The GPL self-destruct mechanism that is killing Linux (CD, laptops, desktops)

[rcsteiner]

Originally Posted by plwhit

Interesting read, hopefully the people here who treat the OS as a religion will read the entire article before starting their usual speil about Linux....

I have one of Professor Tannenbaum's books. He and Linus have long held different opinions about OS design, but sometimes a basic and pragmatic approach to software design actually works better than the designs created in ivory towers. It happens more than you might think, I suspect.

One of the reasons Linux is so popular is the fact that it works, and it works well. The same could be said for the more cathedral-driven FreeBSD and friends at one time, though it's been long enough since I've bothered to look at those that I have no idea how they fare currently. I assume they still survive.

UNIX is not graphical, though many of its shells are these days. People tend to confuse the OS with the user interface, but for many operating systems other than Windows and classic MacOS there is no real dependency on the GUI ... it exists out of convenience, even in an administration context, but not out of necessity. I work on Solaris boxes for a living, and there is no GUI at all on the servers we run.

An interesting article, to be sure, but the arguments made are far older than the 20-year-old debate that is cited near the bottom, and I still find it interesting that monolithic and hybrid kernels still seem to rule most of the world. Not sure about kernels like QNX which claim to be microkernels, since many so-called "microkernel" platforms are actually hybrids if you go by formal definitions.

[rcsteiner]

Originally Posted by TurcoLoco

I hate to say it but every 10 years or so, they need to abandon whatever they have and create a newer, better OS from ground up instead up patching up an existing far from ideal dino OS.
There is no room for sentimentality when dealing technology.

Speaking from an enterprise computing perspective, that isn't correct. The desire for consistency has little to do with sentimentality, and has a lot more to do with a uniform experience over time, leveraging existing developer skillsets, and the sheer expense of porting complex in-house applications from one platform to another.

One of the big reasons why we still use operating systems like zOS, OS2200, MCP, and others on Big Iron is the fact that the exposed APIs and ABIs are consistent. In the case of the ones I cite, this has been true for 50 years or more.

I agree that the kernel and other elements of the platform can (and should) change with the times, but if an operating system is well designed to start with, massive changes should not be required at the core API or ABI level. Exceptions are made for new hardware platforms, obviously. Smart phones have a very different set of requirements for large-scale servers.

[rcsteiner]

Originally Posted by plwhit

Tell me something, if the corporations didn't have anything to do with computers just where would society be right now?

Most of the early development in networking and computing was driven by the federal government, not by business interests. It's hard to say where we'd be now ... business interests have certain made great headway in some computing contexts, but in others they've held us back for years or decades.

[rcsteiner]

Originally Posted by tvdxer

Linux did rise, not on the desktop, but on servers, workstations, embedded applications, and lately, Android. In fact, if you consider Android Linux, it may soon be the most used operating system in the world. In addition, the open-source philosophy, which Linux has bolstered immensely, spurred such innovative projects as Firefox, OpenOffice, and even Wikipedia.

Much of OpenOffice (and LibreOffice) can be attributed to StarDivision, who actually created the software in a cross-platform manner in the first place, and Sun, who had the foresight to purchase StarOffice from StarDivision and release the code base as open source.

Wikipedia is a classic example of what can be good, and bad, with "open source" projects. For the most part it is quite accurate, and the breadth of its coverage is remarkable, but its open nature does lend itself to edit wars and other similar things. The Linux kernel seems to avoid much of that because it has a much stronger hierarchical structure.

Linux already is the #1 server platform on the web and has been for some time, and the LAMP stack (Linux, Apache, MySQL, and Perl) also has a very strong presence.

[rcsteiner]

Originally Posted by swagger

The linux community can't even decide on a single graphical interface...

API/ABI fragmentation is a problem for commercial developers. They need to have something consistent write to, and that consistent thing much have a large enough market share for the development expenses to be worthwhile making. It comes down to simple arithmetic, really.

It's a real problem for Linux, and it always has been. I said that very thing 20 years ago, and I still agree with that assessment.

Unfortunately, given Microsoft's unethical and illegal behavior in the past (and I say it straight out instead of "allegedly" because they have a long track record of convictions throughout the world, not just in the US), the commercial desktop OS market is seen by most as a nonstarter. When the incumbent can break the law with impunity to maintain their position, why would any sane company want to play that game? Microsoft has the peripheral, PC preload, and developer markets locked up. There's no profit in trying to break that monopoly. Apple is a bit of an exception because they don't complete in the same hardware space, and they also have a fan base that stretches back in some cases for almost 30 years (first Mac release was 1984).

Linux did rise, not on the desktop, but on servers, workstations, embedded applications, and lately, Android. In fact, if you consider Android Linux, it may soon be the most used operating system in the world.

Android is an example of exactly how OSS should work.

The fantasy of a bunch of geeks working for free and sharing with everyone is nice and all, but the result is a bunch of half-baked abandonware and forked projects. Linux has been successful on the server largely because companies like IBM and Red Hat have put in the hard work to make it happen. Android has succeeded because Google has backed it and steered the project forward.

But Linux is all about voluntarily participating, for the betterment of everyone.

No it isn't. It may have been in the early days, but that's not even close to true now. It's a project that is backed by huge commercial interests that see a way to make money from it.

[TurcoLoco]

Originally Posted by rcsteiner

Speaking from an enterprise computing perspective, that isn't correct. The desire for consistency has little to do with sentimentality, and has a lot more to do with a uniform experience over time, leveraging existing developer skillsets, and the sheer expense of porting complex in-house applications from one platform to another.

One of the big reasons why we still use operating systems like zOS, OS2200, MCP, and others on Big Iron is the fact that the exposed APIs and ABIs are consistent. In the case of the ones I cite, this has been true for 50 years or more.

I was merely speaking from a Desktop computing perspective but on the enterprise side, what you stated holds true, naturally. So, no argument here.

I agree that the kernel and other elements of the platform can (and should) change with the times, but if an operating system is well designed to start with, massive changes should not be required at the core API or ABI level.

Has there been such a well made desktop oriented OS? I mean even a commercial OS like Windows almost always has major security holes and other software defects out of the box. It takes like 2 service packs to make it run somewhat decent and secure. I don't know any OS that is codded that well at the core API or ABI level so they keep patching and patching and patching and then discontinue support since there is no money in time spent on the patches and other upgrades which should have been done before it was released in the first place.

If they leave the core alone and do mostly UI changes, then consumers are getting last years stuffing put inside a new bird, if they come up with a new core then good luck getting it work right, secure and fully compatible, etc.

OSes are like languages spoken on this planet. Do we really needs all these languages? Why can't we all speak the same language? Then comes the question which also has the answer as well: Which one should people use?

[ahigherway]

Originally Posted by plwhit

Lets use FACTS rather than hyperbole shall we?

OS Statistics

Based on the comments so far I don't think anyone has read/understood the article.....

It has NOTHING to do with how many distros or the age/mentality of the user or the "deals made with computer producers"....

Well, I've read the article twice now, and I don't see what you're getting out of it that I'm not.

Linux is community-based. Microsoft is a corporation. Linux is developed by people who are interested in computing for a variety of reasons, and they use their creativity to give back to the community. Microsoft wants to make lots of money. (As if they don't have lots of money already.)

As for statistics?
To me, they mean nothing. Just because the majority of people do something (or don't do something) does NOT mean that they are right, nor intelligent. The only thing it means, is that they are in the majority.

My question remains: with so much variety and creativity in the world of Linux (which is growing, btw), why would anyone want to pay for a Microsoft license??? The only answer I can think of, is if one needs to use a specific application that cannot be ported to an open source OS. But even then, there are many apps that are also made for Linux that perform the same tasks.


Peace,
brian

My question remains: with so much variety and creativity in the world of Linux (which is growing, btw), why would anyone want to pay for a Microsoft license???

People pay for it because they want quality.

People don't want to deal with beta (at best) software that gets abandoned when the developer gets bored. They don't want to figure out which forked project they should be using and which ones are going to die after a release or two.

There are scenarios in which Linux makes sense, but the desktop just isn't one of them. The software quality just isn't there. The hardware support just isn't there. And there's little indication that it'll ever get there. I've been running Linux since the 90s and while it has made a great deal of progress, many of the same problems remain - flaky hardware support, buggy end-user software that gets abandoned, radical UI changes for no apparent reason.

It's telling that the only areas where Linux has been able to succeed is where large companies with a profit motive have invested in it and steered projects. (Well, that's not quite true, it's also succeeded in areas like scientific computing where the projects are tightly controlled by universities and the funding comes from government.)

[swagger]

Originally Posted by ahigherway

My question remains: with so much variety and creativity in the world of Linux (which is growing, btw), why would anyone want to pay for a Microsoft license??? The only answer I can think of, is if one needs to use a specific application that cannot be ported to an open source OS. But even then, there are many apps that are also made for Linux that perform the same tasks.


Peace,
brian

I don't know what else to say, brian. You've been given many answers, in this thread and others. It's like you can't read them or something.

If you actually do want answers to your question, you need to let yourself accept the simple fact that not everyone uses a computer in exactly the same way that you do, nor do they want to. You're stuck in this "linux is great for me, so it should be great for everyone" mindset that simply doesn't fit reality. It's been pointed out to you time and again through many threads in this forum alone, yet you refuse to accept that reality.