Distro Chooser
distrochooser.de
distrochooser.de
Anyway, any reasoning that divides distros between easy to use and hard to use is nonsense. Just because I'm used to Linux doesn't mean Ubuntu isn't the right distro for me. The questionnaire logic goes: well, you're capable of running this distro so you should run this distro.
Reality is if you're a developer you either use RHEL because that's what corporate chose or you use Ubuntu because that is the platform chosen by library devs and unless you use it as well you will waste time sorting out issues.
If you're a user then it doesn't matter what distro you use, they're all fine. If you want to play games then your best bet is Ubuntu.
My experience is completely opposite, I like to prototype and test stuff and both distros are shipping heavily outdated libraries and tools, even if I need to compile on some old compiler to deploy why would I miss out on compiler perf/usability improvements, run outdated debugger, etc.? The CI can let me know if I broke anything in the target environment, but I prefer to have my tools up-to-date, and updating on Ubuntu/RHEL is a PITA
Of course if you're determined with a lot of work you can recreate the environment by compiling the exact versions of tools and libraries required (if you're lucky the library will specify that, otherwise you'll have to divine it through magic). But when you're doing this for a living you probably have better things to do.
It's a lot easier to use the distro that already has the correct deps and just get a newer version of a tool that you want to use if you really think that the one in the distro repo is too stable for you. So much less work.
That's quite a jump from your previous sentences and I don't find that true at all. I've had no worse of a time on Debian, arch, or Manjaro for developing when compared to Ubuntu (and its various flavors and derivatives). I never used distro deps because of how horribly out of date, misconfigured, or incompatible they are
OK, so you've spent some time and figured out the problem - it needs this particular version of this obscure tool. OK. No problem, I'll just compile it myself. So after tracking down the source you download it and try to compile it yourself. That tool has it's own dependencies. OK. 10 recursions later (I hope you're keeping all this crap in a local dir and not just blindly overwriting your distro version!) and you compiled the tool.
make
Another library that needs compiling from source - see above. Repeat for each one and its dependencies.
make
Nope! Your python is too new! OK, fix the script to you use the correct (older) python.
make
Nope, GCC defaults have changed! You fix the makefiles.
make
Success! You've wasted a week of your time you could've spent working on your product.
I disagree about it being easier to upgrade tools, it's exactly what you described for backporting, you need to upgrade entire dependency chain because the repo versions are ancient.
Not at all true. Lots of people prefer Mint, for instance. I'm one, and one reason is for me the Ubuntu gui is just awful. Or what if you need really high security?
One of the main things that scares people off of Linux is the hundreds of distributions. Telling them it doesn't matter which one they pick isn't going to help with this.
They shouldn't, their security story is atrocious.
http://www.techrepublic.com/article/why-the-linux-mint-hack-...
Lots of existing users prefer Mint because they transitioned when a roughly equivalent looking GUI existed on the proprietary OS they were coming from. Everyone prefers familiarity - and that's fair enough. Ubuntu uses a GUI that's aimed to target new users [0] - the Ubuntu GUI is far easier to use for contemporary general computer users [1] - if you're an advanced user then the freedom of choice of Linux is still there in the X thousand packages in the archive!
To install a different desktop to the default is pretty simple: install the desktop package and edit the session file, or do it the old-fashioned way and edit a .xinitrc. Any advanced user should be able to do this, in fact it's part of the joy of investigating the options! For example, I use i3.
If you care about high-security then by definition you have to use one of the distributions which has a full-time security team. Looking at CVE dates and when updates are published would be revealing.
[0] Source, I used to work for Canonical. [1] You may find this surprising (I did) but the vast majority of end-users really don't do much with 'right click' and find virtual desktops utterly and totally confusing, they often don't even know about alt-tab.
By answering the 'Operating System knowledge' question 'I've used Mac OS and Windows.' my 13/25 dropped to just these two:
* OpenSuse * Slackware
So I just unchecked that question and moved on.
The choice is to follow RHEL because that's what the mothership forces you into, Ubuntu because that's what the cool kids are into, or Debian because you want to get work done.
I want to get work done.
Oh, please don't start that flame war! You can get work done any distro, if you are going to start distro-bashing, at least bring some proper arguments.
I'm not going to start that war, but please at least reason from a factual basis.
It's a silly war, Ubuntu reaches parts of the Linux (and other OS) that Debian doesn't target. If you think Debian is better you should value Ubuntu for being an on-ramp. Honestly, dividing FOSS (or Linux) users is a bad play - we're all part of the same family, there's more that joins us than divides us.
It seems like it always makes users' lives harder since so much software is targeted at Ubuntu first.
The explanations of what distinguishes each distribution are, to me, a wall of Linux jargon that is all familiar (Slackware was my primary OS for a while in the late '90s and early '00s), but also largely meaningless. From a user perspective, I'm interested in a tool like this precisely because I've got no idea whether I should prefer GNOME or Unity or KDE nowadays, and I don't really care to take the time to sift through the flamewars to find out. So when I'm dropped into the results and the top results are five different flavors of Ubuntu distinguished by what desktop environment they use, and the explanatory text for each largely boils down to, "Portmanteau-of-Ubuntu-and-$DesktopEnvironment is Ubuntu with $DesktopEnvironment", I'm left feeling like the tool has taken me on a short ride around the neighborhood and then dropped me off exactly where I started.
As for the site you could make the questions shorter/more easily readable. A radio button instead of a hyperlink would be much more intuitive for selection (and possibly auto-advancing to the next question once an option is selected).
Try KDE Neon, its Ubuntu based.
Plenty of useful feedback from other users here. The one idea I'd develop from what one other said was improving the results page descriptions. You could actually use a couple of question responses to mildly tailor the descriptions - eg, a less experienced user could have text with a lower level of assumptions made and a more experienced user would get responses more relevant to where they were.
Finally, I'd have preferred to know it was a probabilistic recommendation up-front (as was initially concerned that I'd rule out choices with some responses)
It put Arch (what I use) first, immediately followed by nixOS for me.
I'm really tempted to install and play with it, the premise sounds awesome, however time is not infinite
1) The English translation feels a little bumpy (as compared to the German).
2) The presentation of the various distros should be a little more considerate of newbies (see bunderbunder's earlier comment).
3) A few questions need rethinking, especially the one on the target manager (see blueside's comment) and the one about the community.
Otherwise: great idea, concisely executed.
I chose pacman, as I'm already using Arch, yet the best match for me was Debian with 87%.
Well, I switched the desktop environment from Unity to Mate, because my laptop can now be considered old.
I'm good at fixing computer issues myself 1) Agree 2) Disagree
Also, if you move on to the next question automatically after an answer is given instead of having to press "next" the process would be a lot quicker.
The primary purpose of this tool is to help someone choose an OS, which is a significant decision for many people. If a 5-10 minute survey to help answer this question is too much work, I don't think you are the target audience.
Why make something more effort than it needs to be...? You could use that argument to justify needlessly laborious processes for lots of things. Keep in mind also that the user might be skeptical the advice they get at the end is any good so if it looks like too much work they may give up.
I just wish BSDs were an option here (I know, I know, not the goal here). I'm bullish on OpenBSD.
pick: next button messes with the browser history.