Hacker News new | past | comments | ask | show | jobs | submit login
Wasted Creativity in the GNU/Linux Distribution Diversity (write.as)
36 points by gnuarch 40 days ago | hide | past | web | favorite | 43 comments



Not this self-centered, ignorant complaint again.

Even if this time really is 100% wasted, it's not your wasted time, so it's none of your business anyway. Open source devs don't owe your their time, and you're not entitled to tell them what to work on.

But the time isn't wasted, in any case. If existing solutions met people's needs the alternatives wouldn't have been created. A huge project like a desktop manager or a Linux distribution doesn't get spun up on a whim because somebody doesn't like a desktop background, and it's telling that that's the only difference the author notices.


Note that this position is mutually exclusive with "It's open source so you can just fix it yourself when it breaks". I can go fix it myself, sure, but then it is my wasted time, that there are 8 or 10 major distros. We shouldn't have to fix the same bug so many times.

I don't want 8 or 10 distros. I only want one. I don't even care which one it is. I haven't contributed to that OS in 10 years because so much of my time ended up being wasted.

Or, if you take the position that "Developers don't owe you anything", then that's fair on its own, but it means that it's not an OS that I can depend on for anything. It's true there's no literal debt to be repaid but project maintainers are supposed to be good stewards. They can be replaced, but it's a slow and rare process. I can count on my fingers how many successful open source forks I've seen. Most projects will die before they'll change maintainers.

> If existing solutions met people's needs the alternatives wouldn't have been created.

That doesn't follow at all. There are plenty of reasons for starting an alternative software project which have nothing to do with meeting any user need. "Ego" is a common one -- it serves only one person's interest.


> Note that this position is mutually exclusive with "It's open source so you can just fix it yourself when it breaks". I can go fix it myself, sure, but then it is my wasted time, that there are 8 or 10 major distros. We shouldn't have to fix the same bug so many times.

They seem to be complimentary ideas. I might refuse to fix your bug or add the feature you want, but you're free to take my code and do it yourself. It's no more a waste of your time than it would be of the other dev's.

I also don't see what you're saying about fixing the same bug 10 times. That's not how it works. Each distro is responsible for itself. If you submit a fix to zsh, for example, it's up to each distro to go upstream and get that fix themselves. Same goes for bugs fixed by distro maintainers - it's awesome if they submit the fix to the upstream, but they're not obligated.

> Or, if you take the position that "Developers don't owe you anything", then that's fair on its own, but it means that it's not an OS that I can depend on for anything.

And maybe you shouldn't. "Buyer beware" should apply double when you're getting something for free.

Paid Linux distros exist for a reason. If you're depending on it for something important it's worth paying for Ubuntu or RHEL.

> That doesn't follow at all. There are plenty of reasons for starting an alternative software project which have nothing to do with meeting any user need. "Ego" is a common one -- it serves only one person's interest.

Exactly, there are plenty of reasons. Just because you think they're bad reasons doesn't mean they are, or that anybody has to listen to you.


> But the time isn't wasted, in any case. If existing solutions met people's needs the alternatives wouldn't have been created

This argument has merit. However, it also true that sometimes we developers are overly biased towards creating a new project in order to solve some set of problems, when it may be more effective and productive to get our hands dirty in an existing project.

Obviously, this isn't always the case, and there are other important factors. Perhaps the individuals involved have other goals besides simply addressing the issues in most practical way. But the mere existence of a project certainly doesn't demonstrate that its creation was the most effective method of solving the problem at hand.


>If existing solutions met people's needs the alternatives wouldn't have been created.

I'm not so sure about that. This modern age of change for change's sake must have resulted in hundreds if not thousands of javascript frameworks that appear to be even less diverse than the linux distributions.

Why are these flavourOfTheWeek.js frameworks being created? Is it really that someone's needs aren't being met or is it developer hubris, and the desire to be able to put Founder on one's CV?

I suspect much of the same attitude exists among those starting niche linux distros.


It's not self-centered or ignorant, it's an observation. Are observations inherently an attack now?

You're right that the time isn't wasted though. The fallacy here is that, if someone weren't working on <brand new totally superfluous Linux Distro X>, they would be contributing to Slackware or Knoppix. That definitely isn't what I've seen in the open-source world... People work on what interests them, and that usually means a new take on an old idea. The people who are logging into the Debian bug tracker every day looking for ways to improve Debian are a totally unrelated kind of programmer.


I agree, I view the Linux community the same way I view music. I like that every artist can make what they want, and I'm not owed anything from any of them - they aren't required to have their next album/distro be like their last one, and they aren't required to help me understand their work. Not everything needs to be catered to me. If I don't like an album or distro, that's my problem, not the musician's/developer's.


It's still a bit weird. Everybody wants to "contribute" but at some point there's so much noise and redundancy .. it's backward.

Note that I do understand we're all so different in context, understanding, drive, needs .. so bazaar will occur whether or not I like it.

I still believe we should have a tiny bit more integrative groups. Just a tad. No need to force people and resist their energy, but a bit of constructive interference.


> I still believe we should have a tiny bit more integrative groups. Just a tad. No need to force people and resist their energy, but a bit of constructive interference.

There's nothing stopping you from running a vanilla Ubuntu install.

What other people choose to do is still none of your business, though.


> What other people choose to do is still none of your business, though.

So they're not allowed to opine on it? What if that's what they chose to do?


Fair enough. Sometimes it's hard to tell if people are opining or proclaiming.


Well, perhaps so, but Linus, regarding Linux (although I think he meant mostly the kernel), said,

"I'm deadly serious: we humans have never been able to replicate something more complicated than what we ourselves are, yet natural selection did it without even thinking. Don't underestimate the power of survival of the fittest. And don't ever make the mistake that you can design something better than what you get from ruthless massively parallel trial-and-error with a feedback cycle. That's giving your intelligence much too much credit."

In that light, this strikes me like complaining that we have too many kinds of beetles. Linux distributions just work this way.


That is indeed a powerful mechanism -- but it still sucks to be a beetle, especially if you're not on a dominant branch of your family tree.

Imagine you had the job of recruiting beetles. You'll die soon, and your offspring (if you have any) probably will, too! What's the upside? Your species will be stronger for your death! That's a pretty tough sell. You're going to get the young and ambitious, and they're going to fight each other as much as they possibly can.

That's not the only way to win, or even the best. Pine trees are genetically successful, too, and they don't go out and murder each other.

I wouldn't complain that there are too many kinds of beetles, but I would complain that there are very few trees. It would be sad if beetles were the highest form of life on the planet.


Bah. Diversity is good for the ecosystem.

Do you complain about the plants and animals with only minor differences as being redundant, a waste of evolutionary effort?

Just a generation ago there was understanding that we had let too much power accumulate into too few large, multi-national corporations.

Then we got a chance to start over with the internet and new digital companies. How quickly we repeated the same mistakes. Now we have trillion dollar tech companies and wonder if they've gotten too big, too centralized.

What cognitive dissonance to bemoan the consolidated power of just a few FAANG companies while also complaining that the diversity in open source software isn't the kind of diversity you'd like to see.


I worked on Sun Solaris (it's now dead, I know), and now on Linux. This was like switching from a Universe with one Linux distro, to the current one with many.

The speed and priority of bug fixes and feature development was higher with one distro. Everyone was working on the one thing. Any bug ever found could go straight to _the_ best engineer to fix it, who could replicate it _immediately_ since they were running that distro as well, and they could make it a priority to fix as it affected _all_ customers.

Now consider many Linux distros. A user says "this doesn't compile on NiftyLinux". A) The developer hasn't even heard of NiftyLinux, and doesn't have immediate access to reproduce the bug. B) It's a low priority to fix, since most of the developer's users are on Ubuntu or CentOS.

I've felt this firsthand with the performance tools I've developed for Solaris and Linux. With Solaris I could provide better support. With Linux, there's bugs that are open for months or years for odd Linux distros that I don't have time to explore.


On one hand, time and effort is wasted doing the same thing 100 times with slight differences rather than just working on the same thing. But on the other hand, in my opinion these differences are very important. For example, you listed that only a few desktop environments are necessary. While that may be true for the vast majority of Linux users, the more fringe desktop environments/window managers are great. For example, I use i3wm daily and it's a pleasure to use. While some of the DEs listed in the article may have some tiling, the WMs designed specifically for tiling do it better (in my opinion). That's the glorious thing about there being many tools that do ostensibly the same thing, you can almost always find a tool that fits your niche use case.


But I wouldn’t class i3 as slightly different. Probably an argument can be made that the prevalence of similar DEs ensured the APIs and interfaces that allowed for an i3.


Good point, at least i3 or sway don't do their own distributions – afaik.


There is manjaro i3 edition, and if I recall correctly another more minor distro that does i3 by default, but for the most part it is user installed, yes.


A very top-down utilitarian way of thinking about it. So what if there is wasted effort, people want to do that, so they do.


And the most creative distribution IMO: NixOS isn't even mentioned.


There is 0 substance to that link. How did this hit the front page?


There's the rub. Many distributions hardly add any substance, yet the topic is so popular.


Because people wants to express themselves. Tweaking and repackaging feels like something for the young nerd or the tech saavy layman. It feels like making an OS and OSes are semi god for the computer crowd.


I agree, it's ridiculous. Hundreds of distros but all include the same 5-10 usable apps. Great, I can run GIMP on Ubuntu, Fedora, Solus, Mint, Elementary, etc. It's still GIMP.


I think many people contribute to Linux because there are all these extra distros. People have different needs, and gather around a different itch to scratch. If there wasn't a distro focusing on Rasberry Pi, performance or security, they wouldn't be there.

It's like saying there should only be 10 programming languages. Vast majority of programming languages never takes off, and many are very niche. But you never know which are going to take off and when.


I don't think the problem is too many options. This is one of the great benefits of the free market.

The biggest problem with Linux distros, IMHO, is that nobody has really created a packaged install that's simple for users to get running. Ubuntu is ahead of the pack but there are still many issues... creating a USB boot stick, using Ubuntu under Windows, accessing your network with a VPN -- those a just a few I've experienced.

Until a Linux distro gets to the level of packaged, simple installation like users enjoy with Windows and Macs, the operating system won't take off on the desktop. And that's a real shame. We need open source software more than ever.

Perhaps there's room in the market for just one more distro that can solve this!


> Until a Linux distro gets to the level of packaged, simple installation like users enjoy with Windows and Macs, the operating system won't take off on the desktop. And that's a real shame. We need open source software more than ever.

I haven't tried with Windows 10, but Linux installation has been as easy (often easier) as Windows for decades. If you keep it simple and take the whole disk they're about on par, if you want to do anything complicated like dual boot then windows won't even let you. Both have room for improvement on the partitioning UI's. We don't need another distro to solve it because it's a simple as it can get.

But most people don't install windows, it comes with the machine.

As for TFA, I wonder if their concerned about the amount of effort wasted with so many competing phone, laptop and car models.


> simple for users to get running.

This doesn't matter. Most users don't install operating systems, they just use what comes with their computer.

The real issue for most people is "Why would I use this"?


That's a very good point.


Am I alone in finding the linux install process much simpler than windows? At least windows vista and older were always a pain to install, debian, ubuntu, elementary and solus have all been dead simple to install. Heck even IpFire is so simple it hurts.

Sure you need to make a usb boot stick but there are so many graphical utilities to do just that I've never had to think much about it.


No, I have a much easier time installing Linux on a machine than I do Windows. I recently had to clean install Windows 10 on my laptop and it brought back a lot of painful memories of having to find all the drivers for my particular hardware, store them on a USB stick, and make sure to install them in the proper order with proper restarts in between installs.

In contrast installing Ubuntu just involves popping a USB stick into the same laptop and letting it run for a bit after answering some basic questions.

And don't get me started on how much productivity I've lost to Windows 10 quietly "updating" drivers and breaking my programs in the process.


It seems that the author is only aware of the existence of general-purpose distributions. They did not list any distribution with, say, different approach to packaging, or for use in HPC, security, NAS, embedded, you name it. Also, for some reason they cannot avoid listing more or less all well established general-purpose distributions, while that is really the list between which you are supposed to choose one or two if you care about not trying many similar things. It seems that the only criterion to be on this list is to be well known. Same for the DEs.


It's not wasted at all - people are learning and configuring, tweaking and optimizing their use of a computer. I for one really like the fact that Manjaro is an easy Arch for me. Is it wasted creativity now that I can focus on non-distro work since I found my distro?

Besides, who is the author to tell people to not waste creativity. It's remarkable to have this audacity. Even the notion of "wasted" creativity is just not nuanced.


Why not just Slackware?

The author's list got to be the list in mutual cooperation and competition with the list and others.


Not sure why, but this remembered me of Tiny Core Linux: https://distro.ibiblio.org/tinycorelinux/

It's a very different and clever way of deploying Linux.


¯\_(ツ)_/¯ People are free to do whatever the fuck they want with their time and creative energy.

Here’s to anyone who ever scratched an itch, created something and then gave away the fruits of their labour for free.

A thousand more distributions and desktop environments!


One can dream... Linux would've replaced Windows by now if that happened.


What sales force would have caused Linux to be the default on mainstream hardware?


I mean it is already ona number of Dells. If Canonical and Steam focused their efforts they could produce a reasonable offering. I wish Steam would not focus on an entire Distro and just focus on making a specific distros underlying OS usable on plenty of hardware. AMD had an opportunity to work with Steam and Canonical to plant itself as a serious Linux competitor to Nvidia too.


Google. Shipping Chrome OS.


Chrome OS is not Linux, or I should say GNU/Linux (although this name lost its meaning too). Just like Android is not Linux either.


This makes me want to spend some time creating a new distribution with its own desktop environment.




Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: