simultaneously 4 times more reliable so that cost is by far worthwhile.
I think where you guys are getting mixed up is this.
The rule of thumb is that complexity reduces robustness, generally.
However, that is a rule of thumb.
I can make an application simpler by communicating using UDP. It's connectionless, and therefore simpler, if you don't care about dropped packets, etc. I can make data storage simpler if I don't use parity bits. I can make enterprise software simpler if I don't worry about redundancy or backups.
Every bit of minimized complexity that is required to implement a new feature, is valid.
-
Well, so can everyone else. You don't think any of these other companies have R&D?
Not only that but they have EFFECTIVE R&D where they actually try their products out in front of human beings before releasing them to everybody everywhere.
-
A metaphor for what I said.
You can drill holes at the top of the boat if you want to make it lighter, but don't drill holes in the bottom.
-
You're making the false assumption that it can always be made simpler
You're making the false assumption that you understood my post.
-
You're making the false assumption that you understood my post.
In the OP a bunch of touchy feely arguments are made.
"Linux is terrible, except it's awesome because.... I use it.... " | Ok.... great argument.
"Linux is disaster.... to proprietary software.... " | Except it isn't. At most it slightly altered the behavior of proprietary software companies to shift focus onto support, but they still sell their software. Something that open source and proprietary now have in common. No disaster happened.
"Linux allowed me to blow up my rig.... but I was able to fix it just as easy" | Windows recovery disk. I didn't pay for Windows twice. Not sure what you're getting at here.You try to help a shitty OS join the 21st century, and the best this asshole can say about your product is, "well they're trying something different, I guess that's good, shrug?" What a jerkhole.
Literally the same idea I got from it.
I mean, the positive point he made was that we can fork, do our own stuff, make something great out of it.
Yeah, more stuff onto the pile of "blakey doesn't understand it."
Can you please explain?
Unless he had some subliminal messaging that people understand if they intimately know SystemD.
You're not going to change my mind with silly ideas about complexity, either.
His argument is that it became monolithic rather than modular. And I get that. But so has other things. Windows used to be DOS. Why are you adding a GUI to it? Why add support for a Mouse? Why include multimedia player? OMG You're making an entertainment system!!! And, he's right. It became something more than what it was.
So, as he said later. Fork it and go home. Use the simpler version.
But, if those features are in demand, then they're in demand, and therefore it becomes something more. Going back to @blackeyrat his analogy of the cars falls a little bit short. I would say it's more like this.
Someone developed a lightweight car designed to go fast and for speed.
Someone took that model and they added wings on it, and now it's a muscle car. Does it lose speed? Yes. But it has a different purpose now. It's no longer a derby car.That's the flaw that's being made here.
The OP did not demonstrate how the added complexity was not worth it. He simply damned the complexity. He did not show that the complexity was unnecessary.
Like I said earlier, every line of code is added complexity, so if complexity is bad, why would you ever add code. Well, you'd add code if it doesn't perform everything you want it to perform, if it's missing functionality. And unless he miscommunicated or I misunderstood, from what he's saying, SystemD is now more complex because it added functionality. Making it invalid to say it's too complex. Unless you can show how that complexity is more than necessary to add the new features.
You're post is that complexity is has a cost.
Yes, and that cost can be supported by argument that demonstrates that it was necessary to provide a feature that is in demand. The only argument the video makes is that, "It doesn't play nice with our concept of modularity". That's the real offense. If he had something more than that, then please show it. Please demonstrate how there was no demand for the added features, or that it became more complex than necessary for the added features.
You can't just say, "It's too complex", drop the mic and walk away.
That is @blakeyrat's point.
However, when you remove the context of the video. When you remove the fact that the video doesn't defend this statement, and you just look at complexity as a function in software. Yes, complexity is bad. If added complexity does not add value compared to other software that is less complex but has the same feature set, it is bad. That was not the complaint. The complaint was that, it didn't fit in with the Linux philosophy. And complexity was a hammed in argument for that. And all the religious developers are going to be like, "Yeah, I heard that complexity is evil too. You're right, SystemD is bad."
So yes, I get your argument.
In absence of everything else, you are right, complexity is always evil. But the fact of the matter is that complexity is the building blocks we use. The least complex program is below.
// no code
-
The only problem with SystemD is that UNIX admins are getting old and lazy, and do not want to learn few more commands[1]. So they whine, without knowing
they are talking about.- SystemD is simpler than fucking awful init
- SystemD is the Unix way, whatever the hell it means.
- There is absolutely nothing to invent in the fucking init, they should just spend their useless hours banging on something more relevant
- If one thinks the awful bash scripts, using
grepleft and right to parse inputs, are the elegant way, should get a special featured column in TD
[1] there also Solaris, BSD and other crap that I do not care about if they are fucked because no one gives a shit to keep their ancient
initworking.
-
The OP did not demonstrate how the added complexity was not worth it. He simply damned the complexity. He did not show that the complexity was unnecessary.
Yeah, that wasn't really the point of the talk, to go into details.
But to use Raymond Chen's analogy, new features start with -100 points.
And unless he miscommunicated or I misunderstood, from what he's saying, SystemD is now more complex because it added functionality.
I dunno...what functionality did it add to the system? It's my understanding that you now have one giant ball of systemd where before you had multiple components, plus some things are also different, like binary logs.
You can't just say, "It's too complex", drop the mic and walk away.
That is @blakeyrat's point.
Yes, he's an aggressively illiterate troll. I never said that. His argument was that worrying about complexity is stupid just because it makes it more difficult to fix bugs and add features in the future and dismisses any arguments to that effect. And then got all butthurt when someone pointed out how dumb that was.
So yes, I get your argument.
In absence of everything else, you are right, complexity is always evil. But the fact of the matter is that complexity is the building blocks we use. The least complex program is below.
Yes, you almost understand my argument, except that you think I'm saying never do anything complex. I'm just pointing out that added complexity is a valuable thing to avoid and you can't dismiss arguments based on that.
-
I don't know if they really are failures, but how would you ever know if you ignore why you have a problem and keep throwing random shit on a potato as your philosophy.
Dude.... Maybe OSS is actually a giant game of Mafia, and the Linux guys are stuck on day 1.
-
OSS is actually a giant game of Mafia, and the Linux guys are stuck on day 1.
that's.....
No words...
-
Maybe OSS is actually a giant game of Mafia, and the Linux guys are stuck on day 1.
@vote Stallman
-
I dunno...what functionality did it add to the system?
That's exactly my point, we don't know, based on the information in the video. He just goes on a rant about how adding shit is bad.
It's my understanding that you now have one giant ball of systemd where before you had multiple components
That's exactly it. He's more upset about it no longer being modular, than about it being complex.
And that's great as a philosophy. Modularity is a great thing, and it's definitely one of the selling points of Linux. But not every successful software uses that philosophy.
I don't call up Microsoft and tell them I really don't want a firewall in my OS installation. They provide a default one whether I want it or not. I don't call up Microsoft or Apple and tell them I want a different GUI. I don't fault them for that. They're selling a brand.
Linux is a brand too, it just targets more technical people that have the capacity to handle modularity. That doesn't make them a superior software brand. Both are filling a need.
When someone comes along and makes a distro of Linux that is a prescribed set of modules, that's a brand too.
His argument was that worrying about complexity is stupid just because it makes it more difficult to fix bugs and add features in the future and dismisses any arguments to that effect.
At that point I disagree with him.
It's... difficult, because I have to dissect his argument to pull out what's actually a great point.
The video does a terrible job of explaining anything, of providing any evidence for its claims, and relies on feels.
But hey, it was meant to do that. It was a motivational, not a debate.
But I do get frustrated, because that's the approach I get when talking to most people about Linux. It's good because it's the hipster software, it's anti-"the man". I honestly don't care about that stuff. Just trying to be objective here.
Yes, you almost understand my argument
No, I completely understand your argument.
I'm just pulling out of yours and blakey's a good point about the video.
I'm capable of doing that. Of pulling apart things and selecting good points and throwing away bad points.
And again, the good point is that the arguments about SystemD were "the feels" and cultish, not objective with demonstration about what is specifically wrong.
If I were in SystemD, based on that video, I would tell them to fuck off, because they haven't shown that the added complexity is unmerited.
Now, if I were to have the knowledge of SystemD and find out that it's full of shit, I'd change my mind.
Then he tears something apart without demonstrating objectively why, and then use that merely as a motivational tool to say, "Linux is great because we get to throw shit at the wall", which is not even unique. All great software companies have a strong R&D and throw shit at the wall. Then he goes on to make these feels arguments using two people that don't really have the expertise to be saying what they're saying, and one guy that has obvious biases.
But this is a motivational speech belonging to a group, and I've watched motivational speeches from Jobs and Microsoft (to show you how Jobs was the face of apple, every Apple motivational is Jobs; Microsoft's is a revolving door). They all go the same. No real arguments, just, "hey guys were great because of feelings".
So, in the end, is there a point to saying all this. Not really. But rarely is there a point to any of Blakey's rants. Looking at a motivational for logical arguments is not a very productive thing to do.
So, what's my take-away?
My point is that there is a balancing factor to the "complexity is bad" argument. Even if blakey gets it wrong.
My second point is that there is no demonstrated argument against SystemD, only one that begs the question, but this is a frivolous point.
