The Future direction of Linux

[MurphCID]

First thing: Obey the rules. No name calling, no snark, no rules violations please. The other topic was very informative. I really do not want this one shut down.

So, where do we go in Linux? Are we looking at a Containerized future where distributions just send the basics, and everything else is a flatpak or a snap? I see Linux to some small extent following the direction of Microsoft Windows and Apple Mac OS, but I also see them taking things that Linux has done for years and adopting that. An example is Window tiling.

Are we also going to see a consolidation into fewer distributions, as it becomes harder and harder to get people to give up their time to work on them? And are we going to see distributions such as Mint forced into a drastic shift if Ubuntu goes 100% snap, and immutable? Will the devs be able to keep things the way they are?

Also for desktop Linux will we see fewer desktop environments based on corporate sponsorship of Gnome, and just a few others?

[Knightstmplr1]

Have a read of Clems newsletter today, he and his team are actually addressing these issues by keeping a close eye on what Ubuntu is doing with its focus on containerisation with Snaps etc. and how they have mitigated things Ubuntu has done in the past and how they have an eye on mitigations for future versions of LM as well.

https://blog.linuxmint.com/?p=4554

A quote from Clem regarding this:

"Note: We’ve got a vocal minority of LMDE users. As usual we’ll come up with a great release. I appreciate the fact that they love what we do. I ask them to please remain civil when it comes to criticizing Ubuntu and to understand that we do what’s best for Linux Mint as a whole, both when we work on LMDE, when we work on Linux Mint and when we form long term strategies. The news sometimes gets blown out of proportion and people can get really passionate over very little. If you look at the past you can appreciate how calm we are as a team and how serene we are with our development. We’re rarely affected by upstream decisions. When we are, or when we might be, we’re able to invest, to mitigate them and get to where we want to be. That’s how we’ve got something like LMDE already, whether we ever need it or not. Don’t panic, don’t lobby for rushed decisions based on fears or passion, we know who we are and we know what we’re doing."

[rossdv8]

Until fairly recently I always had a couple of roots on my hard drives, so I could have my current version of Mint and also, a place where I could install the newest latest Mint beta (haven't been into the Mint alpha loop for a long time now) and mess around with it.

I stopped doing that once Mint had got so stable that upgrading to the newest - even via 'unofficial channels' and doing stuff like installing KDE into it, didn't prang it.

However I've been wondering just what installing LMDE into a spare root would do, and if sharing /home between the two might interfere with some config stuff.

It might be something to experiment with, because the more I read about LMDE, the more I like it - except that I dislike Cinnamon (yeah, I know it is Clem's baby - and it has come a long way. I know some people have played with KDE on LMDE, but I'm still not convinced it will work as seamlessly there as it does on Mint Xfce.

Still, I can see a lot of advantages to flatpak vs Snap, and I had already moved from repository to Appimage before Mint really hit the flatpaks thing hard, so I like the idea of a flatpak-centric Mint, and Linux generally. With big hard drives now relatively cheap, having huge roots is a lot more sensible now than it used to be, and we can cope with flatpaks living there easily.
Which is one thing I dislike about Appimages, They normally live in /home, so are not as isolated, even through they are protected at runtime.
The other advantage of flatpaks, and something I think Mint Update Manager does very well, is that Flatties are updated automatically in Mint, while at present, not all Appimages work like that.
So while Appimages 'might' get the latest App updates before they hit FlatHub (it doesn't always apply) you still have to remember to update most of them manually.

Since we switched our business computers over completely to running Red Hat Linux exclusively in 1998, and began supplying computers running Red Hat, as well as providing customer support for them from 1999, I've seen a hell of a lot of changes to Linux and the various distributions.

When something as stable and good looking and easy to get used to as Mint and the Cinnamon desktop comes along, it makes a transition to Linux a whole lot easier for people looking for a very comfortable computer experience.

As much as I also dislike Android and Mac for personal reasons, I do think Apple and Google OS based phones tablets and Chromebooks are beginning to make people realise that 'Linux really is a Thing', and is just as pleasant to use as their phone and tablet.

[billyswong]

If Ubuntu really go full-Snap, this implies PPA is being abandoned. If there is really no more PPA, shifting everyone in LInux Mint to base on Debian directly is not a high cost decision. Proprietary firmware is also available in Debian now. There is procedure for one to run backport kernel in Debian too and I am sure Clem will make that GUI friendly when we no longer have Ubuntu based Linux Mint.

But on the other hand, Snap is not compatible with low level software as far as I know. Ubuntu can't create Snap-version of Kisak Mesa repository. So I guess Ubuntu can't really go "100% snap"?

[t42]

It's not a big difference between full-snap or half-snap, snap is just a piece of technology applied to the practical aim. Problem Ubuntu is facing is the same as other distributions, an enormous package base. Ubuntu and Debian repositories are mostly identical. Today Debian has 1171 unmaintained packages from which 1083 are orphaned from one year to more than 10 years. About 5000 packages are in work-needing and prospective packages family. Greater part of Ubuntu packages (over 23000 packages) is in Universe repository defined by Ubuntu in such a way

Canonical does not provide a guarantee of regular security updates for software in the universe component, but will provide these where they are made available by the community. Users should understand the risk inherent in using these packages. Popular or well supported pieces of software will move from universe into main if they are backed by maintainers willing to meet the standards set by the Ubuntu team.

So snap may be a way to sneak out of the package inheritance problem and present Ubuntu Desktop as a (seemingly) complete product.

[spamegg]

Are we looking at a Containerized future where distributions just send the basics, and everything else is a flatpak or a snap?

No, containerization will be mostly for servers / a few specialized distros. Most people still cannot get over the massive size of the initial Flatpak install for one app.

Are we also going to see a consolidation into fewer distributions, as it becomes harder and harder to get people to give up their time to work on them?

No, I think we might see MORE distros. It's becoming much easier to create your own distro or tinker with an existing one these days. (Take a look at DistroTube Youtube channel to see how easy it is.) There are also distros coming back from the dead, like Solus. But we will also see lots of dying distros, and lots of new distros replacing them. So it will be like the last 30 years of Linux but larger. Fierce open source competition, innovation, survival of the fittest, lots of deaths.

And are we going to see distributions such as Mint forced into a drastic shift if Ubuntu goes 100% snap, and immutable? Will the devs be able to keep things the way they are?

No, and yes. I don't think distros like Mint will be affected much. First of all, Ubuntu cannot exist without Debian, and Debian is still old-fashioned package-oriented. That's not going to change. Those distros can fall back on Debian fairly easily. Second, it's all about the people behind these distros. For Mint to change, Clem and the Mint team's personalities would have to change drastically overnight like a schizophrenic person. Third, no distro can exist without the good old packages. Ubuntu or anyone else with any amount of money and resources could not possibly rewrite the entire Debian base and Linux kernel into snaps. (I guess they could decide to go the Red Hat route... but that won't work for them in the long term.)

Also for desktop Linux will we see fewer desktop environments based on corporate sponsorship of Gnome, and just a few others?

We will see MORE desktop environments. Just look at Pop!_OS and their Cosmic, or Elementary OS and Pantheon, or Deepin for example. There are also non-company based new DEs being developed, like Cutefish. People won't stop creating new things. It is born out of a desire to make your own thing. Talented people who work in open source tend to be highly individualistic and disagreeable in their personality traits.

Just my two cents.

[rossdv8]

I would agree with the above remarks and emphasize that developing more distributions, which was 'necessary' in the first 5 or ten years of Linux ( I was actively involved with several including Mepis, PCLOS and some earlier ones), but is more just a 'because we can' situation now, will be even easier due to self contained packages like Snap, flatpak and Appimage being available, because someone who just wants to make their own 'Desktop' available, no matter what it is based on, doesn't have to maintain a repository or a PPA.

Even people like me that take a perfectly good and serviceable distro like Mint, then tweak it and swap in a completely 'wrong' Desktop / Window Manager combo, can be quite at home and find all the Apps we want in a flatpak or an Appimage or a combination - applied over the basic Distribution.

It has become that easy!

[t42]

Yes, it become easier to create a distribution, but what type of distribution is discussed here?. At the beginnings of Debian it was declared

Debian was meant to be carefully and conscientiously put together, and to be maintained and supported with similar care.

At it is exactly that up to this day. There is no "maintained and supported with care" part among almost all new distributions, the creative part is far more interesting than months and months of struggle with infrastructure and support. Most important factor for user while choosing a distribution is not its features but its infrastructure and maintenance policy.

[Pjotr]

Most important factor for user while choosing a distribution is not its features but its infrastructure and maintenance policy.

Spot on. Which is why I love the Ubuntu LTS codebase of Mint.

[MurphCID]

Excellent point Spamegg and T42, with both money and maintainers being, not lost, but not as active, I see fewer packages in both .deb and .rpm being actively maintained, which further pushes things to containerized things like Snap and Flatpaks. Your point on the number of unmaintained, abandoned, or zombie packages is well taken, and I fear the shape of things to come as distros look to focus on core packages. Look at Fedora not making an .rpm of Libreoffice, but instead using Flatpak to get the Libreoffice package. Now I agree that there are some packages that just cannot be made a snap or flatpak and will have to be kept as a .deb/.rpm package.

T42 has another very excellent point, that as resources become slimmer, the Debians, Red Hats, Slackware, and Arch are going to possibly necessarily become even more, not dominant, but alone I suspect. Sure for those with the skills to do so, you can create a distro, but can you keep it going? As the resource pie gets smaller, I see more and more developers (I call them programmers), choosing to spend their time on distributions which show the long term chance of survival. Plus I see competition from Microsoft Linux, Amazon Linux, Oracle Linux, etc moving down towards the desktop from the Server realm.

So the question becomes what is Linux going to look like in 10 years? I don't think anyone really knows. Same thing with Desktop Environments, I see Gnome, KDE, Cinnamon, XFCE, and possibly Mate being around. Others, I am not sure about such as Budgie, and the other DEs.

BTW I am really enjoying this discussion!

[rossdv8]

For some of us, I guess a fair bit of 'loyalty' to a distribution thing, once the initial excitement of actually discovering Linux had work off, was as much about the Desktop Environment and the Apps that each of the early distros' developers were building.
Even now, taking into account the huge differences between KDE/Plasma and Gnome, or Gnome and Cinnamon, and all the work that goes into each of those three what they are (and of course Budgie and all the others) we are seeing a lot of new stuff that sometimes almost goes un-noticed.

Xfce4.18 for example brought back Panel Profiles. Most users will hardly even notice it is there. Regular users will know they can 'usually' fix a failed Panel with a command like:

Code: Select all

rm -rf ~/.cache/sessions && xfce4-panel -r && xfwm4 --replace

But now Xfce users have a simple tool that lets them create a backup of their Panel, and restore it if the Panel fails to load. Not that this happens t most people, but for some who like to 'mess with stuff', it does on occasion. There are 'I've lost my Panel' threads on the forum.

That tool by the way has some marvelous other potential uses that will probably also go un-noticed.

So, each Desktop Environment has its special features that make it appeal to certain users.
KDE/Plasma Desktop has had handy 'Widgets' or 'Plasmoids' since almost forever.
Cinnamon Desktop features something similar called 'Spices' to make life more interesting.

I'm not sure that we 'need' more Distributions, especially as many of these, like Mint, are piggybacked on top of already established Distributions - but Mint is a case in point where someone with what turned out to be a very good idea / dream, produced a downstream Distro based on an established Distro, and gave us a different take on a set of three established Desktop Environments.
Then, with the help of a team of enthusiastic people, created a new Desktop that is now being used by other Distros.

While I was writing this I was trying to remember a distro I used years ago that I was fascinated by as an example. It just drifted into my mind. KNOPPIX ! Knoppix was one of the Pioneering Linix Distros that introduced the concept or running a complete distribution, including a full set of Apps, from a CD (later DVD, then bootable ISO) that gave rise to what we download to run and install today.

I found Knoppix as a live CD on a Computer magazine cover in 2000. At the time I was still selling and supporting Red Hat with some of our IBM systms, and here was a CD-Rom based Linux that I could just plug into any computer and run. It found most hardware, although it didn't find most printers.
I can't recall which version did this, but the thing could find Windows computers across a LAN, so someone using Linux could easily print to a computer on a Windows machine - in the days when Linux drivers were rare.

So some Linux users like Klaus, Clem and others have always been the type of people who see the future, and open a door for the rest of us to follow.
Some of the people who step through the doorway, will open other doors . .

I don;t think Linux is likely to stagnate any time soon, and while it sometimes seems that there are too many distros, the curiosity of Linux users as a whole means there will always be a few people trying the 'unusual' distros. The big problem as always, is that for many a Distro is a labour of love, and provides very little in financial reward to cover cost of maintenance.
Which is where things like flatpaks can ease the pain!

[spamegg]

As the resource pie gets smaller, I see more and more developers (I call them programmers), choosing to spend their time on distributions which show the long term chance of survival.

I generally don't think this is the case... even the most "commercial" or "successful" (or whatever other word can be used) desktop environments have mostly hobbyist, unpaid and even anonymous contributors. I think Nick had a video on GNOME about this.

Also the tendency to fork and create your own thing is quite strong in programmers. I've contributed to some open source projects for a few years and I understand it. You get dissatisfied with the direction, you try to argue it "from within" but you can't influence it enough... eventually you get fed up and decide to go your own way. For a skilled programmer (with a day job) it's not too difficult to keep your distro going. There have been 1-2 man distros going for years, on near zero budgets. Progress is slow but OK (time is an "infinite resource" because you're gonna keep living anyway).

Getting it polished, popular and financially successful is another matter of course... But then again, how was a team as small as the Mint team with extremely limited resources able to make a better desktop for many years than Canonical with all its resources? Patience, good decision making, not making big radical changes, listening to your users... whereas corporations have a weird tendency to attack their own users, damage their own brands and break existing perfectly fine stuff.

I believe that the personality traits of the people involved is much more of a factor than resources. To make an analogy, Linux creators / users are like the Libertarian party in the US. Very small part of the population, still disagreeing with each other on many things, and highly individualistic.

Plus I see competition from Microsoft Linux, Amazon Linux, Oracle Linux, etc moving down towards the desktop from the Server realm.

Probably not. I don't see Microsoft creating a desktop competitor to Windows on purpose. They already added WSL2 so that's good enough for them. Oracle Linux existed for a long time and I don't think they had any interest in normal desktop usage, I don't see that changing. Not sure about Amazon, they like to get into markets they don't know much about and cause big screw-ups and money losses (like streaming TV shows such as "Rings of Power"). But the same goes for Google (who has way more expertise in software / OS than Amazon), and they made Chrome OS which runs on their specific hardware and not made for the "general desktop". So Amazon could do something like that maybe.

In any case, I don't think many users would use such desktops anyway... I mean, just look at the animosity towards Ubuntu, or anything corporate (those who are OK with / like corporate desktops would stick with Mac / Windows). Add to this the long list of historical "commercial Linux desktop" failures in the past, especially from the late 90s / early 2000s.

I could be wrong though! As you said it's very unpredictable.

[rossdv8]

The big thing for me, with Desktop Linux has been a combination of its stablity, and its incredible flexibility. I know Windows rarely 'breaks' these days, but like many of the other dinosaurs on here I spent far too long living with the BSOD, on my own systems and those of a few dozen customers at any given time. As soon as I could get the best and most progressive thinking mob onto Linux for at least the servers, and then gradually onto Linux and Open Office on the Desktop, I did. Until today, the last time U used Windows was on a new machine in 2007, just long enough to see if it looked interesting, then install Linux on it. I think that was about Mint Cassandra era. Maybe.

Today I had to set up a second hand Microsoft Surface Pro 3 and while it boots Mint and Pop OS, it can't really be used with them, because it won;t recognise ANY of my collection of Bluetooth keyboards on Linux, although it manages to connect to several of then concurrently and allow all of them to work at once. It will also find my Wireless keyboard in Windows as long as i have my donger in the USB port, BUT, then I can't use that and a USB stick at the same time to see if Linux will fond the Wireless keyboard. Catch 29!

So Linux still has a bit of catching up to do, even now.

The thing I like most about Linux on the Desktop though is that I can have just about any desktop I want, just by changing a few themes. I can have a very retro looking Windows desktop. I can select from almost any Mac themed Desktop style that I like. I can even choose from a huge range of Lunux styled Desktop layouts!

And with the exception of Cinnamon, most of them can be mixed and matched.
So by simply Logging Out, selecting between Gnome, Xfce, Plasma, Lxqt, Lxde and a couple of others, then Logging back In, the whole shebang can be made to look and work completely differently in moments. Without having to dual boot or run virtual machines.
I haven't tried messing with the Budgie Desktop on Mint and there are others that may be a bit too weird - but as I mentioned earlier in this thread, I love Plasma, but I also love Mint, and I like Compiz as a Window Manager.
Linux lets me blend all those things together to get what is for me, The perfect Desktop Experience.

For me, I think the Desktop component of Linux should continue to grow its user base just because a user can have either something very familiar to what they just came from, or something outrageously different from anything they have ever seen - and they still have the same stability, security and all the same Apps and tools in the same places.

[MurphCID]

I agree with the points above, but I have a small quibble. I just got a notice from System 76 that the issue with my system was that the Steam app was the one causing the issue. Now in this day and age of Linux, an update, should not bork a system, nor should it become an "unknown" where no one knows what is going on. Since the Steam issue apparently was pretty universal across Arch, Debian, Fedora based systems, it seems like the Steam devs released something that borked both the native apps and the flatpaks. I cannot address snaps since I do not use them.

I realize that unlike larger corporations that proper beta testing is sometimes not as complete as it should be, still in 2023 this sort of thing should never be allowed to happen.

I, for one, applaud Clem and the other developers for taking a slow, cautious, and nuanced approach to Wayland so that Mint will retain its essential stability. Since I do not have that much knowledge of the various parts of the video system between X11 and Wayland, I am sure that at some point things will be switched, and probably for the better.

[t42]

it seems like the Steam devs released something that borked both the native apps and the flatpaks. ...

I realize that unlike larger corporations that proper beta testing is sometimes not as complete as it should be, still in 2023 this sort of thing should never be allowed to happen.

You are right but better say all of it personally to them:

Code: Select all

Valve Corporation
Total equity	US$10 billion (2019)
Owner	Gabe Newell
Number of employees	~360 (2016)

[MurphCID]

it seems like the Steam devs released something that borked both the native apps and the flatpaks. ...

I realize that unlike larger corporations that proper beta testing is sometimes not as complete as it should be, still in 2023 this sort of thing should never be allowed to happen.

You are right but better say all of it personally to them:

Code: Select all

Valve Corporation
Total equity	US$10 billion (2019)
Owner	Gabe Newell
Number of employees	~360 (2016)

I am not sure if I completely follow your comment. But if they only have 360 employees of those you probably only have 60 actually working on code. Am I following correctly? I still do not know nor do I have the tech skill to find out what borked the software. Am I on the right track?

[rheinfels]

The future?

I think the future will be Linux and Linux Mint will establish itself as the class leader through its simplicity and out of the box functionality if everything continues to run so well. I think this is largely due to Cinnamon, which will also establish itself in the Linux landscape.
Whereas I think Linux Mint Debian Edition or another edition (maybe based on Arch, LMAE) will prevail.

Microsoft (Windows) is just killing itself with its decisions and people are realising this more and more only that this time really good distributions like Linux Mint are waiting for them to pick up them up. And they stay.

Ubuntu is becoming increasingly unattractive because it is cozying up to Microsoft and also putting out feelers towards Google. One can assume that Canonical's economic interests are increasing.

If Steam manages to establish gaming under Linux with Steam Proton, gaming will no longer be an argument for staying on Windows. And software companies will adapt to where the people are going. Similar to the smartphone market, many will switch to Apple's MacOS and the others to Linux. Just like Android (Linux) and iOS.

All this is already happening.
With more users, there will also be more developers for Linux.
I think flatpak will also become established.

But I also think that YouTube plays a big role here. Many Youtubers are increasingly pointing out Linux and its advantages. I don't remember anything like this back in the Ubuntu days, when I first wanted to switch to Linux (then Ubuntu). Linux is on everyone's lips. also in the forums and usual Windows forums.

But we don't know all that. Microsoft itself is watching the market closely and is one of the biggest sponsors.

Perhaps Microsoft recognises the potential of Linux Mint /Cinnamon and buys Clem and his team in as the main developer for its new operating system based on Linux. Windows (Linux Mint Edition).


It remains exciting.

[MurphCID]

This is such a great thread. Thanks all. I DO see more Windows refugees coming over as Microsoft adds more telemetry and spy/adware to Windows. And if rumors of Windows 12 being a completely cloud based O/S where you only have a client on your system, we could get more people trying Linux. I think that Mint is on the right track at this point, and we shall see where things go. I am interested in seeing if System 76 can manage to get their Cosmic desktop off the ground at some point. I know they are working on it very hard, but again it is a situation of limited resources, and limited budget.

[rheinfels]

I DO see more Windows refugees coming over as Microsoft adds more telemetry and spy/adware to Windows. And if rumors of Windows 12 being a completely cloud based O/S where you only have a client on your system, we could get more people trying Linux.

That will actually kill Windows. or not?
I mean they do it anyway with the "aggravations" on windows 10 updates and the release of windows 11 and the prevailing constraints there, such as an online account. One has somehow the feeling that Microsoft also actually has no more desire or why has one not all too long ago said Windows 10 should be the last project? *irony off*

I think these cloud base only things will never really catch on or atleast not so fast but I might be wrong. Unless everything is shipped to the internet in the form of a complete virtual reality environment.
That can work for individual services like music or e-mails. But for a complete operating system with everything on it? You can't get more transparent than that. But on the other hand we are already mostly connected 24/7 to the Internet anyways. Whether it's the pc, the tablet or the smartphone.

I think this cloud topic deserves its own thread because it is very interesting.

[MurphCID]

What concerns me somewhat is that when Microsoft does something, it just seems like in a few years many other things like Apple and Linux seem to follow suit, even it they do it "better" than Microsoft.

I can see them forcing people to buy new hardware just like they did with Windows 11. Personally I think it was silly to put those system requirements in place, but they obviously had a reason. With few distros still supporting 32 bit computers, I think the days of the 32 bit systems are numbered. Also I still think that most distributions of Linux will go the flatpak and snap route for packaging. I am very glad that Clem and company are looking at where Ubuntu will be with 24.04.

Edit: Also hardware support, AMD seems to be an ally of Linux at this time, and for the most part Intel is, but I am also concerned where Linux will go when hardware makers add the Microsoft Pluton chipset, or other TPM like requirements which almost lock Linux out of easy installation (at least for newbies) unless you know enough to get into the BIOS, and even then are they going to start hiding the BIOS settings or locking them out?