Novell’s Nat Friedman told InternetNews.com: “The basic concept here is that the standalone operating system is dead.” Friedman is Novell’s Chief Technology Officer. He added: “The days in which people buy operating systems on their own and then build a stack from there […] will look like home-built automobiles in the future – people aren’t going to do this anymore.” This is not the first time some big company predicts the end of the traditional operating system.Novell entered the appliance space yesterday, with the launch of an appliance-friendly version of SUSE Linux. Such an appliance-friendly operating system has been pre-configured for the application that will run on top of it – parts of the operating system that are not relevant to the application it is supposed to run are not included. The JeOS Novell will be launching is based on SLES 10 SP2, and will only include the parts necessary to allow it to boot; you can then load up the application, make a snapshot of the image, and distribute it as a VMware or Xen image. Friedman: “I meant what I said: The standalone operating system is dead and we really see the future of the server operating system as being embedded in appliances, so you’ll see more from us on this topic.”
For some reason, this reminds me a lot of the hype surrounding the Network Computer, a concept backed by Oracle, Sun, and Acorn back in the second half of the 1990s. A Network Computer was more or less a diskless node, following the Network Computer Reference Profile, which used network booting to pull its operating system from a server. Since they contained less hardware, they were supposed to be easier to maintain and cheaper to buy.
Oracle’s CEO Larry Ellison had high hopes for the initiative, but it failed to live up to the expectations. PCs were coming down in price during those days, which meant that one of the advantages of the Network Computer slowly but surely eroded away. In addition, the internet was still too slow during those days (dialup, mostly), and as such, could not sustain the speeds needed to network boot operating systems and applications. These days, you can pick up diskless nodes and thin clients for relatively low prices, and they make good X terminals.
I am usually quite weary of claims about some concept’s demise or death. They are usually made by people who actually have an interest in that concept’s death, not by independent people. In any case, I’m not selling my operating systems just yet.
hmmm … this reminds me of some other companys so called “Focus shift” … well, we all know *that* ended very well, maybe this will end up just as big a success …
Think it is, as was said there has been a lot of this type of stuff said none lived up to it. I remember Gates saying we wouldn’t need more than a certain amount of memory, 540k I think, and here we are with memory in the gigabytes.
The limit they had was 640 kB. But yes. There are videoclips on youtube showing people firing up a mac classic, launching claris works, type some stuff, save it and turn of the machine (I think it included that part aswell), and then compared to a new machine, and the new machine are SLOWER. Sure the machine are like hundreds of times faster, but the software have become so much bloated that you don’t get more stuff done anyway. And it will probably remain that way, coders like to add whatever fun stuff they can because the system allows it anyway. It’s not like we NEED bouncing icons in the dock, or 3d effects when switching windows, but the hardware are there so someone will make use of it.
Coders like to add? I see you haven’t had the pleasure to work in a corporate setting:)
hardware ressources
———————————- = overall usage speed
application requirements
Please have a look at this comment for corresponding annotations:
http://osnews.com/permalink?309755
I think this tendency at least rectifies a niche market for users who don’t want to upgrade their hardware day by day just to keep (!) the overall usage speed they are comfortable with.
That, in part, is because ClarisWorks was very sweet.
I still have a SE/30 around to use it…
You know nobody can ever provide proof that he said that. Futhermore he is recorded saying we will eventually need to move to a 64bit memory system long before this phrase was atributed to him.
I’ve seen him say it in a video.
Edited 2008-04-18 04:00 UTC
“The 640kb should be enough for anyone” quote is just an internet myth. He never actually said that. And in his defense, why would he say such a potentially unprofitable thing anyway. 🙂
Umm, maybe because that was the limit of the operating system he was selling, and he was not selling hardware?
I really doubt we’ll get rid of regular desktops and oses for as long as the web are the only alternative. Seriously web apps sucks, even gmail will be and are worse than a good e-mail client, and well, then what more is there to get? I guess web apps could work ok for a calendar to, but preferably not even for a word processor, even less for something like adobe lightroom or imovie or reason or …
The web wasn’t made for this. People (still) try to design their webpages as if it was a medium designed to be pixel perfect and for exact and accurate design which showed up the same everywhere, guess what? It’s not. The information is what matters and if anything imho it’s the browsers work to present it in a decent way. If people didn’t expect browsers to work like images in photoshop we would have way less problems.
Edited 2008-04-17 22:54 UTC
Yeah, I think you’re right that the desktop OS isn’t going to die soon. On the other hand, he seems to be focusing very heavily on enterprise uses, which are often filled with hundreds of drones using a single application (which are increasingly web based).
Appliances are popular in business, and that is what Novell cares about. Just like RedHat, who earlier said they aren’t very interested in a consumer OS.
Keep in mind that web apps are slowly becoming less browser-based. I think a lot of Office apps are going to be handled perfectly by web apps in the future.
As you said, desktops will be left for the apps that actually require it. Fast CPUs, lots of memory and fast disks are for photo/video editing, gaming, etc. Even then, you get better performance on systems (appliances?) that are solely meant for a single task, like your xbox or ps3.
how can the desktop die? ok, people would rather play games in front of their tv because the screen is bigger. But you won’t want to read text on them and that’s the case with smartphones and pdas as well because the screen is too small.
The best mobile device and gaming console will be the one that recognizes the need to work with the pc instead of trying to knock it off.
Funny, as many will say that an xbox or a ps3 is far more similar – both hardware and system software – to a gaming desktop than a PDA appliance.
After all, an xbox run a stripped down Windows 2000 kernel tuned for best game and multimedia experience, while desktop PC still don’t run up-scaled appliances operating systems.
the standalone operating system is dead.
And a strategy regarding Linux to match this concept.
I thought we’d seen the last of these grand and rather silly protestations from Provo Towers. Or should that be Waltham? They’ve moved so many times, I forget.
Errr, yep. VMware and virtual machine appliances, rPath and all that. Been around for donkey’s years.
The problem with these machines is how flexible they are when you go about deploying them in your own environment. There are always subtle changes to be made, and if one ISV standardises on one OS and machine and another ISV on another then you have too much proliferation of distributions going on. For example, suppose the ISV embeds its own database in the machine and you want it somewhere else for backup and consolidation reasons? Having to run around making your own tweaks kind of negates the advantages and time savings, unless you’re creating your own internal appliances ready to drop-in to your own environment.
Errr, and what exactly is this enterprise operating system? Judging from the recent past it isn’t anything Novell is producing, and with Novell’s actions recently, that OS is basically Windows.
I’m not entirely sure where that is, because Novell has been taking the Netware ecosystem and is dismantling it, with no enterprise network OS replacement to be seen. I’d start looking after those customers first rather than chasing after phantom ones in a market you’re not in and where you have no clear technical direction.
Welcome to 2008 Nat. I, and many other people who’ve been using virtual machine environments like VMware, have been doing this for years.
It’s nothing new really, apart from the virtual buzz phrase, as organisations have always taken images of a running systems ready to deploy elsewhere and create clones. The difference with virtualisation is that it is a little bit easier to manage those images, and to do whole machine backups.
Why? Hyper-V is nowhere at the moment, and once again, you’re supporting your biggest competitor.
Does Microsoft certify and support Windows on top of Xen? Errrr, no, I thought not. It’s also something people have been doing for years, and years, and years with VMware. You’re not doing anything new.
Yer. Microsoft’s idea of running Linux is an officially sanctioned version of it in the form of Suse, which provides Novell no competitive advantage whatsoever in the manner its executives seem to think it does.
Might I suggest that you work to actually create a decent installation infrastructure in your enterprise OS that allows third party ISVs and people to install a piece of software in a sane manner?
I don’t know what it is, but Novell seems to have caught the hypervisor and virtualisation buzz phrase bug, where the mere mention of these two words somehow translates into revenue, profit and success.
I would rate up your comment if I hadn’t already posted in this thread.
SuSE is dead to me.
Suse isn’t dead in that we have OpenSuse, and the code won’t be going away whatever happens. The people around that project seem to be clued in reasonably well and they’ve done some meaningful stuff. The problem is that the people at Novell making the decisions just aren’t making proper use of OpenSuse, and the people, to help them. More fool them, really.
The whole idea of a “personal” computer, is that it can operate on its own.
I suppose that many company’s dream is two have a many fools enslaved to what they are selling, paying periodically, but that creates a system with a catastrophic single point of failure.
We were “liberated” from central administration and those terminals on our desks in the 80s by the PC… and ended up with a PC island nightmare. The LAN and Client/Server applications, each with their own protocols, arrived to save us… and fell on their swords. Hemlines are going back down. The pendulum is swinging back. Whatever metaphor you please, everything old is new again. But history never really repeats itself. Not exactly, anyway. Today’s terminals are qualitatively different from the wyse60’s and AT&T 4410’s and 605’s I grew up on.
Standalone OSes *are* the terminal today. And they can run apps all on their own, too. And to think we used to be happy with just programmable function keys…
Edited 2008-04-18 01:58 UTC
Oh yes, what a gain of comfort. 🙂 Okay, at least a gain of employment, because you need more IT staff to keep the “PC terminals” running than you needed to keep real terminal running.
Actually, our university’s library features two kinds of workstations: Centrally administrated and silently running Sun Ray systems and a bunch of roaring beige box PCs. You can easily imagine which systems the students do prefer.
Well, I do often say this when “PC experts” are coming up with something “new” (at least they think so): It has been there years before they started clicking around. 🙂
In most cases, yes, they are. And that’s why the stand-alone OS running on them is that important. From the past, you surely know that the user’s operation speed (at the terminal) was mostly determined by the speed of the machine he communicated with, and the speed of the communication line itself: s_o = f(s_m, s_c). On today’s “PC terminals”, the OS has an important influence. If it’s full of bloat, dancing elephants, jiggling wobble stuff and music, it cannot run well; but if it’s configured well, maybe intentionally optimized for the usage the system (as a whole) is intended to, then it can be a place of powerful work. The ability to administrate this OS from a central point is another advantage, still giving the option to do individual stuff on one machine that you don’t want to have on another one.
Recently, I saw a full featured PC running “Windows XP” acting as a 3270 terminal. What a waste! Imagine the costs of the PC: The system itself, the power consumption, the “Windows” license… well, sometimes I think some (!) users would be better of with terminals. Why? Because the “heavy” stuff is run on some kind of mainframe. And no, it isn’t dead.
(I do see these “waste of resources” quite often, at the Karstadt shopping center, the Sparda-Bank banking center and the regional tax office (Oberfinanzdirektion).)
And we still do have them on the keyboards! Wow! 🙂
Since the PC was declared dead over 10 years ago!
This will probably sit alongside –
“I think there’s a world market for about 5 computers.”
(Thomas J. Watson, Chairman of the Board, IBM, circa 1948)
… 🙂
“I think there’s a world market for about 5 typewriters.”
(Steven Bergman, Nobody Important, circa 2008) 🙂
Edited 2008-04-18 04:08 UTC
Desktop Linux didn’t pan out, so it’s time to move on to some other wish….I mean “prediction”.
people are buying more and more laptops and handheld interweb devices. How would those benefit from a stripped down os?
everything else supplied by google and similar?
is that i think they may be on to something, but aiming it towards the wrong place.
look at the eeepc and similar. sure, there are a lot of people that have hacked that one to hell and back, but in essence its a appliance.
it comes with all the software most users would need, a tailored interface to make it fast to launch them, and thats it.
thing is this that outside of the geek sphere, the power of the pc becomes its own worst enemy.
if people instead bought appliances, possibly ones that one could interconnect to share data if needed. said complexity would drop. one would have one physical item that did one task, making it easy to sort things out in the users mind. and no background processes, or tray apps. if its not in use, nothing is running, end of story.
In the fukced up IT culture as it is for the last couple decades, appliances are the worst possible thing that could happen. With complete platforms some of us can at least help themselves, with locked down crap offered by major players we are simply their monkeys.
I understand that from the big player’s perspective it sure looks like they got him by the balls, and maybe they do. But if distributed methods of addressing and indexing catch up, everything could turn upside down again.
Thom, you are emotionally invested in this, and this is why you can’t see the demise. The OS market is dead. The OS market will live, but in a different concept. In 20 years from now, no one will know what kind of OS they are running. Because it won’t matter anymore. We started getting towards that since the release of XP.
The OS market is dead.
It is not dead and will never be. Consumers will always need an OS to use because not every software can be provided to them via network. In the future everyone might have free access to the internet but atleast as long as that isn’t the case people will need local installations of software. And even if everyone did have high-speed, high-quality connections I still doubt f.ex. games would be a good idea to be provided over the networks..And what about the security issues? Atleast I wouldn’t want to work on something very important on an app that is run over the network.
In the business world the situation is more or less the same. Even if it was possible to run any single app over the network businesses would most likely wish to run a server of their own for delivering the needed software to the clients and for that they’d still need a server OS. And again, some clients would still probably need offline installations too.
You don’t seem to understand. When we say “it’s dead”, we mean that it’s something that will still exist, but it won’t bother the normal user. It will be a COMMODITY. Just like no one cares about the firmware inside the latest Bluetooth headset or even your TV remote control. Nobody cares. Same thing with OSes. OSes will simply be transparent platforms, and no one will care if it’s Linux, or Windows or Mac.
It’s already dead then because Joe User doesn’t care if its Windows, OSX (Mac isn’t an OS) or Linux. He cares if it does what he needs.
On the other hand, there always will be people who do care.
Just like they said about the car decades ago. Didn’t exactly pan out, did it?
You’re awfully out of touch with normal users if you think they’re remotely bothered about the OS they use. They are bothered only about what they can achieve with it.
As for saying ‘the OS market is dead’. Well, all I can say is being able to mod OSN staff down is really the ticket here.
Edited 2008-04-18 12:42 UTC
You don’t seem to understand. When we say “it’s dead”, we mean that it’s something that will still exist, but it won’t bother the normal user
People have said the same thing about televisions, cars, computers in general and so on and so forth. Not going to happen. And, the OS market is driven by what businesses want, not by home-users, so you just can’t declare OS market dead if home-users don’t care what OS they run. Besides, home-users have never cared about it as long as they get to do what they wish, and if they have some specific app (like my sister specifically wants MSN Messenger) that it runs on the computer they own. That brings another point why OS market will not die: I just simply will not believe that all software will run on any available OS. There will always be some OS which is not supported, and as such there is always market for the supported OSes. You can’t deny that.
Just like no one cares about the firmware inside the latest Bluetooth headset or even your TV remote control
Difference: you use those only for ONE specific purpose. Computers are used for gazillions of different uses AND people install applications on them to further expand the uses of the computer. This would be a valid example if computers were really only appliances, used only for a few very specific tasks. But since that is not the case your example is moot.
Operating systems can never be completely transparent platforms because they have implications for how they will be used, and what you can use them with and for. If you have Windows, Linux and Mac platforms you can never make them transparent because they’re different.
However, all that will be taken care of because in twenty years the dominant OS will be Linux, everyone will be chasing Linux compatibility, and unlike Windows, you can actually achieve that goal.
90% of the Userbase doesn’t know what kind of OS they’re using. Today! They buy a “Computer”, use it somehow and that’s it. Maybe they know what kind of OS they’re running when some problems come up. 😛
Just for entertainment:
Tech Support: “May I ask what operating system you are running today?”
Customer: “A computer.”
or
Tech Support: “What operating system are you running? Windows 95?”
Customer: (a little too excited) “95, 97, 98, I’ve got them all!
After conferring with her husband, it turned out she owned a Macintosh with System 8.1.
🙂
More here: http://www.rinkworks.com/stupid/cs_os.shtml
Excellent, that site was one of the sources that lead to my conclusion. 🙂 Just great.
Nat is Novell CTO of _open source_, not the overall CTO — that is Jeff Jaffe.
So… allow me to dream a little.
I wonder if it will be possible someday to, using virtualization techniques load up several operating systems at once, switch between them and use their software independent of what OS you are running it in.
Imagine Linux, OSX and windows all running side by side and you can use any program inside of another OS.
Impossible? Reality-to-be?
I do think the tendency will be something different, more the kind of Java applications, i. e. making the applications independent of the OS so they can run everywhere (provided a kind of runtime environment or interpreter).
Today’s virtual machines, alternative binary interfaces and emulators do already provide a comparable kind of operation, but they still require a “supervisor OS” that controls the whole operations environment. Maybe some kind of hardware will provide means to really run different OSes side by side (multiple real CPUs, own memory, combined I/O channels and interconnection).
No, not quite… more reality-to-have-been. If I remember correctly, we did this on our ESER mainframes… they could run “client OSes” within the “supervisor OS”. So you could, for example, have a EC1056 running OS/ES and within this, running PSU for the UNIX developers. Something similar has been possible on the IBM 360 mainfranes many decades ago.
I don’t think this is impossible. In fact, this is already being done to a certain degree with apps like “andLinux” (http://www.andlinux.org/index.php). BTW, andLinux is not a virtual machine application.
“The Standaone OS is dead”
Not as long as it’s illegal to sell both OS and core applications together.
Only if it involves leveraging an existing monopoly to gain a foothold in another market. Attempting to spread such obvious falsehoods as that, Almafeta, weakens one’s credibility without strengthening one’s case.
Edited 2008-04-18 15:49 UTC
… because I would like to see the days when computers were computers return, for those of us who are genuinely interested in computers.
Alas, modern PCs are more like home entertainment centres, which have to suit the whims of people who are more concerned about watching YouTube and visiting FaceBook than doing anything really interesting with their hardware.
Doubly unfortunate is that OS design has to take that into account.
He forgot to mention that all the apps would be written in Java and they would write a Java version of Linux that would be running in virtual machine that is written in Java. Isn’t that kind of norm every time somebody gets in his head to start making underpowered hardware with no storage media combined with fallacy of streaming applications into them as needed? Only new thing this time is that the applications would be virtual machines.
Network PC already failed miserably years ago and DVB Multimedia Home Platform never got to the point of being stillborn. It seems that they just stopped printing brochures for it and hoped that nobody would ask why their digitalTV never got the interactive features they promised.