Opinions on Linux Kernel

PDExperiment626 · 2007-08-20 01:34:08

Hi, I've been using linux for about ten years now. I have some experience with every winOS except vista (thank god), OS X, OpenVMS (vax and alpha) and a few other *nix variants (I used to be employed by DEC back in the day . Anyway, my love for Arch is on par with my love for OpenVMS as an OS; yes, I love OpenVMS.

Anyway, I am really not keen to give up my linux operating system; but recent developments with the linux kernel have me concerned. It's no secret that since 2.6.21, stability has been a huge issue on laptops. In my own experience dynamic tics/hpet drivers will consistently cause hard-locks on my machine. I won't even go into my feelings regarding the decisions being made about incorporating the CFS scheduler over SD. All I will say is that with CFS, if I play music on my machine, my music player can be the only thing running. Underload, video and sound streams become unusable due to all the static under a CFS scheduler. I have never experienced this with any variant of the staircase scheduler... ever. At the moment, I am running the 2.6.22 kernel with the standard hz timer, con's final patchset (and a number of other patches) and no preemption.

From what I've seen, forced preemption also is not a stable option for the kernel at this point. Regardless of the scheduler and kernel config, forced-preemption will eventually cause my machine to lock on video streams. I took a look at ingo's realtime-preemption patches at one point; and if I could get my machine to run for 3 minutes without deadlocking under that patchset, I was doing good. I'm not exaggerating.

What I've been seeing in the linux kernel development has me very concerned as a laptop user. Kernel features and releases keep moving forward while the bugs of previous releases are scarcely mentioned. You have issues now where distros are having stock kernels being installed that are causing deadlock on laptops. This isn't an isolated incident; it's been going on for months and it's not being acknowledged by the kernel devs as a real problem. If these problems aren't fixed by the next time I upgrade a laptop, I will be forced to go to a bsd variant or to a... mac (it burns my fingers to type the name...).

I am wondering what other arch user opinions are on this, as arch seems to be an OS oriented more toward the optimization savvy linux user. Am I overreacting or do others feel as I do? As it stands now, I will be sitting on a SD 2.6.22 kernel for the forseeable future.

Last edited by PDExperiment626 (2007-08-20 01:37:50)

jdhore · 2007-08-20 02:13:59

I completely understand your concerns, but i haven't had the problems you've had with the CFS Scheduler. I haven't used 2.6.23 very much (It's a bit slower than 2.6.22, but i think it's mostly because 2.6.22 is a package and 2.6.23-rc3 i compiled myself), but it's actually been a little less CPU intensive on my day-to-day tasks than 2.6.22 was/is. I'm able to browse the web with Firefox, have Xchat running, be able to talk on Skype, videos in VLC play fine, songs in XMMS play fine and Flash videos are even a bit less CPU intensive than they were.

chaosgeisterchen · 2007-08-20 02:29:20

Good night,

interesting post to read, but I do not think that I have got enough knowledge about the Linux kernel and its architecture (as many users do) to fully understand the problem(s) which occured.

For me - as the (usual) end user - the kernel development has shown, that Linux becomes faster, more stable and even more comprehensive with every new kernel release. But my point of view changes by readings postings like yours, believing that you are not keen on spreading FUD on purpose, like CK probably did not too long ago.

The new scheduler is indeed an interesting development, the Linux kernel is facing, because I am very much of the opinion that we have to do everything we can in order to reduce power consumption under Linux to make it usable for mobile computing. It's probably possible to cut it down to around 88-91% of what an optimized version of Windows XP is consuming, depending on the task one is facing while working with ones notebook. Mobile platforms are those, which Linux has to concentrate on mainly, as notebook sales soar with desktop sales dropping, let alone the fact that mobile hardware has always been more complicate to be supplied with appropiate drivers.

On the other hand, Linux shouldn't turn its back on regressions and old and ever recurring bugs, which have to be fixed. Linux sells itself branding itself as super fast, super secure and super reliable - apart from the fact of being totally free in terms of free beer and free speech. One of Linux' most profitable advantages is its performance on older hardware which resigns upon being faced with Vista. More and more people will stand before the decision: Either paying hundreds of dollars for new hardware to achieve the same performance under Vista they enjoyed under XP/2000/etc or converting to Linux, leaving old patterns of computing behind and enjoy an faster, more reliable and super secure system with gorgeous eyecandy. I am confident that the free and open source alternative will score, if we both move forward in terms of supporting brand-new hardware properly (I know that the major hardware vendors have to participate in this task, but soaring market shares will force them to support Linux) and cleaning up regressions and bugs, keeping the hardware requirements low by using the force of scalability.

Linux can and will be once the ideal operating system for each and every mainstream platform (i386/i686/x86_64) computer and I am willing to support this development.

But first I'm looking forward to reading comprehensive postings from people with kernel knowledge, as I am more and more interested in this topic.

Regards
~cg

Misfit138 · 2007-08-20 02:59:25

I too have no programming knowledge or experience, but I do see differences in performance from time to time. When I first installed Arch (0.7.2) I was amazed at how fast it was. Then after using the beyond kernel, I was even more impressed. After beyond was abandoned, I went back to the regular arch kernel and I definitely noticed a loss of responsiveness. Now, after a pacman -Syu a few days back, the repsonsiveness is back again, and my system feels as fast as ever. This could be attributable to the kernel update which occurred on my box, or the KDE update, ( I have no way of knowing, due to lack of knowledge) but this system definitely is noticeably snappier than a few days ago.
I hate to read about the politics behind kernel development; I find it extremely frustrating to hear of how Linus Torvalds has such a strong opinion and influence on matters, while claiming to be a hands-off "guide" rather than a "leader". It was disheartening to me to hear that the CK development was basically rejected by Linus, and then copied by another programmer (CFS), to whom Linus gave his blessing, and will now be implemented in spite of its supposed inferior quality and perfomance in comparison to SD.

iphitus · 2007-08-20 08:44:14

PDExperiment626 wrote:

Anyway, I am really not keen to give up my linux operating system; but recent developments with the linux kernel have me concerned. It's no secret that since 2.6.21, stability has been a huge issue on laptops.

And it's no secret, that since early in the 2.6 cycle, the kernel's stability has generally gone downhill.

From what I've seen, forced preemption also is not a stable option for the kernel at this point.

Last time a compiled a kernel, forced pre-empt was not the suggested option for a regular desktop. Unless you create/modify video/audio, your laptop does not need to be a realtime machine, and you don't need forced pre-empt, nor rt patches.

My concerns are:
- politics. It's inevitably becoming political. I don't know can be done about that, it happens to any large project.
- egos. Often LKML isnt the most professional place. nuff said.
- stability. it is going downhill. I wish I'd taken statistics, but after each release, there's an increasing amount of people complaining of breakage here on these forums. Possibly development is too rapid and aggressive.
- current development model. I don't think it works well enough. There's a multitude of various patches I also believe should have caused 2.7.

James

FeatherMonkey · 2007-08-20 12:00:12

Doesn't it come down to what CK said I know it was dis-agreed with, but lets be honest who is sponsoring the development of the kernel if its not Enterprise eccentric distro's/manufactuers.

So surely as much as the devs said we run linux on Desktops so we make sure it runs. The diff is when the Enterprise ecentric people bug report these guys need listening to, wages depend and lively hoods on there sponsorship.

Which goes back to the comments in regards to free, now firstly before I get flamed I'm not saying its right I'm not saying its wrong. As has been pointed out one of the main things is hardware suppport, now you only have to look at the flagging of tainted kernel when you load nvidia.

Now I'm not saying its wrong but there's a strong feeling in the linux community in regards to OSS being good and closed the anti-christ. To me this is what is holdng back linux this belief that we can force the world to go open source over night this just isn't going to happen. If we wish to change the way people do things then we have to gain a large say.
Now this is never going to be achieved whilst we do this. only recently I had a lively debate in regards to this. I asked would a single linux way help software houses and then proceeded to point out that binaries could just as easily be put into tar.gz with distro specific instructions. I got hit back that this wasn't a pkg format, I argued I accept the point but if we made it easier for the software houses to keep there code closed. Perhaps then if they see we are supporting there business model they may support us. I see a big difference between the business model of closed selling versus a support orientated OSS.

Then if we remove this attitude of open is only good is it not possible then software houses will embrace linux, combined with driver support. Take another difficult subject again I'm not saying its wrong or right, but why the hell are we getting all this extra download bandwith if its not for streaming media. Here's the problem DRM so many of these streaming medias are protected with DRM, so the linux user is left a 2nd class citizen. Combined with the latest storage media which if I'm not mistaken at the mo involves an XBOX to achieve this(Though my facts maybe wrong here).

I mean so often I see the argument that software houses won't support us because there's not a single way of install, but I never see the counter argument that you can change a deb to rpm to tar.gz and what ever other way you wish to do. So is this not perhaps they just don't like the Open source business model.

All in all what I'm saying is first the desktop user needs to be the ones supporting the development of the kernel for us to achieve this(How many of us really do buy the distro disc's). Now will this happen in the near future, mmm bit of chicken or egg going on here, to get an invested interest we need to get better hardware support, to get this IMO we'll need to accept that some guys just like to use Closed to protect there interests. Whilst we keep downloading and not supporting the distro's we'll be left with sponsorship by companies to keep the development going.

There's a reason why 75% of linux users in the desktop survey still use Windows, and its not because of open sourced software. Open source is here to stay but to make waves it needs to find a solution to get Closed source Software/Driver houses to develop for us, But as much as Open source is here to stay so is DRM and closed source.

Just my opinion and like others previous but I have little knowledge in regards to Kerenl devel but to me this isn't a question of development but more a Social question. About how to make the common desktop user pay the kernel devs their wages so they do depend on us, rather than the big corporations?

Last edited by FeatherMonkey (2007-08-20 13:36:34)

PDExperiment626 · 2007-08-21 04:27:45

It's been interesting reading all your comments; and I agree with many of the things being said. Again, it's not my intention to spread FUD regarding the kernel or related issues; I'm just trying to get a range of views in hopes to get a clearer understanding or what may or may not be going on with the kernel politically and technologically.

In regards to CFS. I am glad to here that some people are seeing performance boosts with the .23 kernel. Personally, I became infatuated with the CFS scheme of using BR trees in its scheduling algorithm; but as I said, I just wasn't able to make it work as well as SD does on my laptop. Although, to be fair, my wife also runs arch on an AMD64 workstation; and I have not noticed a significant performance hit or boost between the usage of CFS and SD on her machine (the kernels were otherwise built with identical configs). On her machine, I have not heard the static in video or audio streams that occur on my laptop when I run CFS. But then again, she built her machine oriented along the lines of being a high-end graphics workstation as opposed to a conventional desktop with lower-end hardware.

As for preemption, I've always been a no-preemption type of uses. I prefer the lighter weight kernel, better throughput and better stability of a non-preemptive kernel, as opposed to the slight improvements in UI performance I see with preemptive kernels. I gave various levels of preemption a shot out of curiosity, to see if I wasn't just being a server-config bigot. After using various levels of preemption on my laptop and my wife's workstation; I've concluded that I still prefer non-preemptive kernels for these particular machines. In the end, I want a non-preemptive kernel with a good processing and I/O schedulers as opposed to heaps of preemption with mediocre scheduler performance characteristics for my workloads.

I agree that the orientation toward server performance has financial validity for the linux kernel. That being said, Linus seems to proclaim himself as a technological idealist in that the kernel should be above such financial interests. I.e. he makes decisions about what goes into the kernel and what doesn't based on merit and not necessarily financial benefit. I also understand what people do and say can be vastly different. Nonetheless, while the server market does correlate to revenue for linux kernel development; linux has definitely been trying to get more desktop market share in the last few years. The reality is, given the current state of Vista and Ubuntu, there is a likelihood of linux gaining a significant chunk of that desktop market share within the next few years. In this scenario, I am not worried about kernel stability; someone will take the initiative to make the kernel run at least somewhat stable on an average desktop. What I am worried about is a group like ubuntu taking the kernel and recoding it so that its more tuned toward desktops. In that situation, you are looking at the possibility of Ubuntu making their desktop kernel proprietary and even worse... closed source (case-in-point launchpad). It would be windows all over again.

Anyway, I know that any number of things may or may not happen in the future; and that ultimately, what happen, happens with the linux kernel. I just don't want to be caught in a situation where I'm caught off guard by a potential outcome, political or otherwise, in regards to linux. At this stage, I am not sure what the real motivations behind kernel development are. Are they financial, technologically meritorious, nepotist, egotist, technologically over-zealous in nature or some some mixture therein? While it would be helpful to have a clear perspective on this; I fear that getting such a frank truth about such will be nearly impossible.

shining · 2007-08-21 08:08:22

iphitus wrote:

- stability. it is going downhill. I wish I'd taken statistics, but after each release, there's an increasing amount of people complaining of breakage here on these forums.

Just in case : are you taking into consideration the probably increasing amount of users ?
Anyway, I personally couldn't notice a decrease of stability over the years.

PDExperiment626 wrote:

At this stage, I am not sure what the real motivations behind kernel development are. Are they financial, technologically meritorious, nepotist, egotist, technologically over-zealous in nature or some some mixture therein? While it would be helpful to have a clear perspective on this; I fear that getting such a frank truth about such will be nearly impossible.

Believe me or not, I trust Linus more than Bill (or Steves for that matter) on this one.

pecan · 2007-08-21 11:45:55

Call me naive but I trust Linus with the Linux kernel, hes a lot more laid back then the openBSD leader. One thing I have noticed is the kernels rapid development compared to BSD, personally I like the current pace of things. I use a modern desktop, and stability is nice, but if it means sacrificing the rapid speed at which the kernel adapts to newer and better things then I don't think i'd be content.

iphitus · 2007-08-21 12:21:32

pecan wrote:

Call me naive but I trust Linus with the Linux kernel, hes a lot more laid back then the openBSD leader. One thing I have noticed is the kernels rapid development compared to BSD, personally I like the current pace of things. I use a modern desktop, and stability is nice, but if it means sacrificing the rapid speed at which the kernel adapts to newer and better things then I don't think i'd be content.

I will. Believe it or not, there's multiple BSD's. FreeBSD is quite rapid. Dual booting it on my laptop currently.

Cerebral · 2007-08-21 13:09:26

PDExperiment626 wrote:

That being said, Linus seems to proclaim himself as a technological idealist in that the kernel should be above such financial interests. I.e. he makes decisions about what goes into the kernel and what doesn't based on merit and not necessarily financial benefit. I also understand what people do and say can be vastly different.

Even if Linus bases all his decisions on merit, if the only people submitting patches of merit are people paid by enterprises to write the patches, then the kernel itself will get an enterprise-y bent, regardless of whether Linus intends it or not.

raymano · 2007-08-21 17:20:03

I do agree that the stability of the kernel has decreased. Especially in the new kernel 2.6.22 where things that worked in the previous kernel are now broken. As the developer of FaunOS I can give you specific examples that have delayed our next release.

I do think that the Linux kernel developers sometimes tend to over analyze issues/designs and not necessarily give the user community what it wants but rather what the kernel team thinks the user community should have. I mean, why should there be so many Linux kernel patch sets? The reason is that Linux users, desktop users especially, are not getting everything they need from the standard Linux kernel.

This is not a technical issue. It has to do with how the kernel releases are managed and the process in which kernel decisions are made. There is only one Mac OSX kernel and it seems to work both for enterprise and end user applications just fine.

FeatherMonkey · 2007-08-21 17:37:12

I think you highlighted it, desktop Mac owners actually pay for the development of the kernel, if they suddenly began crying my kernel doesn't work it would be hell, heads would begin to roll. I suppose what I'm saying I don't think I'd be stretching it to say that the Mac enterprise in comparison to the Mac Desktop is even close to the relationship with Linux server to its desktop. I certainly would of thought there would be more Linux servers than Linux Desktop(Perhaps I'm wrong).

But as CK said the LKML isn't exactly the most friendly place, and the patches are coming from enterprise, why because the Desktop distro's aren't or can't sponsor the development.

Even if we all complained refused to use the linux kernel its not going to affect there wages, a lot of the Desktop community don't pay for it, unlike the Mac model.

Plus a kernel designed for specific hardware rather than designed for generic hardware not sure its an apples to apples comparison here.

Last edited by FeatherMonkey (2007-08-21 18:25:49)

chaosgeisterchen · 2007-08-21 19:45:17

I think that the main problem has been pointed out very clearly in this thread. It's the wrong persons actively developing the kernel, as they have different intentions than the default desktop user. Enterprise-centric development will make the kernel probably more stable but it won't necessarily increase desktop performance - it will more likely decrease it. Sacrificing speed for the sake of stability isn't the way to go and I think that it should be the target of Linux to increase both aspects.

The Mac-Model has attracted my attention for some while now. It's rather perfect, giving the user exactly what he wants - a system which 'just works', with no hassle in terms of kernel issues. It might not be the case that Macs are the fastest machines on the planet but I experience them as very pleasant to work on, because everything runs so incredibly smooth that I sometimes hope, Linux would run just as smooth.

But Linux has to serve a much larger scope of hardware and with it platforms, so its development is a lot more complex. I just hope, that Linus knows what he's doing and knows in which direction we have to go in order to let Linux become the best kernel on the planet (if it isn't already).

@stability:

I have experienced quite a lot of problems with certain drivers along several kernel releases. For instance, the acx module won't build under 2.6.22 and has been a hassle with almost every kernel release before. It's a problematic driver, but if the kernel development would be straightforward, these regressions should not have happened. In general, Linux bares an incredible possibility of breakages everywhere, because small changes can affect the whole enormously. This has to be taken into consideration when arguing about Kernel regressions.

After all, there are little developers, which make their living working on the kernel. If the number of full time desktop-centric kernel developers would increase quite a bit (say 100-120 of them), kernel development would look quite a lot different. But this is not (yet) the case, I'm afraid.

Ramses de Norre · 2007-08-21 19:48:57

raymano wrote:

[...]
I do think that the Linux kernel developers sometimes tend to over analyze issues/designs and not necessarily give the user community what it wants but rather what the kernel team thinks the user community should have. I mean, why should there be so many Linux kernel patch sets? The reason is that Linux users, desktop users especially, are not getting everything they need from the standard Linux kernel.
This is not a technical issue. It has to do with how the kernel releases are managed and the process in which kernel decisions are made. There is only one Mac OSX kernel and it seems to work both for enterprise and end user applications just fine.

It's not like the possibility to have patch sets is there for the OSX kernel... Maybe there would be a lot of patch sets for it too if it was open source...

Last edited by Ramses de Norre (2007-08-21 19:49:22)

PDExperiment626 · 2007-08-22 02:18:22

Comparing against the mac development model I think is a bit of a red herring here. A mac isn't just hardware or particular software, it's the whole unification of both software and hardware. Testing stability and performance on a fixed hardware platform is infinitely easier than doing what linux attempts to do. This is a strength and a weakness of macs in general. While their performance is good with their own software products... stability is horrible when trying to use software from MS. It's been a few years since using a mac; but I do not have fond memories of using them. Plus, I am more oriented toward the academic end of ideals; and the idea of moving away from open-source is not in alignment with those ideals.

This does beg the question: does the linux kernel need to split into two development branches? While this would be a developmental nightmare (splitting dev efforts), it may be where the kernel needs to go as the variation amoung the targeted hardware architectures of linux is growing too large to be handled effectively by one kernel development route. It's a real development conundrum because an exceptional server optimized OS will never be a good desktop OS; VMS is a good demonstration of this. Yes, there are *BSDs out there that do both well; but linux seems to be on a path to be the unequivocal best server OS. To really do this, linux will have to become more focused toward servers than any other OS out there and that means desktop performance has to suffer. Really, I think the linux kernel devs need to settle on where they want their kernel to go relative to the hardware spectrum that is currently present; it can't be all things to all architectures.

What I want is a straight answer from the kernel devs as to whether or not they are going to continue to push for their kernel to be server oriented. Right now, they aren't committing to the desktop/laptop hardware; but they haven't openly said it will never be their target platform either. I believe that the laptop/desktop market has plenty of financial equity to lure linux kernel devs... just as the server market does; but a decision needs to be made on where the development is going to go and that decision has to be made clear to the users.

At this stage, I have seen a lot of actions indicating that desktop/laptop stability is really a background concern for the underlying kernel development effort (see my previous posts in this thread). If that is indeed the case, linux will go down the same path as windows... pushing more and more features at the expense of stability... to the point the OS is no longer usable by the average user. It's not a matter of trusting Linus or other devs over Bill or anyone else; there are many reasons why windows went down the path that it did. I am not convinced that anyone is above making the same mistakes as MS did with their OSes; on the contrary, the beginnings of it have already happened with linux. It may be that in the end, I will have to use something like a Mac just because I know that there is a cohesive development between the hardware and software for that platform. We'll just have to see.

Last edited by PDExperiment626 (2007-08-22 02:24:16)

iphitus · 2007-08-22 10:15:32

chaosgeisterchen wrote:

@stability:
I have experienced quite a lot of problems with certain drivers along several kernel releases. For instance, the acx module won't build under 2.6.22 and has been a hassle with almost every kernel release before. It's a problematic driver, but if the kernel development would be straightforward, these regressions should not have happened. In general, Linux bares an incredible possibility of breakages everywhere, because small changes can affect the whole enormously. This has to be taken into consideration when arguing about Kernel regressions.
After all, there are little developers, which make their living working on the kernel. If the number of full time desktop-centric kernel developers would increase quite a bit (say 100-120 of them), kernel development would look quite a lot different. But this is not (yet) the case, I'm afraid.

tiacx is a poor example. It's an external module, uncontrolled by kernel developers. When i talk stability, I mean the core, vanilla kernel.

James

chaosgeisterchen · 2007-08-22 10:30:02

@iphitus:

Okay, agreed. tiacx is driving me nuts, though. Concerning the vanilla kernel, I have not yet encountered fatal problems which kept me from continuing working with it. So much from me, an average desktop user.

@PDExperiment626:

You come up with some interesting points and demand quite a lot from the kernel developers, but your demands are indeed appreciable. I think that you would be torn to shreds anyway upon posting stuf like that on the LKML, but it's evident, that kernel development follows a direction which many people do not agree on (without actively changing it or having the chance to change it). I could very well imagine a desktop and a server branch for kernel26 development (or going further, reorganizing some major aspects of the kernel and developing towards some 2.8 release?), both using the same foundation but differ in several aspects which make them optimized for what they are used for. Furthermore, I would appreciate special support for some architectures, especially enhanced support for x86_64 could push Linux even farther beyond what Windows makes possible.

This is the theory, porting it to practice would be the next step (and very comprehensive to do..).

It might keep you from having to purchase a mac.

shining · 2007-08-22 10:47:17

A little question : do people criticizing the kernel development here really know what they are talking about ?
(mostly about it being only orientated to enterprises)
Do you at least follow said kernel development closely ?
For now, that looks just like FUD to me.

FeatherMonkey · 2007-08-22 11:31:50

@PDExperiment626
Here's my tuppence, firstly you ask which way will it go, I think you know the answer better than I, you mentioned what Linus thinks. Which reading between the lines means he seems to want a Kernel that does both. I think I'd agree with what you said at the moment its perhaps a little of a master of none, jack of all trades.

Though I think that would be a harsh as it certainly seems to be doing a very good job at both, now we can hypothise what may or may not happen. But unless as chaosgeisterchen put it a change in the kernel devel team is taken then will it change, will this change with out sponsorship from lets say Ubuntu? Splitting is going to create 2 devel trees, does the team have the resources, how do you give them the resources?

To me this whole situation is far from clear cut, say enough of us make our voices heard, which I actually am not sure can be done. Can the devs afford to ignore there sponsors, is there a company on the desktop distro's who can replace that sponsorship?

Now unlike you I'm very much a pragmatic user, I use closed source video drivers and I'd feel a hypocrite to be condemning closed source whilst I do(Which I feel many OSS fans do, do). Next as much as we hate to say, most if not nearly all new users are ex-window users. Now it just works for them, they go into the shop they buy the latest hardware, the new printer, the new video card and it works. So many times you see a percentage fail, they don't grasp this isn't the case with linux.

Running linux takes management, choosing the right hardware device etc, not being scared when the odd boo boo comes through on the updates etc. Then you have so many what I call commercial apps that just dont have a linux equivalent, spreadsheet limits, no decent 3d cad, no photoshop, no decent PDF creator etc ... I can go on here hopefully you got the gist. So thats a very limited user base, we've been left with. Next even if you do manage to find the equivalents is there an eqivalent to something like access, so you now need to train people in another app or even several when there's ton's of windows people already who can do the job on Windows. Now to me linux has a far better chance of getting at business's when they start, but it takes management researching pinters, scanners, software etc..

So Business's lacking the resources, the manpower, or can't find the equivalent just aren't really going to increase the user base. Now perhaps the SMB's startups are increasing but we're not going to see a revolution.

Which leaves the real Desktop user the home user that do sit on windows, why because it works it plays there DRM content as the user we aren't in control. To have my hands tied because of someone elses principles does go against my Ideals of freedom. Take the blu-ray drives and I'm afraid perople do like keeping up with Mr Jones, why I guess thats one for the physcologists.

Now the difference with servers is it's less likely to need graphics etc ... why because you're more than likely going to be accessing it from a system that has graphics or may even just be console based.

Now to me the market we can really make a dent in is the household desktop, but whilst we condemn, have no solution for closed source will it happen, or will we see the possible converts fail because there new fangled hardware failed, they can't find the equivalent app say photoshop or non-linear video app .

Now I actually bet you can find many of the commercial apps I mentioned for the Mac. It may well go against yours and many others principles but a solution needs to be found so that the linux desktop can actually support closed, why because money makes good apps money buys time.

Now strangely enough you mentioned Launchpad, now i needed to do a little research firstly they have released a component this year. Next what about Bitkeeper then? http://www.linux.com/articles/44147?tid=2&tid=25&tid=3 From what I can see actually designed by a Kernel devel.

@ iphitus
Which leads me to you do you think its gone down since they changed revision controllers then? As from what I can work out going on when you implied the kernel started slipping would put it at about this time.

Honestly I think we need to accept closed source find a way the software/driver houses are happy with, get support for a higher percentage of off the shelf components, then and only then we can start making waves with OSS. Why because we'll actually have a user base that counts rather than being a minority and a hobby distro. Until this happens we'll be left with the guys that can sponsor it, who in the majority of cases are enterprise orientated,

Take the advances with Ubuntu, now you finally get a maybe convert they go to their favourite online TV channel http://www.bbc.co.uk/iplayerbeta/find_out_more.shtml now you see a mention of the Mac but where's the mention of linux, what happens they either dual boot(Because linux doesn't do it all) or go get the disc of friend and install XP. We can't force OSS principles on converts thats not why these people are trying linux, they are trying because they are sick of Blue parity screens, tasks clogging there system with background tasks but they still expect a similar service to Windows.

Linux is an alternative to windows yes but in many areas it still lets the common user down, hopefully this will change in the future. As pointed out there is a big difference between the Business model of closed source selling versus OSS support model, you can't go around to these big companies saying what you do isn't what we want. My self I want Adobe to support us, I want 3dcad to support us, I want a decent non-linear app and I don't care if its closed or open. So its closed but with every app the user base will have the potentiol to increase and eventually then we can preach OSS principles. At the moment we have no choice but to rely on OSS which does fail on many apps at the end of the day Money drives development.

@shining so you think post 11 is wrong why?

Last edited by FeatherMonkey (2007-08-22 11:32:53)

iphitus · 2007-08-22 12:48:42

FeatherMonkey wrote:

@PDExperiment626
Running linux takes management, choosing the right hardware device etc, not being scared when the odd boo boo comes through on the updates etc. Then you have so many what I call commercial apps that just dont have a linux equivalent, spreadsheet limits, no decent 3d cad, no photoshop, no decent PDF creator etc ... I can go on here hopefully you got the gist. So thats a very limited user base, we've been left with...

Openoffice spreadsheet covers most use. And OOo spits out perfect PDFs of what it can read out of the box. In addition, most applications can save to pdf in some form or another.

Which leaves the real Desktop user the home user that do sit on windows, why because it works it plays there DRM content as the user we aren't in control. To have my hands tied because of someone elses principles does go against my Ideals of freedom. Take the blu-ray drives and I'm afraid perople do like keeping up with Mr Jones, why I guess thats one for the physcologists.

bluray discs play on linux, takes some effort currently, but can be done. This shall improve with time of course. I don't see any urgency, i've never seen a hddvd/bluray disc in my life. But despite this, development is occurring in this area.

@ iphitus
Which leads me to you do you think its gone down since they changed revision controllers then? As from what I can work out going on when you implied the kernel started slipping would put it at about this time.

No, that's got nothing to do with it. You're just looking for a closed app to support your argument.

Honestly I think we need to accept closed source find a way the software/driver houses are happy with, get support for a higher percentage of off the shelf components, then and only then we can start making waves with OSS. Why because we'll actually have a user base that counts rather than being a minority and a hobby distro. Until this happens we'll be left with the guys that can sponsor it, who in the majority of cases are enterprise orientated,

There's nothing stopping Adobe from releasing photoshop. If anything, they're encouraged to release it, even if it is closed source. Only a foolish few would expect them to open source it.

Take the advances with Ubuntu, now you finally get a maybe convert they go to their favourite online TV channel http://www.bbc.co.uk/iplayerbeta/find_out_more.shtml now you see a mention of the Mac but where's the mention of linux, what happens they either dual boot(Because linux doesn't do it all) or go get the disc of friend and install XP.

Pretty hapless example. Most video channels do play on linux. And as for iPlayer, there's been a lot of action, and it's going to be fixed up iirc.

Linux is an alternative to windows yes but in many areas it still lets the common user down, hopefully this will change in the future. As pointed out there is a big difference between the Business model of closed source selling versus OSS support model, you can't go around to these big companies saying what you do isn't what we want. My self I want Adobe to support us, I want 3dcad to support us, I want a decent non-linear app and I don't care if its closed or open. So its closed but with every app the user base will have the potentiol to increase and eventually then we can preach OSS principles.

And there's nothing stopping them. Nothing whatsoever stopping them from releasing closed apps. Take the many closed games on Linux. We're not shunning those, we're not giving those the cold shoulder. We welcome those. The same would be afforded to Adobe or any other commercial app which is released on Linux. You're argument is based on nothing, and is also offtopic -- it's got nothing to do with the kernel now.

I think you're also impatient. Things happen with time. It's not like Linux has a short launch window to take off. People predicted the end and the doomsday for linux once Vista came out. Didn't happen. If these software giants are going to release their applications on linux, it's going to happen over a long period of time. There's no rush, and there's no "urgent need now or we're doomed".

James

Last edited by iphitus (2007-08-22 12:49:45)

shining · 2007-08-22 13:16:22

FeatherMonkey wrote:

@shining so you think post 11 is wrong why?

What I said doesn't invalidate post 11. The "if" is important.

FeatherMonkey · 2007-08-22 13:29:36

iplayer has no intention of doing anything, the providers dictate DRM. As do the other 5 channels I have access to in my country. And you are right I can play many streaming medias.

As for looking for I wasn't I was actually looking I didn't know, I was curious whether it might of been related not that it was closed, more the time frame.

I could add personally I don't see any problems but then as stated I'm far from knowledgable.

Many people complain about lacking features in a commercial environment on the spreadsheet I can't find it now, but its something about importing. Spitting out PDF isn't the same as devlopment I do agree there's plenty of alternatives for many apps but in comparison some leave the user feeling like they have had a bath with there clothes on.
As an example Just some complaints I can find.
try to mark text blocks in OO
no otf support at all
poor ps rendering
poor object manipulation

Perhaps I have gone OT that wasn't intentiol but what i'm trying to say perhaps badly is we need to make the Desktop User base increase and at the moment with the alternatives it is lacking. Though the way I've expressed it may be wrong. We can't have an input to the development of the kernel until there's finacial input coming from the desktop is all I was saying.

When you say only a foolish few I'm not so sure its a few. No I'm not impatient but to deny there's a hole in the roof when its not raining doesn't make it so. Whilst we are behind in hardware development, it'll stay a 2nd class citizen.

I never believed Vista was going to be the death of Linux but I am left feeling that it may be a hobby distro on the desktop.Take the uproar regarding Novell that was just pathetic, from what i can see they've done a lot for the community as whole but that all got washed away for one sin.

All along I've mantained this is a communty problem not a development probelm we need to change the community.

Edit
If every one of us gave a £ to the kernel developers for each kernel then who would they listen to.

Last edited by FeatherMonkey (2007-08-22 14:10:36)

chaosgeisterchen · 2007-08-22 14:26:51

shining wrote:

A little question : do people criticizing the kernel development here really know what they are talking about ?
(mostly about it being only orientated to enterprises)
Do you at least follow said kernel development closely ?
For now, that looks just like FUD to me.

I have subscribed to the LKML-digest mailing list but never came to read a whole message. So I have no real knowledge about how the kernel is working and I am not going to criticise the work of the developers, it's just that I would sometimes like to see another target towards which kernel development goes. Currently I do not see a main target they are approaching, it's simply replacing, updating and adding features and drivers, which is great but it might be interesting, how kernel development would look like if its target was to create something specific (e.g. a possibly slim kernel, dedicated to run x86_64 faster than everything before, focusing on desktop usage). But only money could drive them to follow ones ideas unless he/she contributes the code the person wants to be included by themselves (or create a patchset for the vanilla kernel).

Actually, one could fork the kernel anytime, take some devs and remodel it as they like, if all of them aren't satisfied with what Linux signs off.

And please do inform me, if kernel development really has a overall target they want to apporach.

Regards,
~cg

shining · 2007-08-22 14:49:05

from lkml, reported here : http://kerneltrap.org/node/14008

From: Linus Torvalds [email blocked]
Subject: Re: [ck] Re: Linus 2.6.23-rc1
Date: Sat, 28 Jul 2007 10:12:32 -0700 (PDT)

On Sat, 28 Jul 2007, Jonathan Jessup wrote:
>
> Linus, there is a complaint about the Linux kernel, this complaint is that
> the Linux kernel isn't giving priorities to desktop interactivity and
> experience. The response on osnews.com etc have shown that there is public
> demand for it too.
No, the response on osnews.com only shows that there are a lot of armchair
complainers around.
People are suggesting that you'd have a separate "desktop kernel". That's
insane. It also shows total ignorance of maintainership, and reality. And
I bet most of the people there haven't tested _either_ scheduler, they
just like making statements.
The fact is, I've _always_ considered the desktop to be the most important
part. And I suspect that that actually is true for most kernel developers,
because quite frankly, that's what 99% of them ends up using. If a kernel
developer uses Windows for his day-to-day work, I sure as hell wouldn't
want to have him developing Linux. That has nothing to do with anything
anti-windows: but the whole "eat your own dogfood" is a very fundamental
thing, and somebody who doesn't do that shouldn't be allowed to be even
_close_ to a compiler!
So the whole argument about how kernel developers think that the desktop
isn't important is totally made-up crap by Con, and then parrotted by
osnews and other places.
The fact is, most kernel developers realize that Linux is used in
different places, on different machines, and with different loads. You
cannot make _everybody_ happy, but you can try to do as good a job as
possible. And doing "as good a job as possible" very much includes not
focusing on any particular load.
And btw, "the desktop" isn't actually one single load. It's in fact a lot
of very different loads, and different people want different things. What
makes the desktop so interesting is in fact that it shows more varied
usage than any other niche - and no, 3D gaming isn't "it".

Arch Linux

#1 2007-08-20 01:34:08

Opinions on Linux Kernel

#2 2007-08-20 02:13:59

Re: Opinions on Linux Kernel

#3 2007-08-20 02:29:20

Re: Opinions on Linux Kernel

#4 2007-08-20 02:59:25

Re: Opinions on Linux Kernel

#5 2007-08-20 08:44:14

Re: Opinions on Linux Kernel

#6 2007-08-20 12:00:12

Re: Opinions on Linux Kernel

#7 2007-08-21 04:27:45

Re: Opinions on Linux Kernel

#8 2007-08-21 08:08:22

Re: Opinions on Linux Kernel

#9 2007-08-21 11:45:55

Re: Opinions on Linux Kernel

#10 2007-08-21 12:21:32

Re: Opinions on Linux Kernel

#11 2007-08-21 13:09:26

Re: Opinions on Linux Kernel

#12 2007-08-21 17:20:03

Re: Opinions on Linux Kernel

#13 2007-08-21 17:37:12

Re: Opinions on Linux Kernel

#14 2007-08-21 19:45:17

Re: Opinions on Linux Kernel

#15 2007-08-21 19:48:57

Re: Opinions on Linux Kernel

#16 2007-08-22 02:18:22

Re: Opinions on Linux Kernel

#17 2007-08-22 10:15:32

Re: Opinions on Linux Kernel

#18 2007-08-22 10:30:02

Re: Opinions on Linux Kernel

#19 2007-08-22 10:47:17

Re: Opinions on Linux Kernel

#20 2007-08-22 11:31:50

Re: Opinions on Linux Kernel

#21 2007-08-22 12:48:42

Re: Opinions on Linux Kernel

#22 2007-08-22 13:16:22

Re: Opinions on Linux Kernel

#23 2007-08-22 13:29:36

Re: Opinions on Linux Kernel

#24 2007-08-22 14:26:51

Re: Opinions on Linux Kernel

#25 2007-08-22 14:49:05

Re: Opinions on Linux Kernel

Board footer