That explains it.
I read another good quote about Ubuntu on Reddit. I especially like their P.S., although I still like to try this or that distro from time to time.
I bought an entroware laptop last year. Last week I decided to format and install mint to see how it goes. It booted once and crashed after an update. I tried installing it again. I got the same result. I put Ubuntu back.
Ubuntu is like the Mac for Linux, it works the best.
PS: I think distro hopping is stupid. It’s like testing hammers for fun.
You need to test hammers… they dont all work equally well in your environment, and there are different hammers for different tasks.
Noone would use a sledgehammer to drive a nail, but you might use it to drive a rail spike.
I don’t think people who simply use their computers, and are perhaps recent fugitives from Windows, should feel any incentive to hop between distros. That means we all need to be informed which distros, if any, are (or soon will be) considered sufficiently user-friendly and reliable for widespread use at home, school and work.
The situation keeps changing. … because the world of computing is dynamic and turbulent … some would say chaotic. It is like trying to read the weather.
So how does a new user survive? You tell me.
From the day I first used computers ( in the 1960’s) it was an experience of things slipping off the horizon as fast as I learnt to use them.
The natural reaction is to try and freeze things… it does not work. … you have to go with the flow of developments… and that means changing hardware and distros and apps and work patterns , continually.
When you join Linux, you are joining an evolving set of entities. You have to evolve with it.
From the day I first used computers ( in the 1960’s)…
The 1960s were sixty years ago. It’s high time for operating systems to settle down. The choice must include at least one that doesn’t need an account with a commercial supplier and ensures that our personal and other data and communications are totally inaccessible to government and commercial agencies. This should not be too much to ask.
They have. The one thing that helped with ‘settling down’ was Unix. It did 2 things
- made OS’s portable across hardware
- standardised the utilities ( ie GNU core utilities and Posix)
Nothing much has improved since . Kernels have no standards, we rely on Linus to hold it stable. Addon apps can be anything. Package systems are still evolving. Init systems are a dogfight. I dont see any ‘settling down’ today… maybe GUI design has settled.
If operating systems settle down, they become more like concrete and, quickly, like stone. Windows and Mac are prime examples. Crl, trust me that in Linux we have not fossilized like those two but retained the freedom to try new things, new ways, and to make new mistakes. Try the one I’m using right now, MODICIA O.S. It’s pretty civilized–settled down–but it’s still way more flexible than the dark side operating systems. Giving up choice is like volunteering for prison.
The practical reality must be that Linux Desktop will present no prospect of ousting Windows and the other intrusive commercial alternatives until some organisational requirements typical of industry practice are met: for example, the provision of documented quality control, maintenance, rapid advice and repair services. Other essential design criteria such as ease of use have already been met by Linux Mint and some other distros.
I too was using commercial computers during the 1960s, not just Unix, but PDP8, UCSD Pascal and its HP variant, Wang and the rest.
I’m not interested in this relatively ancient history. What concerns me today is that, for example, when the Ubuntu libpciaccess update bug struck recently, an “ordinary” user would observe only that the screen becomes blank at startup. Ordinary users might not possess (or know about, or think of trying) a USB boot key, and anyway they might not know about changing the boot order. Ordinary users are not interested in an evolving set of entities. Android is evolving, at least in finding ways of driving you mad, but at least it works.
A special feature of current desktop Linux compared to Mac, Windows or ChromBook is that when something goes wrong you can’t yet take your machine back to the shop. Relevant advisory websites have varied approaches, but many propose nothing but geek stuff. A bit of a geek myself, I became particularly aware of the difficulties experienced by others through participation in a local free repair shop and in recycling old computers for people who can’t afford to pay.
Right from the start of electronic technology, people who were fascinated by it made or tried to make their own radio transmitter or receiver, television set, HiFi, etc. This was, and still is, an interesting hobby, useful at one time because manufactured equipment was beyond the means of many people. User-friendly Linux Desktop (as opposed to enterprise implementations) still seems to be somewhat at this hobby stage. There’s nothing at all wrong with that; making and playing with working model steam railway locomotives has long been an instructive and respectable long-standing rewarding pastime, as are so many other arts and craft activities. But that has nothing to do with the need for private individuals to be able to submit their online tax declaration.
Addendum 2025-09-28
I just came across another bug with Ubuntu – running version 25.04 on my new office machine. I need to run the commercial bibliographic reference package Mendeley Reference Manager from Elsevier, which is supplied only as an Appimage, and promptly discovered that Ubuntu has been unable to run these applications for a few years now. None of the miraculous solutions I found on the web worked.
Our free repair shop has an old recycled machine running Mx Linux 23.6, ready to be given away. The Mx package installer has AppImageLauncher (which Ubuntu doesn’t), and it took only a couple of minutes to get Mendeley recognised and running.
Mendeley is a major application used by researchers and others, both at work and at home (in retirement in my case). Whatever we think about the way Elsevier and rivals operate in the academic literature market, it’s free for home use. One can’t suddenly switch something else, because the desktop application synchronises with web storage and other facilities and, apart from other obstacles, hundreds of linked PDF files can’t easily be transferred. Distro hopping isn’t too easy as a solution, when you actually use your computer to the extent that there are are quite a few other applications to be hopped.
I like that couple of sentences. It clearly distinguishes creative activity from repetitive work.
Creative people enjoy what they are doing and in most cases dont care if or how much they are paid for it. Working for a living wage is something else… it can be quite investigative and logical or it can be boring and repetitive, but the one thing workers want is reliable tools that minimise the effort.
I think the major computer companies try to provide reliable tools, but they fall into the trap of freezing things. That is a trap because the more you cling to the past, the more you miss out on modern advances in technology. … we see things like IBM providing JCL in a VM inside another OS so people can run their ancient JCL procedures forever…or people wanting to freeze apps in a container environment so they are not exposed to updates. Those are traps.
The challenge is to provide consistently working tools in an evolving system. Microsoft try to do it by releasing a new distro every 2 years… yes each is a new distro… they just call them all windows to placate people.
Linux as a whole doesnt even try. Apart from the kernel team there is no centralised planning in Linux. Freedom reigns. Any distro can be as creative as it wishes. App makers are the same. It works because of the power of numbers… if a large number of developers try, a few will be successful. But it does not necessarily lead to consistent tools.
For the normal user with few computer skills it is a terrible choice… Microsoft support or self support. Neither is satisfactory. There is no computer system yet that ‘just works’ without support.
There is a ‘digital divide’ between people who cope and those who fall behind. It is widening, despite efforts to bring computers into education at an early age. In theory one could solve the Linux takeup problem by education, but in practice the educators fail us by opting for the easy corporate supported option… I know of schools where they throw out all the hardware each year, and buy the newest model … just to avoid doing updates.
We all want hassel-free, but that is ridiculous.
Or they have to support old and outdated hardware and software because BILLIONS of people use it and they’d be blasted by the press, by the users, and by the Linux community for abandoning their user base. I’m sure Microsoft and Intel would both love to drop some of the legacy aspects of their products. What happened when Intel tried to create a RISC chip? What happened when Microsoft required a TPM chip for Windows 11?
I do like the creativity of some of the Linux distros, but there are many times I wonder if there would be more real progress and more adoption if there was less “wasted” effort. It isn’t really wasted, but man have there been a lot of dead ends and missteps. Upstart, Mir, Unity, Ubuntu One, Ubuntu Touch, and those are just the first few that come to mind from just Canonical. As I recall, they were criticized for all of those ventures.
Were they a total waste of time? Not really.
Have they made leaps and bounds since abandoning them? Kind of.
Yes, it is like evolution in animal and plant populations. They explore every niche by diversifying their gene base. Many attempts to diversify die out, but some thrive. It is the genes that survive… individuals are mortal.
Same in computing… it is the knowledge or the code base that survives. Individual instances all eventually disappear.
Do we best support it by freezing or by effective modern substitutes?
They did / attempted it several times…
Always failed as a server platform… There are heaps of embedded controllers (e.g. RAID controllers) still using Intel RISC chipsets… I couldn’t name them all…
Intel’s IA64 was merged into HP’s PA-RISC (or vice versa) - which also incorporated a bunch of stuff that HP got when they bought out the IP for Alpha64 (off DEC / Compaq)… I’m pretty sure DEC sued Intel for some stuff they were trying to do in IA64 - and maybe even got a settlement - but HP shushed all of that…
Intel have tried manufacturing / rebranding ARM in the past…
What are we left with ?
- sparc - Oracle are doing their best efforts to kill it off (but Fujistu own huge chunks of the IP - I don’t know what they do with Sparc)
- mips - almost dead - but the Chinese licensed the arch and still going with things like Longsoon (probably embedded in stuff like Huawei)
- PowerPC - still going fairly strong - but mostly just IBM AIX and AS/400 (some gaming consoles used it like the PS/3)
- alpha - dead (it was awesome - may she RIP).
- arm - still kicking and massive - mostly down to the aweseome powersaving features - helpd by Apple channelling a heap of dosh into ARM when the Apple Newton came out (Apple still own a chunk of ARM) - everything Apple is now ARM (arm64) and pretty much 99% of Android is arm - due to how power savvy that ARM RISC arch is…
- RISC V - still kinda experimental - can’t wait till it’s more mainstream and I can get a board that will do mainstream Linux workloads
- 880x0 - another Motorola attempt at RISC - I mostly only ever did stuff on 88010 - Data General UNIX on AViiON - they were cool machines - not as cool as SGI on mips32 or mips64 - but DG-UX was a pretty awesome SysV UNIX… DG moved to Intel - but they worked with the likes of SGI to implement awesome stuff like NUMA - some of that NUMA stuff was pretty cool…
Some manufacturers made RISC laptops (besides Apple) one company was Tadpole, you could get a laptop running Solaris on Sparc, AIX on PowerPC…
That was way back when you could run the vendor UNIX, or Windows NT for RISC… Microsoft only ever ported the Windows 2000 codebase to Alpha when it was Windows NT 5 “beta”, but dropped it when Windows 2000 came out… I actually supported several DEC Alpha boxes running OpenVMS, Digital UNIX / Tru64, and Windows NT 4…
This thing looks pretty cool - 100% better than the shonky PineBook :
https://www.kickstarter.com/projects/argonforty/upton-one-the-true-raspberry-pi-compute-module-5-laptop
Dunno if it’s based on the Pi5 (full board) or the CM5 (compute module 5)… You can get a Pi5 with 16 GB of RAM - shame it couldn’t have more cores…
Exactly right.
linux can have a spectrum of distros ranging from rock solid to experimental … in that sense linux has solved the problem of business users wanting dependability… use RHEL if they must and pay for their support.
Intel and other’s RISC ventures didn’t pan out largely due to a cost issue. My point was it’s easy to be a Monday morning quarterback. Hopefully that phrase makes sense in other countries too. With the userbase Microsoft and Intel have they have obligation to make things work. Look at the hubbub about extending support for Windows 10. How many Linux distros with official support, RHEL and Ubuntu and Suse and others, support their software for as long as Microsoft has? A couple do for sure that I know of. How many of us would suggest end users or companies continue to use those 10-year-old versions of Linux? Not many. There are cases where it’s too painful or costly to move to something new. We’ve all been there. Yet Microsoft is the devil for forcing users to upgrade.
In an ideal world where everyone insisted on their software being portable the “too painful” syndrome would not exist.
It is the user’s responsibility to avoid lockins… choosing something because it is easy now, can lead to pain later.
It’s not the problem that Microsoft forces people to upgrade, it’s a problem that Microsoft forces people to throw away perfectly functional hardware, because it doesn’t have a particular unnecessary chip.
On top of that, every new version of Windows is, as far as I see it, a downgrade, as opposed to an upgrade.
I’ve had a few upgrades of Linux Mint on my current hardware; no problems or frustrating and unneeded changes at all.
I read somewhere - quite recently - about Microsoft’s workflow for Windows NT (i.e. pre alpha Windows NT 3.x) - the codebase was compiled on some semi-obscure Intel RISC platform - then ported to x86 (i.e. i386)…
Fascinating stuff - to me anyway…
Apple did a better job of porting their RISC code-base from PowerPC, to Intel CISC, then to ARM64 RISC… the microkernel is probably fairly easy… And that Darwin was a BSD that NeXt ran on Motorola 68000, then i486… When Apple got Darwin - they just had to port their microkernel to a new ARCH and build…
Microsoft are still flirting with RISC - Windows 8 could run on ARM and x86 (and x86_64)… And there’s new stuff they’re doing on ARM - I’ve seen reports of people getting Windows to run on non-sanctioned armhf and arm64/aarch64… And people getting aarch64 Linux to run on things like Exonys ARM CPUs…
All I want is “convergence” so I can run Linux (with a shell) on a RISC (i.e. arm64) CPU on a portable device - that is also my phone, but can power a full desktop experience - which I first encountered running MaruOS on a Nexus 5… but having to use the Android Linux kernel is rubbish - i.e. they don’t compile in NFS support! WHY? How much kernel space did they save by omitting NFS?
I dont get it? Debian and Ubuntu are supposed to support risc-v. What extra does a portable device need?.. or does this “support” claim not mean that it works?
You’re again confusing “RISC” with the open source hardware instruction set / CPU “RISC-V” - they’re not the same thing…
RISC = reduced instruction set computing… i.e. faster clock cycles, less instructions… it’s simple really - if you can omit a complex instruction - with a reduced one - for simplicity’s sake :
mutliplication
CISC will do the multiplication with dedicated instruction
RISC will do addtion on a loop till its finished the end value of the loop (or something like that)
(the less instructions on the CPU die - the more efficient it can be)
It might seem wasteful - but - RISC can often do things faster than CISC… I dunno - that’s how I learned about CISC VS RISC when doing Computer Science at Uni - and - reading lots of nerd papers in mags like Byte 
RISC = mips, sparc, powerpc, alpha (and more)
CISC = x86, motorola 68000
RISC-V = an opensource hardware initiative to develop an open RISC instruction set and CPU architecture …
And - the main issue with mainstream Linux supporting ANY non x86_64 is drivers…
Like I mentioned before - a whole bunch of SBC (single board computer) manufacturers leverage hardware/driver kernel support by using the Android kernel… or keeping the GPU driver internal and proprietary and only releasing support as BLOBS…
In nearly ALL ARM platforms, not dissimilar to AMD incorporating AMD graphics into the CPU die, - i.e. “Vega 8” (not to be confused with discrete GPU like AMD Radeon) - to use features - you need drivers… Some SOC (system on chip) licensee manufacturers of ARM chipsets won’t release their code…
Raspberry Pi foundation are open as possible as they can be… I’ve been playing with ARM SBCs for 10+ years - RPi is the best platform… it’s the most open…
But other licensees of ARM CPU IP aren’t - why? I guess they can make money off Samsung with non-disclosure agreements that help Samsung write better drivers?
I can understand why Apple are closemouthed on what they’re doing with arm64… But to be honest - they’re doing a vasty better job with ARM than any other vendor…
But some shitty little fabricator in some city in China keeping their tech a closed secret? Why? Did it help you? Meanwhile China is stealing IP left right and centre anyway… WHY CAN’T WE ALL JUST SHARE AND GET ALONG