“The fragmentation of Android is very real and very problematic for end users, developers, mobile operators, device manufacturers, and Google. However fragmentation does not mean Android is going to ‘die’ or ‘fail’ as some seem to think. On the contrary I think we can count on Android playing a significant role in our world for a long, long time. I also am confident that Google has already lost control of Android and has zero chance of regaining control. This post explains why I’m so confident about this.”
Android fanboys will say that fragmentation isn’t an issue because most users have no idea and/or don’t care what version of the OS they’re running. However, we’ll see what happens WHEN (not if) a huge vulnerability in Android is found that affects multiple versions of the OS. I promise you, if somebody’s phone gets p0wned because it’s running an old version of Android that doesn’t have the fix for said vulnerability because it was not backported by either Google or the carrier, they’re going to care when some hacker starts using their phone to make/send random phone calls/text messages to somewhere in Russia, or elsewhere.
p0wnage isn’t a bug in Android, its a feature :
http://viaforensics.com/security/nopermission-android-app-remote-sh…
So it’s as shoddy as OS X, then.
You mean like rage against the cage found on all pre 2.2 devices, aka 1 in 10 of Android devices.
I wouldn’t say that the problem you describe is really one of fragmentation, but one of carrier control of the software/updates. IE/ everyone could be on Android 4.02 or whatever is the latest, but if every carrier is in charge of updates you’re going to have problems getting the phones updated in a timely manner.
What is really needed is a unified update manager, simular to Windows update, ports/Yum/Debian repositories. You do that so every phone has a single way to update itself. If there is a problem you just need to tell everyone to hit menu/system/update or whatever. And if/when the carriers ( or ideally google or ASOP) would put the update there when it a was available for their phone.
Problem is, putting aside carriers and OEMs, Google themselves seem wishy-washy in enforcing this, even on their flagship Nexus handsets. Take the whole yakju/yakju-xx firmware debacle for example:
http://forum.xda-developers.com/showthread.php?t=1376856
I do agree with most of what you said, though.
Edited 2012-01-17 03:27 UTC
Hence, fragmentation.
And how/why do you suppose this would work, when vendors are tweaking the OS per carrier demands? Meaning, by the time vendors get done with it, you’re not dealing with the same base OS from device to device.
The article is linked to Kindel Systems.
Marketing genius here. Look at the adds on the site.
His main business mixed with addsense and Kindle support.
“Lots of people accidentally spell Kindle Kindel and come here by mistake.”
One thing I can’t help but notice, on desktops at least, is that the more unified the software solution, the more popular it tends to be. And the more divided the software solution, the more it goes trudging on for year after year after year after year, holding its own, but not making a dent in the status quo. Windows absolutely rules the desktop. Linux sits at < 0.5% – 0.75%… for over a decade now.
Apple may seem to buck the trend. But another degree of freedom comes becomes important in that case. They hold both software and hardware fixed. Solaris/Sun is probably a good example of how you can get a leg up that way. Albeit in the server market.
Windows leverages the appeal of an integrated desktop solution sliding along on the coattails of Intel’s hardware dominance.
Mobile devices look kinda like the desktop model. And Android looks a bit like the Linux model. Somewhat better on the hardware end, maybe.
One can make an argument that fragmentation *can* be good in this or that case. But it’s awfully hard to make an argument that pointless fragmentation isn’t bad.
Android can’t afford to allow pointless fragmentation.
When I start seeing articles explaining how fragmentation of a particular platform isn’t really bad, it is difficult for me to interpret it as good news for the platform.
When you say Windows rules the desktop, do you mean Windows 8, Windows To Go, Windows 7 Starter, Windows 7 Home Basic, Windows 7 Home Premium, Windows 7 Professional, Windows 7 Enterprise, Windows 7 Ulitimate, Windows Home Server, Windows Fundamentals for Legacy PCs, Windows Vista Starter, Windows Vista Home Basic, Windows Vista Home Premium, Windows Vista Business, Windows Vista Enterprise, Windows Vista Ultimate, Windows XP Starter Edition, Windows XP Home Edition, Windows XP Home Edition N, Windows XP Professional, Windows XP Professional N, Windows XP x64 Edition or Windows XP Media Center Edition?
Or have I spotted a flaw in your argument?
I’m sure you feel you’ve been very clever. But I mean “Windows”. That family of desktop OSes which maintains enough coherency that both users and vendors consider it to be one OS, more or less. Even over multiple generations of that OS family.
I’m only allotted 8000 characters, or whatever, per post in this forum. So I’ll just provide a link to the 317 Linux distros that Distrowatch maintains statistics on:
http://distrowatch.com/dwres.php?resource=popularity
Now, that could be construed as being a bit unfair of me since many of those distros are built upon the foundations of other of those distros. There are not *that* many core distro families. But there is more incompatibility between the current incarnations of those core distros than you see between contemporary versions of Windows.
What is much worse is that even if you pick one distro… 6 months down the road you are going to encounter incompatibilities that MS would consider unacceptable in their products, delivered by Linux vendors with the advice “You deal with it, users”.[1]
This philosophy is disseminated from the kernel devs, who refuse to implement a stable driver interface, at the top. And is picked up and emulated by the intervening layers which lie between the ethereal levels of kernel development, and the grunt levels of Linux administration and home usership.
And in another 6 months, we’ll get another round of that. And in another… well… you should be getting the picture.
Look. I dislike Windows as much as anyone. I don’t allow it in my home. (Nor MacOS X, for other reasons.) But while I’ll go so far as to point out the facts, I’m not going to stick my head into the sand to have a prolonged conversation with you. You’ll need to come up for air and agree to have a civilised conversation in the open, with all the obvious facts gathered ’round us so closely that they can’t be denied.
Hey, I’d like to deny them, too.
Anyway… no, I don’t think you’ve discovered a flaw in my view. I think you’ve trotted out a meaningless cliche, popular among advocates of non-Windows OSes, that unfortunately, doesn’t really bear close inspection.
-Steve
[1] Exceptions to this are the enterprise versions of Linux. RHEL/CentOS/Scientific Linux. I’m recommending these to people, increasingly. Some teeth-jarring is still involved. But only every 2 to 4 years. And you don’t have to encounter any for 7 if you don’t care to.
Debian, your argument is invalid.
You’ll need to specify what you mean. Debian is as bad as any of the Linux distros. A sometimes ~2 year, but very uncertain 2 – 4 year release cycle. Forced upgrade a year after… whenever it is they decide to release the next version. Teeth-jarring incompatibilities at that time. No option to go to 7 years on one release, as with RHEL/CentOS/SL.
OK. So maybe it’s better than Fedora. But if you think Debian’s existence somehow conflicts with my stated argument, I’d recommend a splash of cold water to you.
-Steve
Edited 2012-01-17 04:56 UTC
Distrowatch is a geek site and is not indicative of general Linux usage.
There are only four basic distros (Debian, Red Hat, Slackware and Gentoo). Every other distro is a fork or spin of these.
There are only two mainstream Linux distros Debian-based (Debian/Ubuntu/Mint) and Red Hat-based (Fedora/Suse etc) which realistically cover well over 95% (probably >99%) of desktop users.
There have been large educational and corporate rollouts involving tens of thousands of individual machines rollouts. AFAIK every one of these rollouts has used some type of Debian or Red Hat base.
To the non-technical end user the Linux experience on any mainstream distro is virtually identical. KDE and Gnome offer a user experience similar to any other modern desktop OS.
True, but irrelevant to the current context.
The real problem, as I described, is not so much the incompatibilities between contemporary versions of distros in any 6 month period. (Although that is still a very significant problem.) But the ongoing, pointless, reckless incompatibilities between Distro X Ver. A and Distro X Ver A+1. Every project involved thinks they are allotted a major incompatibility.
Linux is still a developers playground. Unsafe for the average desktop user. With a dedicated and savvy sysadmin, who’s given a sufficient supply of anti-ulcer pills, it’s great.
Until they try to upgrade by themselves. Remember, they’ve gotten and applied a lot of advice about getting their systems to work right with the software they want. (Often not considered, or even ethically approved of by their distro’s devs.) Pointing to various repositories, installing various codecs, and making various tweeks to work around problems.
And then… when they go to upgrade… boom!
If I organize a help desk for non-technical end users having these “virtually identical experiences” would you be willing to man the phones for free?
Thought not.
Edited 2012-01-17 05:25 UTC
You obviously have no real experience of Linux to make such absurd claims.
The average user can use any mainstream Linux distro without the slightest problem.
You are either a bullshit artist or a total moron. If a newbie sticks to Mint, Ubuntu or Mandriva their life is drama free. No sane person recommends Slackware, Suse or Fedora to a novice (however they are extremely easy to use once configured).
Ubuntu has a software store that is effortless to use. Alternatively you only need to tick Ubuntu Restricted Extras and all codecs, flash, MS fonts etc are automatically installed and configured.
Mint comes fully configured out of the box.
In comparison a clean Windows doesn’t do anything useful unless you have no ambitions beyond using Notepad and Internet Explorer
Upgrades of any mainstream Linux distro go without a hitch. If you have a separate user partition you can even easily switch between distros and still retain all your personal files and settings.
The average user has no real need to ever upgrade their distro. If it ain’t broke don’t fix it. A Linux distro is perfectly usable long after the support period ends.
If they use Debian they can upgrade their system via rolling releases for as long as their hard drive works.
Most Windows users never upgrade their OS. Most of them wouldn’t be able to do so anyway. I have met many long time Windows users who have absolutely no idea how to perform the most basic tasks such as a defrag or instal a programme.
I would much rather support a variety of Linux distros than a a bunch of average windows users. For start I wouldn’t be dealing with all the crapware found on the most Windows machines.
Well, let’s see. I’ve been a Unix admin since 1988. A Linux admin since ~1996. And currently admin ~100 desktop Linux users. Linux is the only OS I’ve used on my own desktop since 1997, aside from the occasional jaunt with Open Solaris or some *BSD variant.
Aside from your grandmother’s and uncle’s computers, what Linux desktops do you administer?
I have a wealth of experience which I’d love to share with you. The situations I deal with would likely crush your spirit. But I persevere, and have succeeded, on the balance, with Linux desktops, despite its intrinsic problems. And despite having been spat upon, regularly, by two bit Linux pollyannas who don’t think I’m pure enough.
Go ahead. Give me your stats.
Edited 2012-01-17 07:19 UTC
Don’t burst into tears Princess.
You claim to be a Linux admin and then try and scare everyone with highly implausible nightmare scenarios. Do you use Fedora daily builds for production environments?
I’m not saying Linux is perfect but most of your comments are pure BS.
Fact: It only takes 30-60 minutes to get a fully functional desktop system up and running with any of the mainstream Linux distros. In contrast a new Windows install is basically useless.
Fact: It is extremely easy to upgrade or even totally replace a distro as long as you use a separate home partition.
Fact: It takes hundreds of hours to have even a reasonable understanding how an OS works “under the hood”. [It takes thousands of hours to be an expert]. Linux is very different to Windows. If you allow an untrained Linux user administration privileges they are guaranteed to fsck their install. Ignorance is not the fault of the OS.
Fact: It is not necessary to upgrade to the latest and greatest version of a distro. In reality the hard drive in a home PC will normally fail before the distro support period ends.
Fact: Enterprise level distros eg RHEL have extremely long support cycles of 7-10 years (as long as Windows).
Fact: Linux provides the source code so you can patch or update any code even after the support period ends. [Try doing that with Windows].
Fact: You can’t expect a a purist FOSS (Debian) home consumer (non-LTS Ubuntu) or bleeding edge (Fedora) distro to be used in a production environment without some hassles. That isn’t their intended role.
Who are you trying to fool? The year of the Linux desktop never arrived.
– A LOT of hardware have issues under Linux, or require drivers that are difficult to install. Not to mention never tailored to whatever distro the user is running.
– Distribution fragmentation is one thing, but what about desktop fragmentation? GNOME 3.x, Unity, KDE 4.x, XFCE… It’s all a mess.
– You claim there is nothing you can do on a default Windows installation. Well, at least you can install flash/java/multimedia codecs painlessly by answering yes when going to a site. I mean, Synaptic is great for technical users but grandma can forget about it.
– None of your “facts” about how easy it is to use Linux applies for normal users. You are looking at it through your own eyes, you fail to see Linux from a non -techincal users point of view.
– When faced by superior arguments from sbergman27 you react by claiming that he is not a Linux user? Just… Wow.
Mr. Bergman isn’t making any superior arguments. He is mixing and matching use scenarios with different target audience distro’s and painting a far bleaker picture of Linux than is actually the case.
I for one don’t care how many enterprise desktops Mr. Bergman administers. He might administer 100, 1,000 or 1,000,000. That is a large scale deployment with specific, specialised use cases for the enterprise.
The wonderful horror stories of fragmentation from an end user perspective simply doesn’t exist. Be it a Linux distro or Android version X. If an end user has version X, he has version X.
Non-tech, consumer end users work, live and breathe version X and they will continue to do so until either the software they want to install from the friendly software center doesn’t install anymore or a techie tells them they need version X+1.
All the horrible things that can go wrong with dist-upgrade, do go wrong on a daily basis. I just wonder why ordinairy, everyday people opt to go the dist-upgrade route in a consumer setting. Just put that damn /home directory on that separate partition, blow away the .directories who’s settings you don’t want to keep and do a clean install in about 30 minutes and be done with it.
Yes, there is an awful lot of hardware out there that doesn’t play nice with Linux. The big question in my mind is always why people keep bothering to torture themselves with non working hardware under Linux for days on end, when a wonderfully working replacement part usually can be had for scraps. If you can’t afford to replace a small piece of hardware, you simply can’t afford computing.
Do some research before plonking down money and taking home a shiney piece of excrement that doesn’t give you any Linux pleasure. My machines and the ones I administer for my family simply don’t give any grief hardware-wise. Then again, they only receive supported hardware.
While appealing to authority and belittling all two bit Linux pollyannas must feel good, it doesn’t help a prospective new Linux user. We do after all help some of the grandmother’s and uncle’s with their computers. Does Mr. Bergman do the same?
This just coming from a two bit Linux pollyanna, who simply doesn’t see the big horror story for home users in the non-enterprise world.
Good. Because the Linux desktop market has been stuck at 0.5% to 0.75% for over a decade:
http://gs.statcounter.com/#os-ww-monthly-200807-201201
Despite all your hand-waving about how easy it is, it’s never gotten past square one. Especially embarrassing for a free (as in beer) set of offerings. This strongly suggests that there are some problems. Among them are the fragmentation issues I’ve mentioned. (Horizontally, between extant distros, and temporally, between successive versions.) A disregard of the user experience. (Something else is always “even more important” in key areas.) And a stubborn refusal by members of the perpetually tiny Linux Desktop community to even admit that there is a problem.
Sometimes I don’t know why I bother.
I’ve noticed this as well. 10 years ago, in most cases, we used to be able to just pick a distro of choice and install third party applications without too many issues. These days you need to do your homework and match the distro and version to where the application has been developed and tested on or you’re certain to bump into library incompatibities which make you end up wasting a lot of time trying to fix. SDK stability seems to be something completely alien to the Linux developer crowd.
But the ongoing, pointless, reckless incompatibilities between Distro X Ver. A and Distro X Ver A+1.
Until they try to upgrade by themselves. Remember, they’ve gotten and applied a lot of advice about getting their systems to work right with the software they want. (Often not considered, or even ethically approved of by their distro’s devs.) Pointing to various repositories, installing various codecs, and making various tweeks to work around problems.
And then… when they go to upgrade… boom!
Why should a Linux distro do any better with version A to A+1 than, for example, a heavily augmented (read third party applications and tools) Windows XP to Windows Vista? Dist-upgrade is replacing an entire OS, not just a “simple” update of the existing one. Which non-tech oriented end user is able to replace any OS? For these circumsances, LTS versions were invented. 3 years of stability and no dist-upgrade head aches. This only poses a problem when someone has updateritis. (For non-Debian/Ubuntu/Mint/etc. distros, replace the terminology with the appropriate equivalent.)
I have three desktop users to keep happy and dist-upgrade is expressly forbidden. I handle the OS replacements, they can do the update manager within the same version. I’m guessing this scenario is the same for every non-techie, for most available OSes. They can click on the in-version update icon, but replacing the OS is simply out of their league.
[ snip ] meaningless cliche, popular among advocates of non-Windows OSes, that unfortunately, doesn’t really bear close inspection.
I don’t think that cliche is as meaningless as you make it out to be. Win32 is the most stable API in existence, but MS has and does muck about with their driver model every few years. (That Windows XP was as long lived as it was, is just an abberation.) I’ve seen quite a few pieces of hardware “die” along the way (from Win 9x to Windows 7), just because the driver model in Windows had changed and the device vendor opted not to port the driver for economical reasons. While most programs (except for low level system tools) seem to survive quite nicely between versions, it isn’t all milk and honey in the Windows camp either.
Doesn’t exonerate Linux (distros) for sometimes seemingly needless churn. But your demands for 5 + years stability is a bit unrealistic. Not even MS has that in the cards for ordinairy mortals.
Consumer, Hardware, and Multimedia products
Microsoft will offer Mainstream Support for either a minimum of 5 years from the date of a product’s general availability, or for 2 years after the successor product (N+1) is released, whichever is longer. Extended Support is not offered for Consumer, Hardware, and Multimedia products. http://support.microsoft.com/gp/lifepolicy
The devil here is the 2 years after the successor product (N+1) is released. MS is ramping up development speed again, after the Longhorn disaster. So you’ll get 5 years at best and most likely 4 years at worst. This is for consumer versions of course. They do treat businesses better.
I’d add this got even wore after 64bit switch. Linux is now arms and legs above Windows in support for 3-5 years old hw on 64bit systems.
Edited 2012-01-17 16:59 UTC
The point I’m trying – and obviously failing – to make is nothing to do with Linux on the desktop.
You’re right to say that Windows rules the desktop. But just counting XP, Vista and 7 installs (that make up over 99% of all Windows installs) gives us, I reckon, more than 30 different versions of Windows.
Windows is far more fragmented than Android.
NetApplications, the most famous source for the <1% Linux desktop usage claims, reported Linux desktop use grew by 50% over the last 6 months to 1.41%. Their competitors reported a similar surge.
You need to update your argument for 2012.