Discussion Pastiche to Linux distros

[DK1994]

Linux has always been more for technology hobbyists who want to learn how software works inside and out, and then create new software afterwards (whether it'd be individual programs or a whole OS). While there's nothing wrong with people using it for that purpose, and I respect the people who do, the problem is it causes more challenges than necessary for people like me who want to use it for average computer usage. While I haven't gone back to Windows or Mac (and I'll only do so if I run them in Virtualbox), there are a number of people who have, because they're more stable in terms of standardization and hardware support. Some more experienced users will end up going back to Windows or Mac because they're concerned about where the ecosystem is headed, or because some members of the community tend to have a holier-than-thou kind of attitude. Because of this, I'd like to see up-and-coming development teams create a series of stand-alone operating systems that have a single kernel, software library, desktop environment, package manager, and configuration, all-while looking and feeling like a Linux distro, but not having any Linux code in them. That way, users can have something that's stable and potentially wide-spread, but not have to deal with data-mining, surveillance, back doors, or other issues that are common with Windows, Mac, iOS, Android, and Chrome.

While making them stand-alone would make them standard, the developers would still need to gain driver and hardware support, so I'd recommend them writing their own drivers for hardware companies (as it'd be no skin off of the company's nose).

As for marketing, I'd recommend advertising on billboards, benches, bus stops, in magazines and newspapers, or even showcasing them at community events so people have a better understanding of how they'll work.

After development teams get driver and hardware support, and enough users through advertising, that should begin to get some game developers interested in branching out to these operating systems. Maybe as a way to solidify the potential of their presence in the gaming scene, the developers of each OS could have their own gaming division (that is, if they manage to get enough money from people buying their OS).

Because these operating systems are better off being open-source (due to proprietary software causing the issues created by Microsoft and Apple (as mentioned above), and the free software community acting more like a pseudo-religion), please don't encourage programmers to create forks because that's exactly what's led to Linux becoming unbelievably fragmented, which in-turn has led to a lack of game, hardware, consumer, business, or 3rd party software support. Instead, encourage them to create their own software from scratch. If they create better software than what's available, users will gravitate towards that, and crappy software will gradually and forcibly be phased out. Its not impossible to create software from scratch, but forking is essentially a cop-out because it's easier to do. Despite what people will say about some forks have turned out great (such as LibreWolf for example), the developers could've made something very similar and more efficient than what was being used before, but from the ground up instead.

Those are just my thoughts on this. If anyone else would like to see this happen, lets start a brainstorming session.

Update: The more I thought about this, the more I realized that forking isn't bad per-se. While it is true that Linux distros getting forked has lead to the issues I mentioned it having above, as long as the fork doesn't become reliant on it's predessor, its fine. For example, if someone were to fork Linux Mint and removed everything that keeps it tethered to the Linux kernel, I have no problem with that. I apologize for my original comments regarding this.

 

[crashtech]

Writing a full-featured OS from scratch is a monumental undertaking.

 

[DK1994]

Writing a full-featured OS from scratch is a monumental undertaking.

While that's true, it's worth it in the long run. It was a challenge for Wozniak to create the OS for the first Apple computer (and that was back in the 70s), and despite numerous criticisms and controversies over the years, look at where Apple is now. Its still a huge challenge to create an OS from scratch today, but modern computers and programming software have helped allievate that stress.

 

[Fallen Kell]

You do realize the amount of work you are suggesting right? I don't really see the gain of attempting to remove the Linux kernel from a "linus OS" as a whole. The vast majority of the issues you speak of are not a problem on the major stable distributions of Linux (i.e. Red Hat Enterprise Linux, Scientific Linux, Oracle Linux, (up until this past month) Centos, or SUSE Linux). These large and supported distributions that work on long term stable releases with documented support life cycles and have tons of hardware and software support (including how to get Steam setup properly).

I agree with many of your comments/complaints about some of the distributions out there having little to no support for various hardware and all kinds of issues with dealing with various software. I'm no so sure about your whole issues with trying to remove the Linux kernel from the OS. I mean you basically want to redevelop or fork everything about the linux os and replace the kernel for unknown reasons (other than holier-than-thou of some other users). I have been using Linux as my primary OS for 20 years now. I do still have a windows system, mainly to support games that do not work properly under Linux, but the vast majority of everything else, I do under Linux. Currently I am using CentOS 7.x at home, but use RHEL at work. I am currently looking at what is happening with CentOS 8.x and future to see what I want to do, but the beauty of it is that I have a choice, unlike in a proprietary OS where I would just have to take it or leave it if they made a change I do not like, I can decide to switch over to one of the other distributions or even fork my own if I do not like the direction of the current distribution I am using goes (which I do not like with what IBM is forcing Red Hat to do to CentOS with making it effectively a beta/testing distribution and not directly compatible with the RHEL versions, but like I said, I have choices and am simply waiting to see what pans out before I make a decision).

 

[DK1994]

You do realize the amount of work you are suggesting right? I don't really see the gain of attempting to remove the Linux kernel from a "linus OS" as a whole. The vast majority of the issues you speak of are not a problem on the major stable distributions of Linux (i.e. Red Hat Enterprise Linux, Scientific Linux, Oracle Linux, (up until this past month) Centos, or SUSE Linux). These large and supported distributions that work on long term stable releases with documented support life cycles and have tons of hardware and software support (including how to get Steam setup properly).

I agree with many of your comments/complaints about some of the distributions out there having little to no support for various hardware and all kinds of issues with dealing with various software. I'm no so sure about your whole issues with trying to remove the Linux kernel from the OS. I mean you basically want to redevelop or fork everything about the linux os and replace the kernel for unknown reasons (other than holier-than-thou of some other users). I have been using Linux as my primary OS for 20 years now. I do still have a windows system, mainly to support games that do not work properly under Linux, but the vast majority of everything else, I do under Linux. Currently I am using CentOS 7.x at home, but use RHEL at work. I am currently looking at what is happening with CentOS 8.x and future to see what I want to do, but the beauty of it is that I have a choice, unlike in a proprietary OS where I would just have to take it or leave it if they made a change I do not like, I can decide to switch over to one of the other distributions or even fork my own if I do not like the direction of the current distribution I am using goes (which I do not like with what IBM is forcing Red Hat to do to CentOS with making it effectively a beta/testing distribution and not directly compatible with the RHEL versions, but like I said, I have choices and am simply waiting to see what pans out before I make a decision).

I'll admit I've mostly been using Mint, but I've heard RHEL is more for advanced users, and I don't have the experience in Linux to run that. Granted, I've been using Linux for about 10 years now, but I'm not a technical-oriented person, nor do I honestly care to be anything beyond a casual user. As long as a distro is easy to use, and doesn't have the backdoors, data-mining, surveillance, and proprietary software that's commonly found on Windows, Mac, iOS, Android, or Chrome, I'm happy with it.

I had mentioned that the numerous forks has led to fragmentation issues, which is why support for 3rd party software, hardware, and games hasn't been gaining traction as quickly as it should, or at all. That alone is one of the reasons why I'd like to have the kernel removed from each distro, and replaced with individual ones, so each distro can be it's own OS. To be clear, I don't mind individual programs being forked (such as LibreWolf being forked from FireFox), nor do I mind forks of a whole OS. What I do mind is various forks all running on one kernel.

I also mentioned the issue of Linux being made for nerds who are looking to become programmers. As long as each OS is made for the average computer user, but it still is open-source, then the differences between both types of users can be bridged. That's another reason I want to have the kernel from each distro removed.

A third and possibly final reason for wanting this to happen is wanting to have more professionalism when it comes to technical support. Again, I'm not saying every single person from the Linux community is a butthead, but there's enough of them who are which makes it a well-known problem. By establishing a tech support service for each OS, the people providing support can be trained on how to handle each problem in a professional manner, and can be fired if they don't present themselves that way.

As far as you saying you like choices, I've also been thinking of setting up a cooperative where each OS would be under a banner, rather than being powered by one kernel. That way people can still have choices without running into compatibility issues, or realizing that support is sparse for their OS.

 

[Fallen Kell]

The kernel isn't the reason for the incompatibility with games. The incompatibility comes from all the other software libraries that the distribution has, which need to match with what the games were compiled against. The biggest culprits would be the libc and libgcc, quickly followed by any of dozens of ancillary libraries for graphics, math, and others.

You can replace the kernel in most linux distributions with one that you compiled yourself. In fact that is done all the time, especially in the high performance computing world where removing features of the kernel that are there for compatibility and flexibility can result in performance gains.

As stated, the issues you are mostly complaining about are more about the particular distribution you have chosen to use, and not the OS as a whole. If you had asked these questions 6 months ago, I would have highly suggested trying CentOS due to it being a free version of RHEL. For your particular use case, you can now get a RHEL demo/developer's license and use RHEL (just realize you will not have full support like email/phone support for problems issues with this particular license and just have access to their online documentation and website portal).

The jury is still out as to what project is going to pickup the pieces of CentOS to fill the void that now exists. AlmaLinux has a beta release and seems to be the furthest along, but Rocky had the backing of one of the original founders of the CentOS project. So again, it is hard to say which is going to win out in the long run, especially with AlmaLinux getting a huge cash infusion of $1million a year from CloudLinux, but Rocky seeming to have some of the ideological backing from the project leaders.

Again, this is all a good thing, as it means no one entity can actually kill a product that is well liked, as IBM (who now own Red Hat), seems to have tried to do with the CentOS product (Red Hat had previously taken over CentOS, with CentOS initially being a free version of RHEL built from the source code that Red Hat had to release as part of the GNU GPL and stripped of Red Hat's branding).

Yes, it can be confusing, as this saga above shows. But the main message is that if you want something stable, you pick a distribution that is known for having well documented product lifecycles with well used beta distribution branches (Red Hat also funds the Fedora project distribution which they use to beta test new software, new versions, and new designs so that anything they add to RHEL and CentOS had months if not years of testing). If you want an OS that support most devices, you stick to ones that are most widely used across the industry. If you want an OS that supports gaming, you stick to ones that can support Steam. If you want an OS to tinker with and has the latest features/ideas, you go with a cutting edge distribution that has nightly/weekly releases (but recognize that it comes with new problems/issues nightly/weekly as well).

So, again, my suggestion for distributions to use goes back to my initial list I posted above for someone like you who just wants something that works. Mint is based on Ubuntu (like how CentOS is based on RHEL). Ubuntu is based/derived from Debian. As such, Mint and Ubuntu take their main fundamental queues from what happens in Debian. Debian is typically considered to be one of the most non-new-user friendly distributions (which is why Ubuntu was first made to try and make it more friendly, and why then Mint was made to try and make Ubuntu more user friendly...).

You seem to be of the I want it to be like Apple or Microsoft, where most things just work, and I don't need to dig into low level stuff to get my hardware supported and want to play games on it, kind of mentality. This screams ANYTHING BUT DEBIAN based distribution to me. If you want something right now to try, use CentOS 7.x (NOT 8.x which is where the changes are occurring, which is going to make CentOS a more experimental distribution on the cutting edge and thus with all the problems of using things that have not gone through months or years of testing first through the Fedora distribution that Red Hat has used to test out new things first for 1-2 years before incorporating into RHEL once it was stable and reliable).

 

[DK1994]

The kernel isn't the reason for the incompatibility with games. The incompatibility comes from all the other software libraries that the distribution has, which need to match with what the games were compiled against. The biggest culprits would be the libc and libgcc, quickly followed by any of dozens of ancillary libraries for graphics, math, and others.

If the incompatibility comes from the various software libraries each distro has, then there should be just one software library for each distro. Lets not forget how there's numerous distros running at once, and each are different from each other, which makes it even more complicated mess, because nothing is standard. This is why developing software for linux, and making them compatible with each distro, is a nightmare, and why even Lennart Poettering, who was one of the main developers for Pulse Audio, said "it is very difficult for programmers to know which audio API to use for which purpose." Making each distro it's own OS, and then making Linux a cooperative, would've made everything standard and work properly. Because that didn't happen, now Linux might as well be the OS equivalent of Frankenstein's Monster.

As stated, the issues you are mostly complaining about are more about the particular distribution you have chosen to use, and not the OS as a whole.

That's putting words in my mouth, because I never mentioned any of the problems I had with Mint in particular. Yes, I've used Mint more than any other distro, but I've also used LUbuntu at one point, I've tried using Debian, and I ran some other distro I can't remember the name of. I could look into running Cent, but it would just reinforce how each distro is different from each other and shouldn't be running on the same OS.

Yes, it can be confusing, as this saga above shows. But the main message is that if you want something stable, you pick a distribution that is known for having well documented product lifecycles with well used beta distribution branches (Red Hat also funds the Fedora project distribution which they use to beta test new software, new versions, and new designs so that anything they add to RHEL and CentOS had months if not years of testing). If you want an OS that support most devices, you stick to ones that are most widely used across the industry. If you want an OS that supports gaming, you stick to ones that can support Steam. If you want an OS to tinker with and has the latest features/ideas, you go with a cutting edge distribution that has nightly/weekly releases (but recognize that it comes with new problems/issues nightly/weekly as well).

I should be able to use an OS that does everything, instead of jumping through hoops to find out which one does what

You seem to be of the I want it to be like Apple or Microsoft, where most things just work, and I don't need to dig into low level stuff to get my hardware supported and want to play games on it, kind of mentality.

That's also putting words in my mouth, because I said in the OP that I want each pastiche to be open-source so it doesn't end up like the next Windows, Mac, Android, iOS, or Chrome (granted Android and Chrome are both open-source as well, but they're still developed by Google, which is a Big Tech company). While I did say I wanted them to be made for the average user, that doesn't mean I want it to be like Windows, Mac, Android, iOS, or Chrome.

I should also mention the lack of market share (aside from Android and Chrome) also holds Linux back. Some people in the Linux community even admitted they see each distro as a niche, and they're proud of them being that. One guy on Linux.org even claimed "If Linux ever became mainstream, it will be another polluted Microsoft product." This explains why some of those people can be buttheads to newcomers. Because of that, most distros never go anywhere in terms of market share, and most-likely never will. Instead, they'll just exist as a small user base.

 

[crashtech]

Hi OP, I wonder how much of your own time and money you'd be willing to expend on such a venture?

 

[DK1994]

Hi OP, I wonder how much of your own time and money you'd be willing to expend on such a venture?

I'll admit I haven't looked into either of those just yet, but to be realistic with an estimate, I imagine the cost to fork an OS would be about $150,000-300,000 while building an OS from scratch would be about $700,000-1,000,000. The time needed to fork an OS would probably take 2-5 years, while building an OS from scratch would probably take 7-10 years. You can correct me on those figures if I'm wrong (but my gut tells me I'm not too far off). I'm aware those figures are huge, but even though it will take me a while to get the money to do this, I am working on getting it. I'm an artist who's currently working on drawing videos for various platforms, and I'm also in the process of getting freelancing gigs as a secondary job. On top of that, I plan to turn my drawings into various pieces of merchandise and sell them on screen-printing sites, such as Zazzle. By busting my butt long and hard enough, I should hopefully have the money by 2026.

Obviously doing this will take an entire team to complete, and there will be many unforeseen challenges that will come up relatively consistently. Being that this will be my first time helping to develop an OS, it will definitely be an adventure.

 

[mv2devnull]

I'll admit I haven't looked into either of those just yet, but to be realistic with an estimate, I imagine the cost to fork an OS would be about $150,000-300,000 while building an OS from scratch would be about $700,000-1,000,000.

CloudLinux did announce that they will invest $1,000,000 per year into AlmaLinux.

CentOS/Alma/Rocky are not "from scratch". Not even a clear "fork", but just rebuild from available sources. (There are some compulsory edits to remove trademarked content.) There were comments on CentOS mailing lists that a million per year is not enough.

AlmaLinux has a "Stable Release" already. Rocky Linux had to delay their RC; now at end of this month.

 

[DK1994]

CloudLinux did announce that they will invest $1,000,000 per year into AlmaLinux.

CentOS/Alma/Rocky are not "from scratch". Not even a clear "fork", but just rebuild from available sources. (There are some compulsory edits to remove trademarked content.) There were comments on CentOS mailing lists that a million per year is not enough.

AlmaLinux has a "Stable Release" already. Rocky Linux had to delay their RC; now at end of this month.

I'll admit that's an astronomical amount of money to think about, which is why I was just thinking about charging for releases, and marketing the OS through various advertisements. By popularizing it through ads, I should be able to develop a following for it, and then get back what I put into it. To be honest, I'll use Mint until I find a BSD I'm comfortable with, and after that happens, I'll just leave Linux behind. That doesn't mean I'm no longer interested in building an OS, because I thought about combining Mint with the BSD kernel so it can be it's own operating system.

Update: After looking up reviews for GhostBSD on DistroWatch, some people have said it's very much like if Mint was running on the BSD kernel, so me making an OS like that would be redundant. However, being that OSX was formed from FreeBSD, I could make an open-source alternative to it, market that, and potentially level the playing field with Apple (and yes I know that sounds ballsy to say).

Secondary update: After reading this thread over, and thinking about how many of the things I said were inconsistent, now I realize I came off as unprofessional in my approach, and decided to deactivate my account as a result. Thank you to everyone who participated in this thread, even though you may no longer think very highly of me.