One of the maintainers of the Linux kernel has said that a lack of ‘credit or money or anything’ for those who test the open source OS could threaten its long-term stability.
Lack of testing ‘threatening stability of Linux’
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
92 Comments
2005-04-24 11:13 pm
2005-04-24 11:46 pm
2005-04-24 11:55 pm
2005-04-25 1:59 am
2005-04-25 2:12 am
Anonymous
1.) Why would you ever install gentoo for someone who doesn’t know what gentoo is?
2.) Almost every linux distro has an installer, except LFS and Gentoo (and Gentoo does have installers, just not in the main distribution; and it’s guide is an installer, you just have to type it in for it 
).
3.) You’re right that most people can’t even Startx. But why does this matter? They don’t have to… This is what: gdm, xdm, kdm, entrance(?); are for! Also, most of these same people aren’t really comfortable installing windows programs either, and the ones who are…are a little too comfortable with it (if you get my drift).
You’re right that in general linux/bsd are gonna be more difficult for people accustomed to not understanding while using their windows computer. However, this style of computer use has become a problem in general lately: Much of windows security issues do lie in completely ignorant users. So maybe it’s time we stop catering to this dream of a place where users can be clueless and start telling them to educate themselves.
Most of the problems people come up with for naive users switching end up being hot air; because the same problems exist in Windows:
Bad hardware support: Oh, and I suppose windows drivers are a haven of stability and good support software? Ever heard of ATI? Nothing but complaints about their control manager. Generally everything is supported on windows; but sometimes the support is just so insultingly bad! No driver is better than a bad driver.
Confusing interfaces: Microsoft Word. I know a lot of people pick on it, but really it gets on my nerves quickly…damn auto-formatting. Confusing interfaces are everywhere though, sometimes they exist because they make a lot of sense once you learn them.
Too many applications for one task: MP3 converting. How many free windows apps are there for this again? 600,000ish?
Anyway, that’s my rant.
2005-04-25 2:16 am
2005-04-25 3:18 am
2005-04-25 4:41 pm
Anonymous
“So, let’s see, this free software on my laptop is almost as stable as VMS, almost as secure as Trusted Solaris, almost as fast as NetBSD, and almost as usable as OSX and its FREE! What a deal! 
“
Almost as stable does not mean it is more stable.
Almost as secure does not mean it is more secure.
Almost as fast does not mean faster.
Almost as usable does not mean more usable.
The obviousness of these statements seems to have escaped you, as I can not see why else you posted what you did. In any event, we appear to be in agreement.
2005-04-25 6:50 pm
2005-04-25 8:53 pm
2005-04-25 11:01 pm
what have those multi billion dollar companies been doing? Dont they test the kernel and fix bugs, you would think they would. Do they just mooch off of all the free developers? Or even better how about they stop development in the 2.6 kernel, and make a 2.7 kernel to do development in.
Morton was saying that people who test the kernel don’t get enough credit for the valuable task they perform. Not giving enough credit make people less interested in doing it. Lost interest means less testing, ergo… less bugs squashed.
This is what the man was saying, he wasn’t saying the kernel is in serious need for testing, or that stability is in stake in the medium-term, just that everybody should give more credit to those who test (or money, which is nice too).
I think what Mortan is specifically refering to is the lack of testing of rc releases, which in the Linux world currently are not actually release cantidates, but more of pre releases.
This can be easily fixed by making releases that never plan on being released by naming them -pre or beta appropriately, instead of rc.
This would encourage more testing of rc releases.
“I’m basically keeping my head down until the dust settles and the shrapnel stops flying,”
it says “threatening the stability of linux” in the title, but andrew never says that. he says it could possibly become a problem in the future. i find lately that the tech news world is becomming as sensationalistic in their articles as the MSM. I would almost perfer a steady stream of “Linux is (not) ready for the desktop!”
I think the bigger threat to linux is openly disagreeing with Linus’s practices in regards to the kernel. Especially this Hindsight stuff i.e. use of bitkeeper. They need to keep their dirty laundry to themselves if they are concerned about the future of Linux. The user base doesn’t want to see instability in maintainers of a product that they just setup to house their mission critcal database. These guys need to grow up
yeah, cause ZDNet is known to be one of the most disreputable news sources, and andrew morton is a well known wintroll.
There is no reason to spread any FUD: Linux is extremely stable, especially since the 2.6 kernel. There is no version of Windows that comes even close to the stability of Linux and there are numerous studies published that support that fact.
If Linux out performs Windows and it’s a fact before it was supported with numerous published studies. Then linux is doing great, right.
This isn’t the army, we don’t need a leader. OSS is a bazaar. As long as there is someone we can gather round and trust that’s all we need.
What threatens science and technology is rampant ignorance in journalism.
Andrew Morton says …a lack of ‘credit or money or anything’ for those who test the open source OS could threaten its long-term stability.
Upon hearing this “news” you journalists go running around screaming: The sky is falling, the sky is falling!
Lack of testing ‘threatening stability of Linux’
There is no lack of testing. And it is not threatening the stability of Linux. However, it could threaten the stability of Linux if we agree with Andrew Morton’s opinion and don’t give our testers ‘credit or money or anything’.
But you will find there are a lot of testers who already get ‘credit or money or anything’ for their hard work.
Us GPL fanatics love our developers and testers. Even though we forget to pay them sometimes.
One of the most frustrating things (for me) has been the fact that Linux has been advancing so quickly but constantly that one is almost forced to upgrade every few years in order to be able to run any current binary packages.
I use Mandrake 8.2 on my PPro systems at home, and it’s been an extremely stable system (and yes, even its older version of KDE runs well on a 64MB box, and is usable if slow on a 32MB box), but it’s getting harder and harder to find software for it these days.
I suspect the state of flux is also frustrating for those who are doing the testing — you never know how long your results will last, and you’ll probably find yourself testing the same thing again in six months.
Growth and change aren’t bad things, but they can be the source of some frustration at times.
As long as there is someone we can gather round and trust that’s all we need.
What the hell do you think a leader IS?
..how lots of people for a couple of years have worried about how Linus himself may endanger the future of Linux. Morton has done more damage to Linux’s reputation in the last 6 months than Linus has done in the last 6 years.
If they don’t reel Morton in, he’ll totally destroy the kernel.
I think the problem with Linux is the classic case of biting more than it can chew. There is widespread false pride among the Linux community. Often you see things like “Hey look at this, we can do so and so thing” or “Hey look at this, we can run on this x architecture.” Immediately, this will be followed by comments such as “Our competition can’t do the same.”
However, generally the reason for this is not an inability to do this by the competing product but rather a deliberate choice made by it’s developers. Like jumping of a bridge, just because you can do it, it does not necessarily mean that you should.
So what we get with Linux is line after line of new code creeping into the kernel slowly but surely changing the kernel a unmantainable mess. I have always mantained this and continue to believe this, Linux cannot be everything to everyone. Linus simply needs to start moving things out of the kernel and only allow implementation of specific features. At the rate the kernel is going, we will soon have that “never stable feeling” Debian Sarge users have been feeling the last two years. Seeing bug lists of the 2.6 branch, I fear we might already be there.
Agreed.
Anybody have any reasonable explanation as to why Morton has been *consistently* undermining both Torvalds (who no longer needs any help) and the kernel itself for quite a while now?
There is nothing wrong with criticism but Morton’s comments are (a) often senseless, such as these and (b) a known and much loved source of quotes for “Microsoft Press”. One would think that even if he had everybody’s best interests at heart and valid complaints, he’d find a way or style or venue that wasn’t quite so destructive to express them. I’m suspicious that he wants the main Linux kernel and an ‘lt’ patchset. However, his methods make a Morton kernel even more disreputable and useless than a Torvalds one. So it really *seems* like a kamikaze on the kernel as a whole.
I don’t actually accuse him of these things – I’m saying that’s what it *looks like* and he needs someone to play Perens to his Torvalds.
What are you talking about?
I feel completely uninformed
. I was reading OS News (almost) daily for more than the last 6 months and I don’t remember reading anything ’bout Andrew Morton endangering the Linux kernel.
OTOH, why should I trust someone completeley un-identifiable?
. Provide links to support your claims, or be moderated down like a troll.
And my kernel works just fine, thank you
Lighten up on the “gotta get more hits” titles.
this is great to finally hear, WE NEED MORE TESTERS. MORE TESTERS FOR ALL THINGS. Bringith thi bug reports! Small or larges, minor or major share your constructive criticism today!
If you go to http://www.osdl.org you will see that they do current testings of the latest kernel. It is mostly database type testing.
I’m no fan of the 2.6 kernels, and i’m no fan of the current kernel development model, and i’m no fan of Andrew Morton. Linus, with his recent attacks on Andrew Tridgell has lost a lot of respect in my eyes – it was uncalled for, childish and shows that there was more to Linus using bitkeeper than meets the eye.
The 2.6 kernels are now becoming the playground of the large corporations, they are starting to dictate development. What they want for an “enterprise” kernel, they usually get. We had computer associates just very recently say that there’s too much “desktop stuff” in the 2.6 kernel. You know what? Computer associates: I don’t give a fuck. Go find another kernel to trash. Linux was a hobby kernel, used by (mostly) desktop users/developers. It’s we who started using the Linux kernel and getting the ball rolling, that made it semi popular. Only when it because semi popular did the big, rich bastard corporations start jumping on the bandwagon, demanding their share. I say tell them to piss off. Linux survived before they became ‘involved’ in its development, and will survive afterwards. The only real player that’s been positive to Linux has been HP. IBM will screw Linux as soon as it gets a chance. Computer Associates I wouldn’t piss on if they were on fire (it’s an aussie term hehehe). Sun are just two faced serial liers.
Andrew (imho) has done so much damage to the Linux kernel development model, and I suspect it’s irreversible. It’s been tailored to meet the demands of big business now, and not the average user. Obviously, that’s where the money is. I don’t trust the OSDL either. We’re going to see less and less relevant development to the desktop usage, and more and more to the enterprise development area. It’s starting to get to the point where the 2.6 kernel should be forked. And I don’t mean a 2.7 kernel tree, I mean someone else take the whole damn thing, fork it, and poach developers who care about the code, and aren’t up for bribes from the big corporations like what has happened now.
I don’t need RCU. I don’t need Dynix shit. I don’t need jfs/xfs. The vast majority of users don’t need it. Better to get rid of corporate crap like that, stub our noses at big corporations, and continue on as a hobby o/s.
That’s just my 2c worth.
Dave
You’ll be able to drop {insert *bsd like kernel here} onto a “Linux” distribution rendering the Linux kernel irrelivant. Shoot, maybe the best of all worlds would be Ubuntu with the Windows kernel and underlying components. I would probably use it. I don’t use Linux for the kernel, it’s everything else that makes the OS worth using. The kernel is the only thing that I’ve ever had issues with. When’s the last time someone complained about bash or gnome-terminal? Never, it’s always the kernel this and the kernel that.
Ok, lets change it. I’m in!
Security flaws in Windows typically aren’t because of problems with the NT kernel, but rather due flaws in the various services and DLLs. The stability of Windows is also often more affected by the quality of the drivers loaded than by inherent flaws in the NT kernel, and as such you’ll see a large variance in the stability experienced by users.
Presumably one could take the integral parts of the NT kernel and construct a more robust POSIX environment subsystem, and use this to provide the experience the other poster wanted without any of the baggage of the typical Win32 experience people bitch about.
It’s not going to happen of course. It seems more like that ReactOS will provide that before anyone tries to strip Windows.
linux is not stable. the rate at which 2.6 changes is scaring people who need stability. 2.4 was fine .. but 2.6 is crazy. BSD anyone?
“There are too many buggy drivers that aren’t being tested and are causing kernel panics.”
Solution:
[ ] Prompt for development and/or incomplete code/drivers
I’ve been saying this for nearly two years now. A good example being the complete and utter lack of a simple standard kernel debugger has long been a glaring omission in the Linux kernel. Combined with the absolute brain-dead decisions of the kernel maintainers (new features constantly being added to 2.6.x for example) lead me to have nothing in the way of faith in anything to do with Linux, and I’d go so far as to say that that I would not recommend it to people that I highly dislike.
This is what happens when an OS is more about religion than about being technically sound. Long live most other operating systems.
As I interpret it, he’s not saying that Linux is unstable *now*, he’s pointing out a potential issue that could impede future releases of the kernel. In other words, he’s trying to draw attention to it so it doesn’t fall by the wayside and get forgotten about, until suddenly the Linux kernel is an unstable P.O.S. and the life flickers out of the project …
… but that’s just my interpretation.
That Morton guy seems to be a real dolt. Furthermore, the values that should really get you going especially in this field isn’t credit, money nor something that empty, it should be that you’re doing exactly what you love to do, and the fact that people around the globe love what you produce, now that’s payment enough in itself I would believe. Let that be you nourishment.
So your contributions to the kernel are greater than those of Andrew Morton? Or is your “nourishment” simply to tell other people what they should or shouldn’t be motivated by?
Wow. The overall cluelessness displayed in this thread is absolutely staggering. Normally I would have jumped in here and started arguing, but this is too much.
Yes the sad fact there are a few journalists out there who are tied to certain O/S’s and they write a load of crap. That’s the best one can say. It is the readers job, to sort out what is FUD, truth, close to the truth or not. Sometimes it is difficult – other times easy.
“I don’t use Linux for the kernel, it’s everything else that makes the OS worth using.”
You mean, you don’t use a linux distribution for it’s kernel.
I think we need to FORK linux after a point release and manage a stable version which will not change bar bug fixes (like just about every other OS).
think we need to FORK linux after a point release and manage a stable version which will not change bar bug fixes (like just about every other OS).
Hi! I’ll help with the fork. I’m 12 and have been using Linspire since I was 11 and a half and so I know the kernel. We’ll make the best kernel ever!
I think we need to FORK linux after a point release and manage a stable version which will not change bar bug fixes (like just about every other OS).
Isn’t that that what the new 2.6.x.y releases are about? Perhaps they wouldn’t be maintained for as long as you want, but they do only accept bug and security fixes.
This can be easily fixed by making releases that never plan on being released by naming them -pre or beta appropriately, instead of rc.
I thought Linus stopped doing this because he was having trouble deciding where pre ended and rc began and also that the releases marked pre weren’t getting enough testing.
Did you even read all the way through my post? What I expressed was solely based on what _I_ would believe one should thrive on, what _I_ would consider mentally healthy. I would never try to force anyone to feel the same way that I do, it’s just that scenarios like this one happen all to often these days, and thus I felt the urge to chip in my two cents.
If your contributions are based on a wish of becoming financially wealthy or famous, then you’re in it for the wrong reason, that’s my opinion.
Just to clear out any confusion, my post was aimed at Anonymous (IP: —.dialup.mindspring.com).
O_o
not a task that easy (even for me as a CS graduate)… good luck kid 😉
The 2.6 kernels are now becoming the playground of the large corporations, they are starting to dictate development. What they want for an “enterprise” kernel, they usually get.
“Some rich men came and raped the land / nobody caught ’em / put up a bunch of ugly boxes / and Jesus, people bought ’em.”
I don’t need RCU. I don’t need Dynix shit. I don’t need jfs/xfs. The vast majority of users don’t need it. Better to get rid of corporate crap like that, stub our noses at big corporations, and continue on as a hobby o/s.
I’ve been saying this for years. Linux fans should have kept it to themselves.
“They called it paradise / I don’t know why / you call someplace paradise / kiss it goodbye”.
Going to get flamed for this but hear goes.
People have asked what about the commerical testing, is it not enough. You fail to see the big picture – it is in the interest of Redhat, Suse and IBM that bugs exist in Linux. When it goes wrong you have to go to them for support, they bill you, you go thats what I need support and around the circle goes. Its the entire business model for Redhat and Suse, and a good chunk of the IBM get their professional services claws into you.
Contrast that with the BSD model, there aren’t commerical companies behind it, so they make sure its done right in the first place, it is tested, they make sure its working and then it is released. There is strict control not a drive to support every bizarre device and wierd hack people want included. Its called controlling cost of defects. You can be damn certain that Apple and Sun approach it in this way as well. Microsoft (somewhat less successfully) too.
People are going to tell me that I can go search on google or recompile my kernel or whatever. But if I am two days from deploying a commerical application in say a telco and it goes tits up due to a linux bug I want someone to call for support then and there – so we are back to that support model again.
> Linux is extremely stable, especially since the 2.6 kernel. There is no version of Windows that comes even close to the stability of Linux and there are numerous studies published that support that fact.
Ok, then show us these studies. In real life it’s another story – Running Linux here on 2 computers for programming and every once in a while the whole computer freezes for no reason at all. This doesn’t happen on all the other computers of the company running WinXP Pro. Win98 was a real nightmare, but WinXP Pro is very stable, I mean it. Have you given it a try at least? You don’t seem to know what you’re talking about, sorry.
Reading alot of this I had to comment. Ive used Gentoo,slackware,mandrake,progeny,suse,knoppix,vida linux,beos,all versions of windows,currently running winxp,ubuntu linux and freebsd on different boxes.The ubuntu box after I figured out I had a defective ide cable lol,has been running for a month with no reboot.As has my xp box and my freebsd box.Most of the security issues Ive seen have been from either lack of information for users or incompetent end users security configurations.Ive found also that having some corporate created or big business kernel add ons can be useful or downright terrible.I’ve always added what I needed or not.Thats one of the best things about linux.Yes for the average user most people want everything preconfigured.Which I think can be done with linux as well much more so than windows.I do think the bug tracking needs reworked.All distros need an easy reporting system.I’m still learning programming so someone needs to write that.Also the beauty of open source.
> Linux is extremely stable, especially since the 2.6 kernel. There is no
>version of Windows that comes even close to the stability of Linux and
> there are numerous studies published that support that fact.
>Ok, then show us these studies. In real life it’s another story – Running
>Linux here on 2 computers for programming and every once in a
>while the whole computer freezes for no reason at all. This doesn’t
>happen on all the other computers of the company running WinXP Pro.
>Win98 was a real nightmare, but WinXP Pro is very stable, I mean it.
>Have you given it a try at least? You don’t seem to know what you’re
>talking about, sorry.
Ok then show us these studies. In real life it’s another story – Running WindowsXP on 2 computers for gaming and every once in a while the whole computer freezes for no reason at all. This doesn’t happen on all other computer of the company running Linux. Linux kernel 2.4 was already very stable, but LInux kernel 2.6 is extremly stable, I mean it. Have you given it a try at least? You don’t seem to know what your’re talking about, sorry
SO? the moral of this reply..its all about the users experience, hence we even run some Windows 98 workstation in a company and they ARE NOT unstable at all. You cannot compare apples with pears by trowing in some oranges and citrons.
SO? the moral of this reply..its all about the users experience, hence we even run some Windows 98 workstation in a company and they ARE NOT unstable at all. You cannot compare apples with pears by trowing in some oranges and citrons.
I was just giving the example of the company where I work, and as you said, one can’t drive conclusions from isolated experiences. However, this is not an isolated case, the company installs and maintains workstations of a whole portfolio of companies and experience says “If you can avoid Linux, this will save you time and frustration” and also regarding WindowsXP, it just works, this is what people want. When we say we can install Linux for half the price, they prefer paying and having something that can use and have their work done. They have no time to waist in learning a new, developper-oriented and primitive OS.
Leadershop is influence and the power to influence, like it or not… those who influence you the most on a given issue are in fact your leaders. Torvalds is in fact the leader as he has more influence on the linux world than any other (both directly and indirectly). As long as he has the influence, he will be the leader (in some way shape or form).
Even if that is not waht he was saying, there is an ellement of truth to it. While the Kernel itself might undergo quite extensive testing, many Linux applications do not. There is no real proper development, engineering, testing and quality contol cycle to much of Linux development. Often it is just a case of ‘if it appears to work, then it’s good enough.’ This isn’t at all the way it is done in large professional software development offices, where it can sometimes take two or more years to ensure that applications do meet a certain standard and will not cause significant problems at some point in the future. My one problem with Linux (as a long term Linux user) is that everything always feels as if it is in a very Alpha, or often pre Alpha state. The majority of testing is done in the real world, with a large body of prerelease software being released for every day general use.
If I had any say (which I don’t) sourceforge would be a central repository for applications and testers – and also for the kernel, to evaluate how applications performed against it, where any applications that were submitted had to go through several levels of testing, development and evaluation before being released to the general public.
I don’t think that anarchy is really the best approach to solving anything – regardless of how much of a hurry some people might be in to try to get one over on MS.
GJ
I’m sorry, perhaps you should read through my post in its entirety. It’s both quite short and doesn’t seem to imply anything you seem to be responding to. In short its point is that you have no idea what you’re talking about, probably don’t even write software, and your opinion about what should motivate others doesn’t create anything.
So maybe before you talk about someone who does contribute to the kernel what gets work done, while implying that they’re a dolt, you’ll actually use your love-powers to contribute.
I have to say that I just love the Anon posting above that was quoting the Eagles “The Last Resort”. Nice touch!!
And who (in competition with linux) does this?
>However, this is not an isolated case, the company installs
> and maintains workstations of a whole portfolio of companies and experience says “If you can avoid Linux, this
> will save
>you time and frustration” and also regarding WindowsXP, it
>just works, this is what people want. When we say we can
>install Linux for half the price, they prefer paying and
>having something that can use and have their work done.
>They have no time to waist in learning a new,
>developper-oriented >and primitive OS.
Here we go again…
However, this is not an isolated case, the company installs
and maintains workstations of a whole portfolio of companies
and experience says “If you can avoid Windows, this will save you time and frustration” and also regarding Gnu/Linux, it just works, this is what people want. When we say we can
install Windows for $300 more, they prefer not paying and having something that can use and have their work done. They
have no time to waist in learning a new, game-oriented and primitive OS.
For once understand that you customers are not mine and visa versa. I think Gnu/Linux is ready for the corparate and home desktop regardless of what everybody else thinks…stop.
“I’m sorry, perhaps you should read through my post in its entirety. It’s both quite short and doesn’t seem to imply anything you seem to be responding to. In short its point is that you have no idea what you’re talking about, probably don’t even write software, and your opinion about what should motivate others doesn’t create anything.
So maybe before you talk about someone who does contribute to the kernel what gets work done, while implying that they’re a dolt, you’ll actually use your love-powers to contribute.”
Well I read a short part of the article. So no I guess I wasn’t responding to that, rather than to my impression of Linux as a whole.
You may not agree with what I said, but I invite you, and indeed any linux user to list the software that they currently have on their systems that could be defined as ‘pre-release’ or ‘beta’ or in many cases often just ‘alpha’. (I don’t think this applies to the kernel. I think the kernel has a very good development and testing model. I just wish this model was applied to all the other software out there that is expected to interact with it).
In any case as I said, as a long term Linux user (does 5 years qualify as long enough?) who for two of these last years has not had any MS OS installed on their systems, I should hope that this would afford me at least some insight and some right to pass comment.
I think it is tough in the extreme that you should meet out such blatant derision, even when people (who themselves are ardent Linux supporters) pass even the mild critisicm of the current state of affairs.
It is probably impossible anyway. Not without large scale support from the major vendors. Since linux is open source, you could never stop people from forking away from any centrally approved and controlled repository, or from contructing projects outside of that repository and refusing to submit them for proper evaluation and approval.
The closest there is to this in the linux world is Debian. But then they take the exact opposite approach, by simply assuming that everything that is older is somehow better.
But at least they have something resembling a testing cycle – even if many have baked projects still do seem to occasionally slip through.
But anyway, as you said this may not have been the point that was being made. Unfortunately after the 5th Konqueror crash in an hour and the second kaffeine crash, I just had to let off some steam.
I guess if you never installed any window managers, and onl;y instyalled one or two base applications, none of this would really be a problem.
GJ
Ok, then show us these studies. In real life it’s another story – Running Linux here on 2 computers for programming and every once in a while the whole computer freezes for no reason at all.
I’m sorry, but I smell B.S. What distro are they running under? What causes the crash? (There are plenty of ways to figure this out, btw.) Does the Alt+SysRq keys work? Do the computers start working again if let alone for a while? And what do you mean by “once in a while”?
I think you’re making up these “examples” – either that or the guys setting up your Linux computers don’t have a clue.
In four years of using Linux, I’ve had perhaps four crashes. I used to get X freezes with some NVIDIA driver versions, but these you can escape from without having to reboot your machine.
This doesn’t happen on all the other computers of the company running WinXP Pro. Win98 was a real nightmare, but WinXP Pro is very stable, I mean it.
Yes, it is very stable. Not quite as stable as a stable Linux install, but about as stable as a more experimental one (Mandriva Cooker, for example). I’ve seen XP PCs freeze and spontaneously restart. As with unstable Linux systems, it usually has to do with funky hardware. But as far as uptimes go, *nix systems still have a strong lead on Windows and its variants.
The main problem I have with WinXP installs is that they degrade over time, like every other Windows version before. This causes WinXP systems to be quirkier as time goes by.
Have you given it a try at least?
The question is, have you really tried a Linux system? Because
You don’t seem to know what you’re talking about, sorry.
Neither do you, sorry.
But anyway, as you said this may not have been the point that was being made. Unfortunately after the 5th Konqueror crash in an hour and the second kaffeine crash, I just had to let off some steam.
I think that application stability is another matter entirely, and that it’s quite legitimate to criticize those who release unstable packages as stable. On the other hand, many Linux users (myself included) feel the urge to always upgrade to the latest version. That’s not always a good habit, unformtunately. I guess the best situation is to test such new programs (i.e. the latest KDE 3.4 builds) on virtual machines before installing them on your main system, but that can be a hassle…
May I ask everyone what the point of argueing about which is more stable, Windows or Linux, is on OSNews? Are you helping any projects become more stable? Are you bringing over anyone from the dark side? Is any of the stuff posted in this forum making its way back to the developers.
How about instead of bitching and argueing with each other on OSNews, you actually get off your fat lazy asses and file bug reports and *gasp* some of these issues may be resolved.
Linux doesn’t need testing.
Linux doesn’t need users.
Linux just needs developers:
-doing new stuff
-advancing version numbers
-making patches
Linux is just fine the way it is.
Its more stable than VMS.
Its more secure than Trusted Solaris.
Its faster than NetBSD.
Its got a better mascot than Windows.
It just works better than a mac.
…and you all you people don’t understand that then you must be americans…
Its more stable than VMS. BS
Its more secure than Trusted Solaris.BS
Its faster than NetBSD.Faster at what? Definitely not routing.
Its got a better mascot than Windows.Purely subjective.
It just works better than a mac.In what way? Certainly not usability.
You seem to like to make broad assertions with absolutely no supporting evidence. Isn’t that the way the Bush administration operates? Maybe you should think twice before a) making such prejudiced comments, and b) shooting off your mouth without a clue.
And you as well seem to like to promote your individual views to undeniable true statements that every sane person has got to agree on. I’d say that’s not a lot better than the previous poster who you were rightfully criticizing.
Everything is secure, until it gets compromized.
Everything is stable, until it crashes.
None of you have had the time to QA these products or have the authority to speak on matters of security or stability. You tested this system on what, 1 PC at your office? C’mon. Real QA involves tens or hundreds of test cases per application. Something like the Linux kernel has to be running several regression tests across several different architectures for each new release. And then all those drivers have to be tested. I can’t imagine the lab they use to test this thing. It has got to be impressive.
And then I’m sure SuSE and RedHat and IBM and HP have their own labs, testing varying distributions with varying patch and rev levels for each software component that makes up a Linux distro. Have you ANY clue how many packages that is? I’ve touched most of that source myself at one time or another. Each of those are tested by their maintainers. Then tested again at RedHat and SuSE and IBM and HP and mom and pop’s Linux shop. And there it gets patched when they find bugs and tested some more to verify those patches. Then each end user is given full access to source code and their apps often include built-in feedback systems so further test cases are generated for comparison when bugs are found.
Andrew Morton was commenting how the many eyes approach is working, but it could always be better with more eye doing more testing and patching. And those people who do this work and find critical bugs should be rewarded for their efforts and recognized a bit more personally by everyone. If they do this often I don’t see why they couldn’t be as influential as Linus or Miguel. Their experience and knowledge is just as valuable, if not moreso, in some cases.
The real issue is the hyping IBM has done – and mostly to attack the traditional Unix community, whose definition of stability goes way beyond 2.6. I still see way more stability in the Solaris platform, yet IBM is out there saying the opposite – they lose credibility, and people stop focusing on stability because IBM makes them believe they don’t have to.
I still don’t understand why you wouldn’t just run OpenSolaris, and skip all the hype (and expense) of paying for the supposed stability of Linux.
I have two reasons.
1. Access to all the source code.
2. Avoid vendor lock-in.
“And who (in competition with linux) does this?”
Apple, and the BSDs (probably Sun too)
Also, just because some big company doesnt test doesnt make it right for linux not to test. Learn from idiot’s (Microsoft’s) mistakes, dont make them
This has been the most moronic thread I’ve seen. Lighten up when someone posts broad canvasing generalisations and stop hounding for seriousness.
Euginia, another fine news post. Ta.
“And you as well seem to like to promote your individual views to undeniable true statements that every sane person has got to agree on. I’d say that’s not a lot better than the previous poster who you were rightfully criticizing.”
I’m sorry, is there really any denying that VMS is more stable than Linux? At best one can argue that Linux is as stable as BSD. To suggest that it bests VMS is a clear overstatement.
Or can either vanilla Linux or the majority of popular Linux distributions really compare to the security of Trusted Solaris? Perhaps SELinux and Trusted Solaris are comparable, where it is quite debatable as to which is superior to the other, but that was not what he was asserting.
As for the rest of my statements, I posed a question as to what Linux was faster than NetBSD at, and it’s inherently obvious that one’s preference of a mascot is purely subjective.
Lastly, I pointed out one possible way of interpreting what he meant by how an operating system works. By that interpretation, there really isn’t any question as to MacOSX’s greatest strength being superior to one of Linux’s notable weaknesses.
Regardless, a refutation is in its nature less broad than an assertion. It is only necessary to demonstrate a single case in a refutation, while it is necessary to demonstrate all cases in an assertion.
The advantage of credit and money (for testing):
http://weblogs.asp.net/scottgu/archive/2004/10/28/249458.aspx
There are more testers at Microsoft than there are developers now.
Microsoft have pulled themselves together quite well in the last few years. I am fully behind them.
Here we go again…
However, this is not an isolated case, the company installs
and maintains workstations of a whole portfolio of companies
and experience says “If you can avoid Windows, this will save you time and frustration” and also regarding Gnu/Linux, it just works, this is what people want. When we say we can
install Windows for $300 more, they prefer not paying and having something that can use and have their work done. They
have no time to waist in learning a new, game-oriented and primitive OS.
You gotta be kidding. You’re out of reality. If you’re a sysadmin/programmer, Linux is pretty good. If you work with any other thing (sales, marketing, communication, finance, etc…) you won’t even be able to start X, this is that simple. I said that it’s not an isalated because I base my experience on all our customers, not just one. Windows is not more “game-oriented” than Gnome, for instance, it all depends on what you decide to install or not, I don’t know why you said that !? Now, “primitive”, I think Windows is older and far more mature than Linux! Half Linux distros don’t even have an installer! Who can use it apart from developpers and sysadmins? The last time I installed Linux was for a customer who wanted 2 workstations for a mini call center for his company and he didn’t want to pay much. We delivered the 2 computers with Gentoo, ready to use, the guys saw that and brought the 2 computers back, furious and asked to install right away for the same day because he already the 2 attendants ready to start working.
You know, when you get bad feedback all the time from customers, it is not about isolated cases. One needs to offer what the customer asks for, period. We have to be open-minded, be able to offer different solutions, but not force the customer, this is being selfish, and it makes the customer mad.
The question is, have you really tried a Linux system?
Yes, there’s no doubt, every week we have to install Linux on new computers for cyber cafes, call centers and all that. It’s my job, if I don’t know how to install Linux and Windows, I can look for a new job, eh eh!
“Everything is secure, until it gets compromized.
Everything is stable, until it crashes.
None of you have had the time to QA these products or have the authority to speak on matters of security or stability. You tested this system on what, 1 PC at your office? C’mon. Real QA involves tens or hundreds of test cases per application. Something like the Linux kernel has to be running several regression tests across several different architectures for each new release. And then all those drivers have to be tested. I can’t imagine the lab they use to test this thing. It has got to be impressive.
And then I’m sure SuSE and RedHat and IBM and HP have their own labs, testing varying distributions with varying patch and rev levels for each software component that makes up a Linux distro. Have you ANY clue how many packages that is? I’ve touched most of that source myself at one time or another. Each of those are tested by their maintainers. Then tested again at RedHat and SuSE and IBM and HP and mom and pop’s Linux shop. And there it gets patched when they find bugs and tested some more to verify those patches. Then each end user is given full access to source code and their apps often include built-in feedback systems so further test cases are generated for comparison when bugs are found.
Andrew Morton was commenting how the many eyes approach is working, but it could always be better with more eye doing more testing and patching. And those people who do this work and find critical bugs should be rewarded for their efforts and recognized a bit more personally by everyone. If they do this often I don’t see why they couldn’t be as influential as Linus or Miguel. Their experience and knowledge is just as valuable, if not moreso, in some cases.”
Like I said, I think this is true of the kernel. But I think you are dreaming if you think this is true of the majority of most Linux applications that are expected to run on top of the kernel.
QA is an area of concern when it comes to application stability – because in the end most people don’t ‘see’ the kernel. They just tend to judge an OS based on how well everything works and how stably the applications on it run.
I hardly imagine I am the only one to say that my experience is that I have hardly ever had the kernel crash on me, but I have had poorly written applications crash on me on a great many occasions, so much so indeed that application stability really is my biggest Linux bugbear. I have also had poorly written applications physically lock up and crash the kernel. I’m sure that is not the kernel developers fault, but nonetheless it does impact negatively on my own and other’s impression on the usefulness and stability of linux in general.
A kernel is nothing without well written applications. Which is why a lot more focus really should be placed in this area. A QA stamp of approval for applications via a sourceforge type mechanism/central repository where apps wouldn’t be released to the wider public until they had been properly tested, would at least encourage people to think about QA and would perhaps cause many developers to interact with users more and maybe raise their game a little.
Is that really such a bad thing?
There is so much talk about Linux one day overtaking MS. But I can’t really see that ever hapening unless something like this is put in place.
QA and testing really should be pushed to the very top of the agenda. That really is the only way to win.
GJ
There is so much talk about Linux one day overtaking MS. But I can’t really see that ever hapening unless something like this is put in place.
I see more in some sort of symbiosis of the two.Combine the strenghts of both windows and linux and you have a killer OS.
If you work with any other thing (sales, marketing, communication, finance, etc…) you won’t even be able to start X, this is that simple.
Funny, X automatically start on my distro, as it does on all major distros.
Half Linux distros don’t even have an installer!
Really? You’ve counted? Are you including every obscure little distro ever made? Because ALL the major Linux distro, and nearly all the others, do in fact have installers.
How can even begin to believe that you actually work with Linux daily when you make such outlandish claims? In fact, it’s becoming pretty clear that you’ve invented yourself a company that allegedly provides Linux to customers, and are coming up with disaster scenarios just to further your own anti-Linux agenda. But it’s not working.
The last time I installed Linux was for a customer who wanted 2 workstations for a mini call center for his company and he didn’t want to pay much. We delivered the 2 computers with Gentoo, ready to use, the guys saw that and brought the 2 computers back, furious and asked to install right away for the same day because he already the 2 attendants ready to start working.
Another fairy tale? Or were you guys clueless enough to give someone 2 workstations before: a) figuring out what kind of software he needed on it and b) telling the customer what you had installed on it?
Yes, there’s no doubt, every week we have to install Linux on new computers for cyber cafes, call centers and all that. It’s my job, if I don’t know how to install Linux and Windows, I can look for a new job, eh eh!
Or you could simply stop lying and start telling the truth.
is there really any denying that VMS is more stable than Linux?
What is VMS? Does it run on a laptop?
Or can either vanilla Linux or the majority of popular Linux distributions really compare to the security of Trusted Solaris?
Fedora Core 3 include an SELinux security model. So I guess it compares with the security of Trusted Solaris and make regular old OpenSolaris look like swiss cheese, huh?
I posed a question as to what Linux was faster than NetBSD at, and it’s inherently obvious that one’s preference of a mascot is purely subjective.
NetBSD might be faster at routing than Linux, but Linux certainly smokes it on scalability. I seen benchmarks that prove this. I’ll leave it up to you to google them.
there really isn’t any question as to MacOSX’s greatest strength being superior to one of Linux’s notable weaknesses.
Ah, yes, what Linux lacks in usability it sure makes up in features. Tell me, does OSX come with a pager yet? Can you move windows by holding down a key while clicking and dragging from anywhere in the window? Does it shade windows when you double-click on the title bar? How many themes does it come with? Can you change icon sets?
And how much do you have to pay for each upgrade? I haven’t paid for my Linux upgrades for years. Technically you don’t have to pay anything ever, if you don’t want to. But I get a kick out of contributing from time to time.
So, let’s see, this free software on my laptop is almost as stable as VMS, almost as secure as Trusted Solaris, almost as fast as NetBSD, and almost as usable as OSX and its FREE! What a deal!
Linux is well known for its stability, particularly since the 2.6 kernel. Development is coming along great. Yes, the geniuses of free and open-source developers working on it deserve the greatest respect and a small donation wouldn’t hurt. This year my donation will be considerably larger and it would be great if you make at least a small donation to the Free Software Foundation and/or the various developers: http://www.fsf.org Thank you so much!
Or you could simply stop lying and start telling the truth.
I’m not inventing here. Come and install Linux on the computer of our bookkeeper and see his reaction.
I’m not inventing here.
A bold claim. Any proof to back this up? Your company must have a web site or something.
Come and install Linux on the computer of our bookkeeper and see his reaction.
If the software he uses isn’t available on Linux, and he isn’t willing to use a Linux equivalent, then what’s the point? It would actually be pretty stupid to force someone to use an OS without thinking this through first, just like I’d be pissed off if someone forced me to use Windows at home. Which leads me to believe even more that you’re making all of this up just to further your agenda.
You’re obviously new around here. In all the time I visited this site, I don’t recall ever seeing a nun, he moos accepting any posts critical of Linux or Open Source.
If you were to believe him, anyone who has bad experience with Linux is either liar or moron.
My advice to you: don’t bother responding to that raving lunatic.
Yes please do show us your web site. It would be very odd for a company selling Windows or Linux PC’s not to have a web site – or some kind of advertising somewhere. Indeed I think it would be positively unheard of.
What is the point of these silly stories? You shouldn’t come to a place like this and tell fibs, there are far too many extremely clever people here who will very quickly catch you out.
GJ
You’re obviously new around here. In all the time I visited this site, I don’t recall ever seeing a nun, he moos accepting any posts critical of Linux or Open Source.
How mature of you, Russian Guy, reprising my post from a different thread and replacing your name with mine. I guess my criticism hit its mark, and you had problems taking it in…
As it happens, I have no problems accepting posts critical of Linux or Open Source when they’re based on truth, not on pure FUD. The problem is that lots of people here like to spread FUD, and I just do my best to dispel it. For some reason the truth seems to annoy you…
If you were to believe him, anyone who has bad experience with Linux is either liar or moron.
Not at all. I’m always ready to help someone who has problems with Linux, and in fact I contribute to help forums on a regular basis.
However, when someone makes obviously misleading claims like “normal people wouldn’t know how to start X”, and then proceeds to provide highly dubious claims about how customers were angry that they were given Linux workstations, I take out the BS-meter and make a reading. In this case, the BS is right off the scale.
My advice to you: don’t bother responding to that raving lunatic.
Weren’t you the one that kept complaining that I somehow insulted you in a previous post? Well, I’m glad that you finally showed your true colors…
If Linux were so damn good, everybody would use it. That simple.
As you asked for it:
http://www.intecma.com.br/
If Linux were so damn good, everybody would use it. That simple.
Wrong. The best technology doesn’t always lead in market share. Case in point: Betamax vs. VHS.