Better or faster?
Are the linux developers getting better or faster?
One is an indicator of QUALITY the other of QUANTITY.
The Linux Foundation, which is something akin to the marketing arm of the open source operating system kernel and its related systems software, has today released its second report detailing how the Linux 2.6 kernel is evolving. The report reveals how it is coding the changes in the kernel and what companies are sponsoring the …
Has MS the breath to compete with this in the long run? Seems not.
They are in danger now to completely lose their smartphone OS market share to others, including Linux based solutions.
Windows on servers is as weak as it ever was.
Their final holdout, the desktop, is still going strong, but they are losing that, just very very slowly..
If they do not change their business model, MS's core business will end up to be where Linux can not go. The DRM world of consoles.
That is... if DRM won't completely die somewhere along the way....
"The number of lines of code added to the kernel has tripled, with the latest release, Linux 2.6.30, having over 10,000 patches."
You say that like it's something to be proud of.
"Since Linux 2.6.24, a stunning 10,923 lines of code are added every day, with 5,547 lines removed and 2,243 lines changed."
Er, we're still talking about the KERNAL here, right? ~11,000 lines of code added every day?
I hereby elect Linux as the most bloated, in-efficient load of shit ever produced.
ITS A KERNAL. It should be about 50K tops.
WTF? Seriously, WTF???
Ok, I maybe missing something here, but if we're talking about the KERNAL, as I understand the word, then there is something seriously amiss. Remember, the kernal is NOT the GUI, browsers, 4.1 surround sound audio drivers, media players and all that other crap, it's just the underlying resources that everything else is built up on: Memory management, task scheduling and management, inter-process communication, file system, device driver hooks (not device driving) and *possibly* TCP/IP (though it really shouldn't be in the Kernal, it's just a protocol)
I therefore fail to see what it is that is changing so rapidly? Why the constant state of flux? What's going on?
Is it simply "If we don't keep changing stuff, we're no longer needed, so no paycheck"?
If not, then I don't know what it is. I'm just not getting it, am I?
Sorry.
Tens of thousands of patches applied. A gazzillion lines of code. Thousands of coders at hundreds of companies, etc, etc, etc.
I love all of those stats. This is huge. I get that.
But where is the 'roadmap'? What are these people *doing*, and why are they doing it? What is slated to be available in release N+1, N+2, N+3, etc. What are the working on, and when will it be available?
Yes, there is clearly a lot of horsepower. And progress.
But who is driving, and - most importantly - where are they going?
Every other OS publishes a roadmap.
Windows on servers is as weak as it ever was.
Care to back that up with some facts? Last I read most reports say Windows are around 65% Linux around 15% (dependant on the report either can be higher or lower). Some reports saying Linux uptake is slowing in this segment as many shops have moved all they wan't to move to it. I still think it will make inroads, but take over? Not for a very long time.
Their final holdout, the desktop, is still going strong, but they are losing that, just very very slowly..
Yup it fell by 1% during a year when most Windows buyers are holding of for Windows 7. Would you buy a new product when you know a better one is just around the corner for the same price?
Linux is great, but please back up your "facts".
We'd need to decide on the number of grammes per square metre of the El Reg standard piece of A4 first. We probably need a standard printer too, variations in ink / toner use could make a significant contribution when weighing the value of large code drops.
There's an important problem though. Taking 66lpp as a standard, it ain't going to be long before someone works out that the most flattering module size is 67 lines.......
"Windows on servers is as weak as it ever was"
I don't know where you work, what you read, or where you get your opinions from, but Windows always has been (and still is) pretty strong on the server front. You're obviously an anti-MS idiot, because you've read somewhere it's the thing to be, so feel free to delude yourself into thinking Microsoft's server market is weak.
Granted, Linux is a significant player in this area (as opposed to the desktop arena), but to dismiss Microsoft as always being a bit player is ridiculous, and shows you don't actually work in the real world.
Bytes of executable code is even easier to track and less prone to distortion by things like long variable names and large block comments. It also has the advantage that even end-users have heard of "bloat" and may treat it with the contempt it deserves when it is used as a productivity metric. (Oh, you're a /manager/. Well you might not see that as an advantage then.)
Real productivity metrics would be things like "features" or "use cases" or "bugs fixed", but (as already noted) these are rather harder to even define, let alone measure. For one thing, they are mostly subjective, reflecting the (to me, obvious) fact that reasonable people may disagree over whether some changes are actually desirable.
What this article is measuring is "churn", which fairly accurately measures the amount of popular support for a project. That's an interesting sociological statistic, which makes it useful to anyone whose line of business is selling an audience of bored techies to hungry advertisers.
This post has been deleted by its author
>> Care to back that up with some facts? Last I read most reports say Windows are around 65% Linux around 15% (dependant on the report either can be higher or lower). Some reports saying Linux uptake is slowing in this segment as many shops have moved all they wan't to move to it. I still think it will make inroads, but take over? Not for a very long time.
...
>> Linux is great, but please back up your "facts".
Nice backing up of your facts there. I'm not familiar with the journals "Most Reports" and "Some Reports" - tell me, how do I go about getting a subscription.
Whilst I your figures are as completely 'plucked out of the air' as the original authors, I think they are probably more correct. I have certainly seen figures from 4 years ago that put percentages of "Enterprise Servers" at 65% and 15% for Window and Linux respectively (http://blogs.zdnet.com/ITFacts/?p=8708) - however not all servers are "Enterprise Servers".
Certainly there are areas where I would be surprised to find that Linux wasn't the most prevalant OS - e.g. Internet servers (web/email/ftp/dns/whatever). That said, I am sure there are more file/authentication/print servers facing inward on company networks than there are Internet servers facing outward.
>>Considering that the number of patches has exploded from 3,616 with Linux 2.6.11 to a high of 12,243 with Linux 2.6.25 and has averaged 10,656 in the five releases since then, it is clear that the Linux developers are getting better at what they do.
Wait, what?
The number of patches 'exploding' surely (Don't call me Shirley) means more bugs are being discovered in the real world and being fixed. If the developers were getting better then there would be less bugs being released and less patches.
I would assume that 3 or 4 times the number of patches clearly indicates a lowering of quality of released code. Unless the patches are all enhancements not bug fixes.
(This is a criticism of the rather slanted article, not the developers – increasing competition forces companies to increase quality, I don’t really use Linux but shudder to think what Windows would be like without at least some competition!)
It's you - it's call an monialithic kernel (i.e drivers are part of the kernel).
Though most us only tend to use certain drivers, we don't have a kernel with all of them enabled - we tend just to use the ones we need and bearing in mind that they are available for so many weird bits of hardware as well as the normal common ones.
I hope this clarifies.. I'm not sure if I'm feeding a troll so that why I choose a grenade.
Mike
Ok, I'll bite...
For one thing, it's "kernel", not "kernal", ffs.
Your definition is something that most of us would class as a microkernel. The Linux wee beasty contains all the source code for all the open-source device drivers, for all the devices supported - which is an awful lot. It also contains all the source code for all the file systems that Linux supports - again an awful lot.
It also contains all the source code to support all the architectures that the kernel runs on.
I could go on...
Cheers
Mark
We have 10 Windows servers and 35 Linux servers i.e. 12% Windows 78% Linux.
Our purchases of Windows licenses were recorded. Our CentOS installations from templates (which must be a really common way to roll out Linux servers) are not recorded.
All market share surveys I've seen would count companies like us as 100% Windows shops, but it couldn't be further from reality.
60% of web servers run Linux. 40% run Windows. It's very easy to determine web server market share by OS reliably, because web server software identifies itself. Source: Steve Ballmer http://www.pcworld.com/businesscenter/article/151568/ballmer_still_searching_for_an_answer_to_google.html
The question is whether web server OS market share is indicative of non-web server market share (the OS running on servers which are not directly on the Internet). In our case it is indicative.
You do realise that most general server OS market share "reports" by the likes of IDC are based on Linux Vs. Windows server *revenues*? Can you spot a problem there?
Regarding the suitability for Linux versus Windows on servers. Does Microsoft compete on performance anymore? Name a benchmark that shows Windows is a better choice.
The only thing Microsoft is competing on is "familiarity". And that is shaky ground.
Since the apparent metric of deployment is subjective experience, I'll just use my company as the standard for comparison--roughly 90% Windows, 9% Solaris, and 1% other (usually some sort of vendor-provided Linux). Since this is just as "scientific" as any other method used in these comments, Windows apparently has roughly 100 times the deployment of Linux!
I'll bite too. Post with your name, no need to be an AC. No one can fault you for having an opinion.
>"The number of lines of code added to the kernel has tripled, with the latest release, Linux 2.6.30, having over 10,000 patches."
>You say that like it's something to be proud of.
It is.
>"Since Linux 2.6.24, a stunning 10,923 lines of code are added every day, with 5,547 lines removed and 2,243 lines changed."
>
>Er, we're still talking about the KERNAL here, right? ~11,000 lines of code added every day?
It kernel - it doesn't shout either.
>I hereby elect Linux as the most bloated, in-efficient load of shit ever produced.
Your choice.
>ITS A KERNAL. It should be about 50K tops.
ntoskernel is somewhat larger than 50k as well. It's unlikely that the kernel in your mobile 'phone is 50k, although I don't really know.
>WTF? Seriously, WTF???
What the fuck. Go on spell it out - it reads so much better. Go easy on the question marks though - they just look silly.
>Ok, I maybe missing something here, but if we're talking about the KERNAL, as I understand the word, then there is something seriously amiss.
Yes, you are. You are probably thinking of microkernels of which the only one I know of is the GNU Mach thing. The Linux kernel consists of all the basic drivers as well as the thing you are thinking of. For example there are many file system types accessible, such as EXT{2,3,4}, NTFS, FAT, XFS, JFS, ReiserFS, MINIX and many many more. You don't have to compile support in for all of these - just those that you require. Also, the kernel is modular. This means that the drivers are hived off into little files and loaded when necessary. They can also be unloaded when they are no longer needed, saving on memory.
My laptop's current 2.6.30-r4 kernel weighs in at a 2.3 Mb disc image. There are 81 modules loaded in memory at the moment but I've unloaded the firewall modules because I don't need them at home really (although the wife's laptop might go rotten sometime!)
Your sub 50k kernel based system - could you tell me what it is please?
I don't know what prompted your diatribe but it seems strange that you seem to feel threatened by a shining example of what Open Source is all about. I think that the ideas behind the GPL etc are up there with Magna Carta or any Constitution designed to give Rights to people. The GNU/Linux kernel development shows that people from all over the world, from many different faiths, skin colours and so on, can co-operate on a massive project and deliver extremely high quality code. Then there are the KDE, Gnome and countless other Open Source project developers beavering away on stuff that they effectively give away for anyone to use. If you use Firefox for example then you are a recipient of the Open Source effort. It's operating system agnostic.
Come on then - reply, in your own name. You're called AC for a reason.
io_uring
is getting more capable, and PREEMPT_RT is going mainstream