What is the biggest problem for linux adoption?

I can set up an Active Directory domain controller on Ubuntu in 5 minutes lol.

This entire thread got derailed from “Commercial viability of Linux” to “Examples of how Linux isn’t Windows”

3 Likes

So… We should have Powershell in our arsenal to manage our non-existent Windows boxes? Gotcha!

Can you please provide a few real-world examples where Powershell would help me in my Linux life administering my Linux boxes though? I am certain it’s great in Windows, but I don’t think I know of any good use cases otherwise.

One has python pre-installed and is native to the platform. The other… Does not. Thought that was obvious? Also, I have yet to find a python environment that stays stable on a Windows install, but granted I have not looked extensively for a solution.

At the end of the day, Powershell is a capable scripting environment, but far too verbose to be used as a proper shell (e.g. where you actually do system administration).

  1. Linux is usually not preinstalled on computers from Walmart/Best Buy/Whatever other big box store.
  2. It’s not unified.
  3. Marketing money.

IMO

3 Likes

#2 Is a rough one, the separation of distributions are, in my view, a huge positive. However, that kind of structure (or lack of structure) is very hard to market.
What could be done to unify the platform without sacrificing that?

:thinking: I’m issued a Macbook Pro. I have Windows Active Directory/DNS, DHCP, IIS, MS SQL, and SCCM servers. I also have Linux Apache, Nginx, MongoDB, Prometheus, and Jenkins servers.

To manage this environment I can either:

RDP into a Windows jump box and use the Terminal on OS X.

Use PowerShell and Terminal on OS X.


Sure! If you came from a job as a Windows Admin and now you manage primarily Linux on, let’s say, Fedora, this would definitely help ease you into the position. You can use PowerShell as your primary driver as well as learn Bash simultaneously.

Or, let’s say you come from a .NET world and everything is getting converted to .NET Core. You can fire up PowerShell on your Linux box and see what scripts need to get converted or if they still work natively.

The link you provided seems like the issue would be solved by using venv. :thinking: I did not read the whole article, though. Also, I’ve experienced that same issue on CentOS and OS X. It does not seem to be isolated to Windows.

I don’t understand the confusion. There are thousands of administrators taking advantage of PowerShell that “actually do system administration”.

Configuring DNS on CentOS and configuring DNS on Server Core 2016 takes an equal amount of skill. PowerShell isn’t Bash, that’s like saying “Well, C# is for actual programmers where as JavaScript is okay.”

They’re both programmers using a tool. It’s really that simple. :man_shrugging:

3 Likes

Powershell is on linux too I thought

Edit: it is

1 Like

It is… mostly. PowerShell Core is missing some core functionality (heh) like AD modules (this was back when I used it).

There have been an insane amount of progress made. DBATools is now 90% compatible with PowerShell Core and it now has native ssh.

1 Like

It doesn’t say Poweshell core anywhere, just says powershell :thinking:

1 Like

In the readme file:

M$ deceived ye again :wink:

1 Like

I knew I should’ve learned to read!
No pictures 0/10

2 Likes

I think it’s great that there are so many choices as well, but I also think it will create confusion for somebody who isn’t familiar with Linux.

We don’t even have to look past our own forum for that, whenever somebody asks what distro they should use they can easily get six or more recommendations with different reasonings. Compared to Windows or Mac where currently there’s really only one choice for each.

Packages are another factor that even still confuses me a little. Which package manager goes with what, where can I get packages, what about my current apps, why won’t this package work Manjaro even though it’s from Ubuntu (both Linux)? Et cetera. I know those answers, but my parents certainly don’t.

Here’s an article from HBR that I think is relevant to Linux adoption, specifically point 4, although the other points are as important:

It’s hard typing wordy posts on my phone at work.

2 Likes

Do you think snaps and flatpak could help in this area? Ultimately we need to get to a point where the distro doesn’t matter, I think were almost there. Though I think even once we get there people will still argue over minute differences.

Yeah, but why should I use Powershell over bash then? I get that it’s a great tool on Windows for Windows-API and Windows-binaries though.

All this suggest to me that Powershell is about as capable as bash/python on Linux. It’s therefore not better nor worse, just… Different, and Windows-y different. Which, to me, smells wierd, but that’s just personal preference. :wink:

I still think the Linux command line provides a more rich environment to get stuff done, but it’s nice to see Windows is slowly getting there. Now if you could only boot directly into powershell or bash when you ssh into a Windows box… :thinking:

1 Like

If distros could commit to one thing like flatpaks or snaps would be great, but I fear that it’ll be like the current situation where seemingly every distro has their own thing.

At which point we’d be back to square one. Except Linux really does need less standards for wider adoption. Not saying limit it to one, but like two or three would be much more helpful than dozens.

I think that is one of the reasons why Macs have gained traction even with people that never grew up with one for school, It Just Works™. From computer to computer and even from phone to computer it’s clean, consistent, and unified. iPhones certainly help, but that’s beside the point.

2 Likes

Actually I’m at the opposite end of the spectrum; Snaps/Flats are a terrible idea, and I hope they never catch on for most software. While they work short term, long term they are terrible.

The reason there are package managers, why we use PPAs, why we use repositories and so on - is because it allows for dynamic libraries. This is what keeps Linux distros small and lightweight, but it is also a big part of why Linux is as secure as it is. When bugs are discovered, the libraries are fixed once for everything. With Snaps / Flats, every package must be updated - and then we go back in time to how Windows work.

While there are some decent use cases for Snaps / Flats, it should always be the exception, not the rule. Or, in short - regular packages are a feature, not a bug!

1 Like

snaps and flatpak aren’t designed to replace repos, they are to supplement repos. PPAs are horrible.

We are talking about linux adoption, so you have to step outside of your own perspective.

flatpaks for example work pretty well, and can (at some point in the future if everyone adopts it) work everywhere. Updating and using them are seamless, its all integrated into the package managers. no need to adding urls to files, no need for repo incompatibility. click, install, use.

1 Like

So just off the top of my head - some of the things I frequently notice

  1. Documentation that doesn’t explain the thought process behind doing things.
  2. Distro Wars and Distro Hopping
  3. Distro & UI Fragmentation
  4. Walkthrough guides with excessive handholding - where readers don’t learn much besides primarily how to copy paste thoughtlessly and hope it works
  5. Hardware support that lags behind other operating systems and the fact that vendors are reluctant to open source certain parts of their device code to implement full linux support.
  6. Core Technology stacks (Rendering & Windowing (X) & sound (pulse) ) that are not up to standard/flawed with modern hardware features or immature (wayland / mir)
  7. Lack of adoption on the desktop as a prebuilt OS - largely as a result of windows market forces, limited commercial desktop interest in linux on the desktop and prior listed problem points

Regarding developing for Linux:

  1. Requires rebuilding code base entirely differently from windows (if you have an existing project)
  2. Added development effort to support 2 platforms (windows is almost always the larger user base for desktop software)
  3. Rather wonky and ‘convoluted’ (if I could call it that" debugging tools)
    • Seriously gdb is NOT very intuitive to anyone that’s used most other debuggers before.
  4. OMG - my users could be using x ,y ,z distros and they all have potentially different versions of libraries that don’t match and different UI libs, etc
    • Flatpak / snap can & does solve this, but it’s simultaneously something I regard as a rather inelegant nasty hack/ workaround the distro centric “packaging problem”.
    • Effectively we’ve merely created a centralized package mirror with snap & flatpak.
1 Like

I just realised something similar came up in another topic.

If you have a chromebook, go look at the help app. Linux lacks that.

Its there in bits and pieces, but its not the best, it has the basics, good information, but can miss the “the help pages dont help” next step. The chromebooks point you to the dedicated chromebook forums for example if more help is needed, or a phone line (obviously not really an option here).

But what if the OS help pages gave you the right links to the right community forums for help, what if they integrated IRC (or other) chat right into the help. Need more help? Click here for the Fedora Help Chat Channel.

Part of the problem on that side is that the use-cases arent well defined, and there are gaps. leaving people not knowing where to go next.

We should brun them all except for a few. (the ones i like obviously)

This isnt to bad i dont think. You have KDE and GNOME. Yeah theres one of two interesting projects and we could probably pick one to be the 3rd “big one”, but i think thats OK. Discount everything else. No one cares. (and if you want it you know where you can find it).

The Linux world really should get behind an “official” hardware provider. My vote is Dell, they actually sell Linux machines, they have devs which help provide drivers and firmware. Push it more.

Systemd was great i think. Part of the fragmentation problem was solved in one hit, and that’s good thing.

Dell :heart:

1 Like

I mean I don’t really care how packages are handled, I was more referring to making the basics streamlined between distros. Like being able to jump from Manjaro to Fedora to Ubuntu and having them more or less handle things the same way would be great, but as it is it takes some effort. The way I am viewing Linux adoption is Linux in the hands of the masses, not just tech users like you or I. In my experience that means simplifying some things will go a long ways.