Hey guys haven't been on forums for a while, but just read this article and thought, that I have to share it!

http://tirania.org/blog/archive/2012/Aug-29.html

My guess is that this guy(as a ex lead gnome developer) is offended that people are pointing to the mistakes gnome team has made.

I am using iMac myself and the machine itself is a masterpiece and it's just a joy how great everything is configured!

But the OSX su**s my ba**s! It's terrible(software installation, software SUPPORT, use restrictions), but the most striking thing is that equivalent applications on OSX compared to Linux runs much MUCH slower and I am gonna fortify it even more I run those applications on Linux distribution which is running in Virtual Box(that of course comes to the question of the quality of applications themselves for each platform)

Also I just hate that OSX have stripped down versions of applications, for example VLC, in OSX say bye, bye to Advanced Controls.

So "many" users have migrated to OSX on desktop 6-7 % while ehm Windows have 90% Not a lot has changed! xD Oh and cloud computing, oh yeah... don't make me laugh.

Fast evolution and configurability of Linux has been and are the best things! Windows - 8, OSX - 10, Linux still 3

You are not alone in that thought.

Somewhere there is a thread about this article, but I could not find it, so here's the article:

http://www.theregister.co.uk/2012/09...ailure_flames/

I have never been much of a Gnome fan, but I was willing to use it up until v. 3. I did a click test--if you are switching back and forth between programs, as I commonly do, it takes more clicks to do so in Gnome three than in its predecessor or in more traditional desktops.

I'll risk getting labeled a heretic, flamed to Hell, and criticized heavily, much less greatly frowned upon, and say what I think:

We abandoned our UNIX roots, and we royally screwed ourselves by trying so desperately hard to not be another UNIX.

Linux has become something it wasn't intended to be by the likes of Lennart Poeterring and developers like him who have stripped out or prevented the backporting of tools existent across every aspect of UNIX by claiming they are out dated, work poorly, and don't contribute to Linux's advancement. On top of that, we have Red Hat which, by all means proved Linux could be marketable, acts like they own the whole damn thing and nobody else has a say in anything, but them.

We abandoned the tried and true like hald, devfs, and hotplug system daemons and toolkits for the wondrous yet overly complicated and rule sloppy udev, and we were quick to abandon another tried and true tool, sysvinit for systemd which now merges in udev, which all sounds nice until you hear it's incompatible with non-Linux systems and mostly incompatible with some Linux systems, especially Linux From Scratch. We could have easily imported devd from *BSD which was the same thing as udev but a much cleaner design and isn't sloppy with rules.

But we didn't. Linux is all about being liberal with evolution, change, and being free... but at what price have we paid for all this freedom and unrest within the spectrum of Linux distributions? There is no set in stone specification for Linux distributions to adhere to like POSIX. Linux Base Standard is a complete fallacy and can not even be upheld by any one distribution, and no distribution can even say they are GNU compliant as well. We have nothing to guide us and nothing to bind us all together. Linux is, for the better choice of words, a blade of grass flowing in the wind endlessly in utter chaos.

We look at non-Linux systems like *BSD which every release cycle starts a full out code freeze and code audit to which a base system can be constructed and managed with everything else patched as needed, tested for compatibility and compile problems, and then a full release is published with little issues other than unresolved issues from the original developers regarding outstanding bugs. Does Linux have even ONE distribution that employs a code freeze and audit other than LFS? No we don't! Most distributions are either rolling release or evolutionary in terms of how packages get added, updated, and managed. No effort is ever made except by LFS to break down the system and make sure everything works from the top to the bottom. if you look at the LFS book they use specific versions of software and even have "Works with LFS 7.2" or whatever because they know the code is stable and works with the compiler because they took the time to freeze their code, audit it, and ensure it works not just for them but everyone.

Could this all be fixed by the same people who caused all this? yes, and it starts with the people who not only maintain the various projects for Linux, but also the maintainers of the Linux kernel, and other projects related to it.

Linux needs to be developed more stably and less erratically. We don't need a new kernel every 7 days, we need a stabilized kernel with usable patches developed over the long term period and not the short term.

System software has to be compliant with some level of LBS, POSIX, etc. We have to uphold our UNIX roots and maintain the fact that we can contribute to other non-Linux systems with the same tools and software not just see them as operating systems and kernels that are outdated and don't matter, as Lennart Poeterring put it, and we have to allow non-Linux systems to contribute to Linux.

Lennart Poeterring is wrong as Hell when he claims *BSD is outdated and doesn't matter. *BSD has done things with great care than Linux has ever done since it's conception. If you look at the way older tools like hotplug, devfs, hald, work together with *BSD's devd toolkit, it's way cleaner than udev ever could be, even if the software is divided between four packages. Those four packages do just as much work, if not equal to or better work than udev can.

One other thing... and this is going to really piss someone off, I know it... I personally think the GPL license the Linux kernel is under is a pitiful farce and doesn't allow for real open source development of features that could be beneficial to the system and beneficial to other systems as well.

Look at ZFS, a great file system that has been used in *BSD and Solaris based systems for about a good decade now. But can Linux get this? No because the GPL doesn't play nicely with the CDDL license ZFS is under. If a license that actively preaches open source and free software can't work with other free open source software licenses just because of petty issues, is it really open source? I think not. In fact I see GPL as a highly restrictive license that while does allow for open source and free developments, is like an exclusive club with member's only access. If you aren't in the accepted list of allowed persons, you don't get in, even if you have the same ideals and fundamentals and your name is on the guest list.

A great maintainer of a Linux long standing distribution, I'll quote said, "Lao Tzu said that in order to lead, one must follow."

What has Linux followed? I ask this wholly. What has Linux followed to become a leader in the UNIX community?

Chaos leads to uncertainty, uncertainty leads to instability, instability leads to unrest, and unrest leads to chaos in a vicious cycle only to start again.

If Linux wants to lead, we first must restore it's roots. No tree can grow and spread it's branches without a solid root system. For that, we have to do the unthinkable, the unwarranted, and even the unwanted and heretical... we have to restore Linux back down to the kernel to it's roots, and maintain some ability to follow standards throughout the system attempting to adhere not just to the standards set by the LBS but to standards like POSIX as well.

I wouldn't say it's dead as a desktop, but it's just not taken seriously because everything is so erratic and chaotic. How can people take you seriously when the system is in chaos? They can't, and they won't.

And unless Linux gets it's act together, matures, tosses out the trash and restores it's foundations completely, you might as well write an article entitled, "Is Linux dead on the Server?" and publish it.

When you are serious, people will take your seriously. For Linux to be taken seriously, it has to become a serious project without chaos, without disorder, and without non-standard compliant junkware making a muckery of the whole.

have you ever wondered why Apple chose to based OS-X off of a *BSD foundation and combine in elements of NeXTStep and OS-9 for an entirely new operating system?

Because *BSD was always stabilized at some point each cycle to where they could take the project down to the source, fork it, rebuild it without any issues, and recreate it as something new. OS-X abides by standards like POSIX and such, and not only that, but OS-X has been marketable too through it's stability.

Could OS-X ever have been based off of Linux? I highly doubt it.

3 members found this post helpful.

@frankbell Thanks, very good insight!

I'm certainly becoming frustrated by third party software on Linux breaking, for good sometimes, due to it being written against an older system. The moving-target nature of Linux means programs like Google Earth (my main and current frustration) can just stop working indefinately because they were written for an older setup (I know it still works for most people but don't be surprised if it stops working).

Hmm

Just checking my Linux desktop:
Pulse: normal
Breathing: yes
Blood pressure: normal
Reflexes: normal
IT'S ALIVE!!!

Hi,

Linus Torvalds on the Linux desktop's popularity problems is another argument that does provide some insight.

As detailed as ReaperX7's thoughtful posting might be, I do not personally agree with it.

I think that every operating system naturally and insistently evolves over time. And, in so doing, it becomes more or less incompatible with what came before it, precisely because that is what it must do in order to move in any direction at all.

Either you change position, or you stay where you are. Them's your choices. Either you choose that the decisions that were made on a PDP-7 in the 1970's should remain untouched, for whatever reason you come up with, or you change position. If you change position, in any direction in any amount for any reason, you're going to both attract people who are glad for the change, and irritate people who for whatever reason don't think that it should have happened.

And, I'm just going to leave it at that. "I respectfully dissent." I think that the moves which ReaperX7 decries were, by and large, in fact good moves. I think that the GPL license is very well thought-out and appropriate. In so saying, I say absolutely nothing about any individual whatsoever: this is an engineering decision.

---

As for the OP's question about "laptops," I think that while Linux certainly can run well on them, many laptops take a lot of short-cuts in their design and rely upon software to cover for it. Or, they simply take short-cuts in their design. Some of these certainly cause problems for Linux; at least, they did for me on my cheap stuff. They're chronically underpowered as a lot.

But, again for whatever reason, laptops are not where you most-commonly see Linux. You see Linux on a rack, on a tablet, on a phone. I don't perceive that the actual market demand for a Linux-powered laptop is that strong. I begin to question the demand for "a laptop" these days.

Hi,

You will on occasion get one or two to agree on something at some point in time. More than not you will have multiple points that could possibly align at some point in time.

In reading what 'Linus' and other have stated as to the evolution of problems for the Linux Desktop. You have one thing that sticks out, 'standards'. If the 'kernel' developers worked on designs as some Desktop developers have worked then the 'kernel' would be a mess. Thankfully some sound design criteria for models were adhered too. I do not always agree with some points that Linus has put out on the Etherworld but at least he stuck to his guns.

Diversity is not always a good thing when designing anything. Standards are a good thing to develop or base your model on. That does not mean we should hamper innovation but at least provide a starting point for young developers. The old 'Why re-invent the wheel?' applies here. If the wheels' performance is poor then take the design and rework to suit the new device. You would still be working with the 'wheel' to develop a better 'wheel'.

“A tool is but the extension of a man's hand and a machine is but a complex tool. He that invents a machine augments the power of man and the well being of mankind.” - Henry Ward Beecher

“One machine can do the work of fifty ordinary men. No machine can do the work of one extraordinary man.”- Elbert Hubbard

Standards, as you mentioned onebuck, do make or break a model.

Like you said, when you have a foundation, you know where to start, and you know where and how far you can, and should, go. When you have poorly formed or abstract standards that exist in a "living breathing" type of existence, then you have a weak, or no, foundation at all.

While it's good that some packages for Linux have evolved naturally, some haven't evolved well or haven't evolved at all, or even evolved too much to where the evolution breaks the design model.

Originally Linux was conceived, from my point of view, to be an open source alternative to UNIX systems. While it did pair up with the GNU project to formulate a base model operating system, GNU was also designed to be a system that supported any kernel design that could be imported to it. This meant if any other kernel and module set was dropped in as a replacement, it could run, because that's what GNU was designed around.

However, this as a standard for GNU. An open source operating system that catered to anyone wanting to use it for their projects and kernels. GNU had been working on HURD for the longest time with the GNU Mach microkernel. It wasn't completed but then Linux comes along and was able to create a working system which was dubbed GNU/Linux and completed the system until HURD and Mach could be eventually be brought up to speed.

This is where, as I pointed out Linux has it's greatest weakness. It has no fundamentally set in stone standards. The Linux Base Standard model has never been fully adhered to or actively supported or even endorsed. As a result we have chaotic projects coming in from these so-called "revolutionary developers", like Poeterring for example, who claim to have the latest greatest medicine for Linux and it's the next "big thing".

Ever heard of a "Snake Oil Salesman"? Back in the 1800s there were all these so called "Doctors" who traveled the American West soliciting "Wonder Cure" elixirs as breakthroughs in medicine. Most of these elixirs were often toxic or useless whiskey based concoctions that did nothing but either get you drunk, killed you, or did nothing at all.

These "revolutionary developers" are the same. They come to the Linux community soliciting the latest greatest software cure for Linux that will make everything better like medicine, yet in the end, we find out, nothing has been made better, everything works just as poorly or if not worse than it did beforehand, and now we're stuck with a bottle of "Doctor Poeterring's Miracle Cure" that isn't worth the bottle it's poured into.

Hi,

We have standards but some people just shoot from the hip. Most professional programmers do a pretty good job but at times some will get sloppy or just plain lazy then things can drift into a glob that needs major rework.

Red Hat is a good model to look at. I do not use nor recommend Red Hat but they do follow standards. And sometimes lead in providing standards papers.

I really think early Linux distribution(s) were developed to be a free 'UNIX-like' OS on the PC at the time of origin. Not a replacement but utility to users to have a free OS that they could associate or feel comfortable using on their Personal Computer.

I will argue that Linux (kernel) does have standards and are very tightly controlled by Linus and other contributors. It is to bad that a model/standard is not available for distributions. We would have limited number or varieties of distributions. I really believe that is why some users align to a particular distribution, it feels right for their needs.

It is only partially true and distro dependent.

<flame_war_bait>
Take Ubuntu from one release to the next and all known bugs and workarounds are replaced by unknown bugs without workarounds.
</flame_war_bait>

But take Debian, and install Stable. Live happy, and install any package from Stable that you like during the next two years until the next stable release.

When the next Testing version becomes Stable, do a dist-upgrade and 99% of all packages are correctly upgraded and your system is new, shiny and current. In other words: it hardly ever happens that I am locked out of my productivity because of an incompatible program.

Debian even managed to squeeze in (pun intended!) a KDE4 version which was actually usable. No minor task at all!

You are right, Windows does it better. Programs from XP run on Vista and 7. BUT it has been numerous times I wanted to install a package on a W2k machine and that package was simply not availabe for W2k. Why do I want to upgrade Debian and not W2k to XP. I have to pay for it!

jlinkels

(Shrug...)

But what are "standards," really? Linux did not set out to be "an open-source copy of UNIX®." It set out to be "Linux®," and that's precisely what it is.

The core fallacy of "standards," when taken beyond the point of engineering reason, is that they seek to lock you to what the status-quo used to be. As though that "former status-quo" was somehow something handed-down from and by the gods. But that was never the case: someone faced an engineering problem in their time and solved it in a way that other people really liked. The actual engineering demand is always for a system that constantly evolves to address new requirements, and that is free to do so. "Requirements" are never frozen in time. There are no fixed targets.

The Linux system today runs on more than 25 very-distinct hardware platforms. Its developers are not looking at new requirements and apologizing, saying, "I'm sorry, but we can't do that for you ... it's not The Standard."

Instead, they are putting their heads together and "doing it," knowing that what they have done may well become "The Standard." Or, I should say, another Standard.

Let us also recognize that nearly every underlying operating-system provided API is accessed by application-programs through the use of libraries, not direct system calls. Many of the pragmatic differences between operating systems (specifically including "this is Linux, this is Unix, this is OS/X") are often absorbed more or less within that library layer.

But the direction in which you choose to move is important.

Old geezers like me remember the ballyhoo for the new automobile model year in the US in the 1950s and 1960s--all the chrome gets moved around and underneath nothing's new.

Too often with things like Gnome 3 and Unity and Pulseaudio, I see chrome moving, just because it can.