So, the author tries to run linux on apple hardware. then he goes on installing a GUI on a server, misunderstands the difference between rpm and yum, judges package managers by command length.
Also mistakes GB for TB
This was a depressing read.
This was a depressing read.
Why blame someone for wanting to install linux on apple hardware (which btw is often better supported than other hardware)?
Why blame someone developing 3D acceleration for virtual machines for wanting to have a GUI on the server that contains test VMs?
Why blame someone for being annoyed having to know the difference between 3 package manager programs?
Why blame someone for talking about the usability of command line options, of which command length certainly is a part?
He's installing the mac mini server in order to run jenkins, redmine and backup. Why would you need a GUI for that?Fortunately he doesn't say anywhere he's going to run VMs on that 4GB ram machine.
I blame laziness. I blame not trying to understand why new programs were written instead of merging new features in old ones. Or failing to appreciate an OS that at least has a package manager which is not a third party one.
But then well, you know what they say. Unix IS user friendly, it just chooses its friends. And I like Linux like that.
whats the point in having a gui installed on a machine that by definition has no keyboard mouse or moniter (Aka a server)
they arent 3 "package managers". yum is a package manager, rpm is a package installer and dnf is a proposed replacement for yum. fyi https://en.m.wikipedia.org/wiki/Unix_philosophy
nearly all linux commands can be completed with around 4 key presses. e.g. aptTAB should give you apt-get
In fact I often set aliases for apt-get and apt-cache because of apt breaking the 4 key press / tab rule you raised.
- If he has a recent model Macbook its KNOWN to be problematic. Searching "2016 macbook linux" most of the top results are about how it doesn't work well. He specifically says he assumed because redhat offered it that it worked with linux. When you can't spend 30 seconds researching your development machine it might be your fault.
- Server guis are known to be limited, hard to automate and just generally inefficient desiring a gui for your server is normally a sign that you just aren't very good.
- Users don't have to use ANY command line package managers there is in fact a gui for that if they prefer. If they choose wholly of their own accord to use a command line package manager they probably ought to read the relevant help document https://fedoraproject.org/wiki/Package_management_system which makes the usage of the relevant tools pretty clear.
I don't think its untoward that red hat employees that wish to set up Fedora servers be expected to read documentation.
Another attack on users, this time pretty direct. I won't discuss the other points as it's again just shifting blame. The reality is that with Desktop Linux there's nobody taking responsibility that things actually work.
Fedora is first of all an unstable test bed for a commercial product running on hardware known to be problematic and the user is a red hat developer.
I realized I had more useful things to do than beta test rhel in 2011 and I started buying hardware that works well with linux in 2003.
Not knowing to how do this is certainly a forgivable thing in joe random business user but it becomes ridiculous when we are talking about a red hat dev. If a device doesn't work with its target platform be it windows, Linux, or mac you blame the oem.
In this case whose job do you imagine is it to donate free labor to a multi billion dollar juggernaut to ensure their machines work well with Linux?
Given the benefit Apple has derived from open source software it would be nice if they would help but they certainly have no real obligation nor do open source devs whose time is inherently a limited resource.
In the end Linux has plenty of room for improvement but supporting everyone's hardware for free with finite dev resources just isn't a rational or achievable goal and lampooning linux as a whole because fedora doesn't work on your mac isn't even useful dialogue.
Rational actors will react to this by reasonably picking hardware that is known to be well supported. Irrational actors will continue to complain on the Internet.
Why should OEMs work on supporting Linux when they sell their machines with other operating systems? That's just passing the blame.
The problem is that Linux is touted all the time as an alternative that is supposed to be even better, when the reality is that lots of things are broken and don't work out of the box. Not even on the most sold laptop models, which are macbooks.
> Rational actors will react to this by reasonably picking hardware that is known to be well supported. Irrational actors will continue to complain on the Internet.
People that want to actually get work done and do not want to or can not fiddle around will not use Linux.
And finding a fix can be elusive, when often times the answer is "your distro sucks, use $OTHER_DISTRO" or "use Google."
> Rational actors will react to this by reasonably picking hardware that is known to be well supported.
But what about when you pick hardware that is supposedly well supported and you still have problems? Run into that more times than I care to remember.
Nobody said linux or open source software was perfect.
Yes, do not inform potential users that they might be taken for a ride, don't mention that 7-years old hardware is not supported, or that Gnome3 doesn't even know how to switch back to the default display. I mean, we wouldn't want to have informed users am I right?
You are pretty much making legulere's point: hush hush no criticism in public, and even if it is valid please report it quietly.
So I assume you would agree if Apple or Microsoft had a similar attitude? (And no need to mention "yeah but with MS/Apple you paid, therefore...", paying entitles you to tech support, and we're only talking about informing users here.)
aside from apple stuff (who have made it quite clear they believe their users want a totally walled garden and are not interested in compatabilty with the rest of the world), the other two are now rare afaik.
so if it happens to you regularily, then yeah, very likely it is nothing more than your distro sucks.
- The oem is not obligated to support linux they aren't to blame
- The open source developers aren't obligated to provide free labor to make your hardware work they aren't to blame
- The user isn't obligated to use linux in the first place but absent a support contract or a promise from oem that the machine runs linux they aren't owed anything either.
Its like the open source software developers collectively held out to the community at large here take for free all this awesome software we have invested man decades in. It works on many but not all machines out there and a legion of users is waiting to help you for free make it work with your particular configuration and hardware and somehow a bunch of entitled individuals heard something on the order of
Here's all this awesome software we wrote for you for free it will replace all this software you are currently pay for on your existing hardware no matter what hardware you have and unicorns will fart rainbows in time with your keystrokes.
Then reacting to this promise that nobody made they get mad and make a linux sucks post on their blog.
That this rant comes from a red hat employee is quite frankly embarrassing.
In that context keeping people from buying machines that don't work with linux and complaining on the internet is equally impossible.
I don't understand how you believe people are obligated to do better work for free for you.
*Please note I'm not implying you or the author are stupid just users in general. This is why most people ought to buy complete hardware/software packages with support and experts should be able to use google and read docs.
I understand the idealistic goal that all software be turn-key simple for every user and every use case. But even if we assume it's a plausible, reachable goal, I just can't get on board with blaming the people working towards this for the failure of not being done. We're simply not there yet. We've barely even started.
I can simultaneously respect the notion that we should be striving towards it as a goal, while still having practical expectations towards the experts of today that have to interact with systems that are imperfect enough to require expert users.
We only have a few warm bubbles where interaction are easy. And there are still expectations of the users, the bar is just lower. Even the most-comfortable of tools makes demands of the users' literacy and ability to figure out how to carry out their particular goals. And we paradoxically put blame on expert users when they use those environments and want to be able, or (more ominously) allowed to do the kinds of things they know are possible but not supported.
So I can sympathize that the tools suck. I get livid at the limitations of my tools too. I can share in the frustration of fresh new roadblocks and suddenly having to learn something new. I hate hitting that kind of roadblock. But I have a hard time getting on board assigning blame when I'm a colleague of the people working to solving these problems, and I don't have the answer either.
I hope it should be clear that I also don't like putting blame on people, and I think there needs to be some collective taking responsibility by the linux community. If someone has problems the answer to that shouldn't be RTFM and PEBKAC, but questions if those issues can be averted directly in software.
Linux etc do not have that luxury so the warts out for everyone to show.
On top of that they are more often than not working blind, as they have limited to no access to the specs of the underlying hardware they are trying to support.
He also did a good job at the start of the article of acknowledging that Ubuntu gives a good desktop experience yet then goes on to rant about how bad Linux is in general after using one other, unrelated, distro.
The whole article reads incredibly ignorantly which I wouldn't have expected given his credentials. I can only assume he approached the exercise with an axe to grind so continued his tasks in a unintentionally negative way. In my experience, if someone is impatient when using unfamiliar platforms and/or wants to find problems with it, then they will find a way to expose problems regardless of the platform.
so the fact it struggled on what seems like ancient hardware is hardly a surprise.
wonder what his reaction will be when he finds he has to go through the whole ordeal again after they end of life his installation in 6 months.