With the release of Longhorn in 2007, Microsoft said it will offer “role-based” versions of Windows in which only the code needed to perform a given function will be included in a particular build of the operating system. Could Longhorn’s new, component-based approach be the start of something big?
Microsoft Goes to Pieces
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
It’s about time that they work towards a modular system. *NIX’s modularity has always been a better approach. I still don’t think that the “role-based” setup is quite right, though. I’m also interested in the marketing techniques they’ll use. If they make it so you have to pay for each version if you’d have a reason to switch I’ll seriously lose what faith I have in MS.
Oh Joy – patching 10 different versions of the same OS
Trust me they will sell you something that barely works by itself unless you purchase something else to give it more functionailty.
Now, consider a small to medium office:
Are you really going to buy separate windows versions for your clients and servers, pay for all that licensing; per person using a workstation, per person accessing the server; per person doing whatever… Unfortunately most people do..
Seriously now, one sloppy burnt copy of any linux distro will satisfy for the clients and provide all the services under the sun, and the tools to develop almost anything… Dear God why??
No. This can’t be. Microsoft trying to emulate *nix. I must say that it’s a valiant effort, but I don’t think its going to work. Indech and Anon really hit the nail on the head. Fork out a $100 for a barely functional OS, then add another, say, $150 to add some functionality. I dont think so. I would rather go with DSL and add the stuff I need for free. Just sounds better to me.
Funny, we’re already seeing this in Win2k3 server. Once you pick the roles you want your server to have the install program grabs the requisite software from the install CD. You still get the generally bloated base OS, but this isn’t even something new in the Windows realm. Including all those different options for functionality will probably push the install to beyond a single CD however, just like Linux.
I really thought we’d see Longhorn sooner. 2007 is a long way away (more than 2-3 years depending on when in the year they deliver). If they deliver in late 2007, Apple will have released 10.4 and probably 10.5 as well. Ubuntu will have gone through 4-6 releases before Microsoft gets their next one out. Is it just me or does anyone else think that Windows XP is going to start looking very old by 2007?
role based is only for business editions.
The consumer/workstation version may still be the same as now, having more features than we desire installed by default.
“. 2007 is a long way away (more than 2-3 years depending on when in the year they deliver). If they deliver in late 2007, Apple will have released 10.4 and probably 10.5 as well. Ubuntu will have gone through 4-6 releases before Microsoft gets their next one out. Is it just me or does anyone else think that Windows XP is going to start looking very old by 2007?”
thats just the beginning. what will really be the thing is that MS will phase out all the old win 98 and stuff and the operating system would have a significantly bigger migration path. Just what the competition needs.
It is pretty sure going to be an exciting game to wach
I hope they don’t read it but they are wrong:
a) It is not an easy task to cut out a program once you have writen it hard linked with other functionalities and Windows is a huge EXE all putted together.
b) They seem not to be able to sell WinXP-04 with their patches applied, so how would they sell 12 versions of Windows at the same time?
c) Nobody installs Damn Small Linux in a server. You install a minimal Debian, because you know that if someday in a near future you need this or this other package to add up another functionality to your server you can apt-get it easily. It wouldn’t make sense to install a limited functionality version of windows if that means you’ll have to reinstall everything again if in a future you want to add functionality to your server.
d) Ok, you have a windows server optimized. There is no explorer and no graphical interface on it. All memory is free for your application: Now you just have to run the setup.exe of your application, customize it through the install process and let it running… no… wait… how are you going to run the setup.exe file without having a graphical interface?
The most exciting thing from an OS POV of the next few years is that a significant percent of the population will be facing the decision: do I upgrade to a completely different OS (Longhorn) offered by Microsoft, or do I make the switch to Mac OS X or Linux, since it’s just as much work?
2007: finally, the end of an OS monopoly?
Also, good thing Java and .NET are maturing more and more: I think cross-platform programming will finally be important in 2007, for just about everyone.
This is about the 10th time I’ve read about this. Can we have some fresh news?
“…The idea of a slimmer Windows…”
“…move toward smaller, simpler Windows versions…”
No doubt in my mind that the system resources will be significantly greater than that of Win2K3!!!
Longhorn “workstation” is still on track for 2006 (according to an article I was reading a few weeks ago) – This article is referring to the server version of LongHorn
It is so much fun to watch these Linux trolls coming here and saying bad about future windows when they have no idea what it is.
Linux has copied most of the things from other OS so there is no point in saying that this was in Linux before because frankly OS is ever evolving software. Everyone borrows idea and btw modularity is not a Linux concept.
Microsoft did well with a single solution fits all and if they are moving to a componentized solution then i bet it will be successful. So to all the Linux trolls please stop flaming or talking crap. Let us wait and see when they ship it. Till then rest in peace…
In 2001 Steve Ballmer said in a commercial that linux is a cancer. Now in 2007 they are going to modularize windows. Competition is just great!
“The most exciting thing from an OS POV of the next few years is that a significant percent of the population will be facing the decision: do I upgrade to a completely different OS (Longhorn) offered by Microsoft, or do I make the switch to Mac OS X or Linux, since it’s just as much work?”
/quote
You’re dreaming. A significant part of the population will still be buying Dells, HPs, ect, with Windows already installed.
“I hope they don’t read it but they are wrong:
a) It is not an easy task to cut out a program once you have writen it hard linked with other functionalities and Windows is a huge EXE all putted together.”
That task could be automated. There are such tools for java
perthaps it could be done for other languages as well. Remember .Net is a clone of java.
“b) They seem not to be able to sell WinXP-04 with their patches applied, so how would they sell 12 versions of Windows at the same time?”
Or not sell any versions at all, just let people pay for whatever lines of code they use, and how often they use it.
Probably could they charge different fees depending on the numbeer of processors.
“c) Nobody installs Damn Small Linux in a server. You install a minimal Debian, because you know that if someday in a near future you need this or this other package to add up another functionality to your server you can apt-get it easily. It wouldn’t make sense to install a limited functionality version of windows if that means you’ll have to reinstall everything again if in a future you want to add functionality to your server.
”
So what makes you think you can’t apt-get it from Microsoft. Just wait a year or two, and you will find out that Microsoft tells you that they actually invented apt-get.
“d) Ok, you have a windows server optimized. There is no explorer and no graphical interface on it. All memory is free for your application: Now you just have to run the setup.exe of your application, customize it through the install process and let it running… no… wait… how are you going to run the setup.exe file without having a graphical interface?”
Just go to their webpage, enter your 72 digit licence code and hold your eye to the retina scanner to log in. Fill in what you need and your system will be automagically configured for you on a given box in your network.
But somehow this seams more like an IBM idea, so perhpas not. They would probably have called it OS on demand.
I have been fantasizing about this for years. I would love to have a Windows installation that was specific to a category of tasks. Of course it would be awful if one had to buy multiple versions at $150 a pop. I would prefer the purchase of the “Full OS” at regular price, and the option of a slim/specific package during installation. Some current Macromedia products ask you if you are a developer or designer the first time they run, for example.
edit: Oops, forgot where I was… LINUX!!!!
Trust me they will sell you something that barely works by itself unless you purchase something else to give it more functionailty.
It depends on how they will implement it .If everything continues to be avaible but in a more modular way and professionals are getting more control over what is installed and what isn’t,being able to enable just those services you actually need to a bare minimum, a feature in itself then it’s not to bad at all.Time will tell.Selling less for more money will most likely be suicide and give open source the free card,maybe they would be better of and sell all stuff to the chinese in that case.At least they show some vision and do show something now,instead of making excuses for not getting things done till the anounced/predicted release dates.I thought they were much more clever then that.Man i hate excuses.And feeling a cow being milked, instead of being a equal business partner.
Till then rest in peace…
Rest rosts,and that costs money.
Now, rather than simply selling task-specific editions of Windows, Microsoft may let systems administrators choose which core elements of Windows to include at installation. That’s good news for administrators, but the question is how far Microsoft will go in letting customers define and extend those role-based configurations. How much flexibility will IT gain – and how much control is Microsoft willing to give up?
The key obstacle is support.How much control can you give in order to give the current support,so it’s profitable but yet not milking the cow.Is it feasonable to let the clientele compile the kernel?In a black box way that lets you choose all the options when lets say you compile a Linux kernel, but yet you don’t have to worry about boot managers,partitions etc because the program takes care of it all? Wouldn’t be to bad,all in 64-bit even better but i think Intel doesn’t like that to happen nor does Dell,to name a few.
But increased global competition and the need to react faster to changing market conditions could provide the impetus for …
That’s the most interesting part of the whole article,it affects us all.
If you’re Longhorny enough, it will be the start of something big…
Awesome article, very good news, yeah, it looks like Longhorn will be something big, very big. Microsoft will easily catch up. Right now, people are stuck with Windows XP with no evolution since years, while KDE/Gnome continusly improved, so right now windows looks a bit “out” compared to Gnome and KDE, but be prepared RMS Fan`s boys ! With Longhorn, Microsoft will easily catch up !
I still don’t have the damndest clue why ANYBODY would use windows as a network server.. Just the sheer fact that most of the “so called” admins can’t even do anything more than click a pretty icon that says “Click Here”… And why waist valuable overhead on usability? administration is for administrators! if anybody expects Ms. Daisy to manage the network for the day then…
I thought it was alraedy role based..? After a clean install of 2k3, your first login prompts you to choose what roles you want for the server like DNS, AD, IIS, etc. and it installs what it needs then. How is this anything new? What? Now we have to pay more for all of them in one? They are all still going to include the same bloated codebase..
I still don’t have the damndest clue why ANYBODY would use windows as a network server.. Just the sheer fact that most of the “so called” admins can’t even do anything more than click a pretty icon that says “Click Here”… And why waist valuable overhead on usability? administration is for administrators! if anybody expects Ms. Daisy to manage the network for the day then…
Even if your “so called” admins have no other skills than randomly checking checkboxes and clicking there, it doesn’t say that the GUI have to run on the server. Typically admins have more than one server to admin, and would probably prefer to do that from one workstation. In fact the servers could/should be totally headless, no display, no keyboard, no mouse.
First off. Noboby said that Linux invented modularity. We are just saying that Microsoft is going to use what *nix had all along. Secondly, I’m not a Linux zealot. I’m actually using Windows more than any other OS (CAD program only runs on Windows).
As far as dishing Microsoft, I just don’t think that people will pay a lot of money for add-ons. That is obviously if Microsoft goes down that route, which I think it would.
Balmer had that wrong then – windows is the cancer.
I have no doubt that when longhorn ships ALOT of users will stick with microsoft. The thing that worries me is the time frame. Thinkhow much OS have evolved in 4 years, in computing more than anyother industry, 4 years is a LONG time.
For all we know skyOS may become the os of choice in that time 🙂
I have no doubt that when longhorn ships ALOT of users will stick with microsoft. The thing that worries me is the time frame. Thinkhow much OS have evolved in 4 years, in computing more than anyother industry, 4 years is a LONG time.
Four years is a long time in computer industry, that is true. But the typical customer of windows is not in the computer industry. They expect stability of the platform. If they spend money on a piece of machinary as expensive as a system upgrade from one OS version to the next, they will expect that investment to last at least 10-15 years. Given that for each new upgrade of windows we have seen since NT4 the additional business value added by the upgrade have been less and less for each new version of windows companies will be more and more reluctant to upgrade.
This means that the rate of major upgrades from Microsoft is allready far too high. A more modular OS could be a way to remedy this, as it may be easier to get people to upgrade minor parts of the OS as the cost of deployment may be more controlable.
The home market is a whole different matter though. Home users upgrade to run the last games and are prepared to take some risks of breaking old applications to do it. If their computer fails to work for a day or two while upgrading its not a big deal.
So we will probably see a modular update system for enterprise users and a more monolithic system for home users.
Now you dont know what windoze you are running so you cant patch every 30 mins. Hurray
“Microsoft trying to emulate *nix.
My thought exactly. Specifically, the source-available versions like the BSDs and Linux.
“I must say that it’s a valiant effort, but I don’t think its going to work. Indech and Anon really hit the nail on the head. Fork out a $100 for a barely functional OS, then add another, say, $150 to add some functionality. I dont think so. I would rather go with DSL and add the stuff I need for free. Just sounds better to me.”
While you are right and this move will cause them pain, they have to do it. Just like Apple had to drop the MacOS architecture for OSX’s layered one. Layers can be stripped…while integrated programs can’t without breaking other things.
This move will help to solve multiple problems for Microsoft;
* Competition in specific segments: They can’t compete — short of forcing or paying people to use them — in the embedded space or in High Performance Computing (HPC). They also are getting chewed up on the server end.
One of the many reasons for this is simple; The desktop. The depenedncy on a GUI is a real problem as it adds quite a bit of overhead that isn’t necessary or even used for rack mounted servers, embedded devices, and high performance computing (clusters, mainframes, …).
Defect reduction: With the current tying of user-level applications to the core OS, the system dependencies are overly complex. This makes it harder to fix when defects are discovered — including security related problems.
Resolving dependencies: You can’t do the equivelent of “apt-get upgrade” on Windows. Windows installer and updates tend to be painful; when you update any system library, different parts tend to break. This was partially solved by the ability of the OS to handle multiple versions of the same libraries though there are still a spagetti string of dependencies that remain from the core OS through to the user level apps.
Process isolation: If it’s integrated into the OS…it can be used as an attack vector. All the checking of messages being passed back and forth will not help when either a defect exists and can impact other integrated parts or marketing insists on a feature that is clearly stupid (see how much Microsoft has back tracked after years of complaints). The movement in *nix to one process level user per-process is one thing that is very difficult to pull off in Windows with the way it is currently designed. Dropping the integration allows for this level of per-process and per-resource protection.
I could go on and on….
They should have done this years ago. Nearly every other OS in use today has already. Just like Apple’s switch to a *nix core and a layered approach, it will take Microsoft many years to make the transition and it will not be entirely smooth.
It makes you wonder why they don’t just suck it up and replace the NT kernel with one of the BSDs. With that, the conversion to an all-*nix world will be complete.
With Longhorn, Microsoft will easily catch up !
In 2 years, where will KDE, Gnome, Mac OSX, and others be? Catching up might not be enough.
Now you dont know what windoze you are running so you cant patch every 30 mins. Hurray
How manny different Linux boxes do you have within one distribution, yet they all use practically the same repository.The (security) updates aren’t perse related to the version number or shade of pale but more to what’s vulnerable.So what’s the problem?
I doubt that this is the start of something big unless by big you mean bloated. Microsofts ability to produce a credible OS has to be in question since there in no empirical evidence to suggest they are capable of such an accomplishment.
I run two production server OSes at work. Redmond w2k3 server and Slackware Linux.
Personally, I would love to tell you how stable w2k3 actually is but my comments would be moderated down rather quickly.
Funny thing, my Redmond 2k3k servers, as was everyone elses on campus, were configured by over-priced Redmond consultants.
This was a mandatory migration due to an Lookout xtraChange server upgrade that requires AD. All servers on the wire had to be running AD for AD to work. Nice trick.
My Linux-based solution works, doen’t require constant patching or upgrading and has unbelievable uptime. I can devote my extra time to assisting staff, faculty, and students in the educational process.
Wow, OSNews has been pathetically infected by the same clueless bunch that roams freely on Slashdot
Eduard Pertíñez, you don’t know what you’re talking about. Modular Windows is a reality here, today, has a price tag and has been a reality for years. It’s called Windows Embedded. There was an Embedded edition of Windows NT 4, and there is a much more modular Embedded edition of Windows XP, not to speak of the Preinstallation Environment and the bootleg Embedded/PE hacks floating around on the Internet. Unattended installation is supported natively by Windows Installer – the short story of it is that, if you produce server software for Windows and want it to have any relevance, you have to deliver it in a Windows Installer package. The Apache Foundation took the clue a long time ago, MySQL was a bit slower to catch up but they eventually did, and we’ll see about the others
To everyone raising retarded points about patching: use Windows at least once. Patching isn’t even component-based, it’s file-based. Take the time to look into the patch database at least once (huge XML file). Singular files are checked – no file found, feature not present, patch not necessary
Please — *PLEASE* — don’t use Redmond when you mean Microsoft, Lookout when you mean Microsoft Outlook, xtraChange when you mean Microsoft Exchange. It makes Linux advocates in general look like this;
http://www.penny-arcade.com/view.php3?date=2002-07-22&res=l
Well, I still remember that was not too long ago that OsNews published an article suggesting that the hobby OSs were a waste of time because by the time they were ready to do real work the new Windows would leave then in the dust.
What dust? Lately it seems every time LornHorn is mention it release date moves a little further away. And I hope there is no-one foolish enough to think the first release will not have all sorts of patches needed for the things you only find out after you software is given out to people who don’t know what they are doing and thus end up trying to do things with your software that you never imagined anyone would try.
Now it is another two to three years before we see LongHorn out there, what will happen to present software in the meantime? Fact it, unless Microsoft ships LornHorn with it as well it will be hard for them to ship a better browser than FireFox with another two years of development on it for example.
Seems like Linux and hobby OSs are just going to keep increasing thier market share.
2007 – is that like “root” in H4XX0R? The future is looking sweet for Linux.
nothing else
2007 – is that like “root” in H4XX0R? The future is looking sweet for Linux.
I think this is about servers. There might be some workstation version of Longhorn before that. But even so, Microsoft clearly have a problem. Today windows XP only have 2/3 market penetration three years after its first release.
This probably means that most of the XP out there is a result of somebody buying a new PC with XP on it, and that means the upgrade rate for existing users is very slow. The reason for this is naturally that ever since windows NT4 the additional business value you get from updating to a new windows release have been less and less for each new windows version that have hit the market. There is no reason to believe Longhorn will be any different.
As the free desktop in the form of Gnome, KDE, OpenOffice, Linux,… becomes more mature it is only a matter of time before some big PC maker will start to sell preinstalled Linux desktops.
By doing so they can lower their prices in an attempt to get a larger market share. Perhaps Lenovo (who just bought the IBM PC devision) will be the first one. They have a big Chinese market that is quite Linux friendly. Microsoft threats by WTO proxy is not likely make them more windows oriented. So such a move would make sense to Lenovo.
If, or rather when, that happens there is a big company from Redmond that is in big trouble. If the only reason to upgrade is that the OS comes with your new computer, people will buy the cheeper Linux variant and then either keep it as it is, or just install their old windows version on top of it. Either way, a loss for Microsoft.
“It makes you wonder why they don’t just suck it up and replace the NT kernel with one of the BSDs. With that, the conversion to an all-*nix world will be complete.”
MS Xenix 2007 will be released as Open Source, having taken advantage of the power of the Open Source movement to do the grunt work, with the proprietary Avalon 2/.Net/Win64 API on top.
Then where did Windows copy everything from? Mac OS.
But wait a minute. Even DOS wasn’t made by MS. They bought it and changed the name.
But wait a minute. Where did DOS come from? It is simplified version fo UNIX.
The End.
Why doesn’t Microsoft cut all the crap they have stuffed into Windows Server and install a bare minimum system instead? They can charge around $200 for this. Then during the installation and through utilities (for later on, after the initial installation), they can give the end user the capability of selecting what to install and what to do with it. The end user then can pay accordingly depending on what services they install. Like Microsoft can have a 4-8 processor kernel support, and charge $400 for it. They can also have IIS, which might cost $100. DNS and WINS might cost $50 each. Active Directory Services might cost $400. This way they can give the system administrator the capability of designing the infrastructure as they see fit while giving them a secure base to build on. Furtheremore, if the administrator ever decides that they want to install some extra services on a particular server, they won’t be faced with the nightmare of having to reformat the entire operating system and paying full price for an OS with this feature so they can get X functionality in there.
Microsoft just seems to never get it. No one asked for this, no one even wants this. Just give me the damned smallest base possible and give me the capability of installing/documenting whatever services I wish to and let me decide on how I want to design my infrastructure. Why is Microsoft trying to make software act like hardware?
Hmm … let’s see. Redmond XPloder is trademarked. So is Lookout extraChange server. So no, I won’t. Sorry.
Besides? Who are you? Relax, breath, and laugh a little.
Freedom to joke, even about our computer operating systems, is such a wonderful thing.
Show me a Windows server that runs at shell level, preserving system resources for processes that actually need them, not an unecessary GUI hogging resources on a machine that no one actually uses except for minor maintenance.
If you can’t use shell (or DOS level) commands to make necessary alterations to server specific operations, then you either do not know what you are doing, or the level of command interface to processes is insufficient.
If you do need a GUI for some operations, then simply start one while performing the tasks, then shut it down again.
Microsoft is always coming up with good ideas, but their implementation of the idea leave a lot to be desired. Fragmenting Longhorn, while apparently a good idea, will quite probably leave more problems than it solves.
At least this time, MS is not simply applying a bandaid to it’s core problems, it intends to hack the problem up into timy little bits and let them lie where they will.
I agree with the premise that if you are using a Windows server anywhere in your network, you have left a gaping security hole, and a managerial nightmare in your system.
Why are some of you so anxious to start a war between these two sites?
You don’t like Slashdot? There are plenty of “Bill Gates is da Bomb” forums on the Net. But if you’re a Windows fan, in case it’s not obvious from the pictures of Gates made up to look like a Borg, and in case it’s not obvious that it’s an Open Source forum first and foremost, then let me be the one to tell you. Find another forum. It’s not for you. It’s for people interested in something else besides Windows.
http://www.fiveanddime.net/billgatesfanclub.html
I still don’t have the damndest clue why ANYBODY would use windows as a network server..
Probably because you’ve never actually had to manage a network full of Windows desktops.
And why waist valuable overhead on usability?
If an idle login screen is using enough resources on your server to affect performance, it’s probably time to upgrade that 486.
While you are right and this move will cause them pain, they have to do it. Just like Apple had to drop the MacOS architecture for OSX’s layered one. Layers can be stripped…while integrated programs can’t without breaking other things.
Windows is already modular and “layered”.
One of the many reasons for this is simple; The desktop. The depenedncy on a GUI is a real problem as it adds quite a bit of overhead that isn’t necessary or even used for rack mounted servers, embedded devices, and high performance computing (clusters, mainframes, …).
As previously mentioned, if a completely idle login screen (all the “GUI” a Windows Server needs in typical usage) is using enough resources on your server to impact performance, you’re probably not interested in performance in the first place.
Defect reduction: With the current tying of user-level applications to the core OS, the system dependencies are overly complex. This makes it harder to fix when defects are discovered — including security related problems.
Please clarify what you mean by “user level applications” and “core OS”.
Resolving dependencies: You can’t do the equivelent of “apt-get upgrade” on Windows.
Windows Update is the closest analogue. Some software developers (eg: Veritas) offer their own equivalents.
You’ll only ever get an equivalent of apt-get upgrade if everything on your server is Microsoft.
Windows installer and updates tend to be painful; when you update any system library, different parts tend to break.
On the contrary, installing applications and updates is trivial and almost completely trouble free, assuming said applications are remotely up to date and you don’t try and circumvent the OS’s self-protection mechanisms.
This was partially solved by the ability of the OS to handle multiple versions of the same libraries […]
Back in the mid ’90s, you mean ?
I think it’s time to update your knowledge beyond Windows NT 3.51.
[…] though there are still a spagetti string of dependencies that remain from the core OS through to the user level apps.
Like what ?
Process isolation: If it’s integrated into the OS…it can be used as an attack vector. All the checking of messages being passed back and forth will not help when either a defect exists and can impact other integrated parts or marketing insists on a feature that is clearly stupid (see how much Microsoft has back tracked after years of complaints). The movement in *nix to one process level user per-process is one thing that is very difficult to pull off in Windows with the way it is currently designed. Dropping the integration allows for this level of per-process and per-resource protection.
Processes run under the context of the user that starts them. That’s the same process isolation you get anywhere else.
IE’s “integration” doesn’t imbue it with magical powers and high-privilege access to the entire system at any time. It runs as a regular user process.
I could go on and on….
Please do.
They should have done this years ago. Nearly every other OS in use today has already. Just like Apple’s switch to a *nix core and a layered approach, it will take Microsoft many years to make the transition and it will not be entirely smooth.
Windows already has a modular architecture. They’re not doing anything more than letting you twiddle a few more configuration knobs at installation time.
It makes you wonder why they don’t just suck it up and replace the NT kernel with one of the BSDs. With that, the conversion to an all-*nix world will be complete.
Probably because the NT kernel is technically superior in just about every way.
“Probably because the NT kernel is technically superior in just about every way.”
That’s quite a claim. How about some examples?
“Four years is a long time in computer industry, that is true. But the typical customer of windows is not in the computer industry. They expect stability of the platform. If they spend money on a piece of machinary as expensive as a system upgrade from one OS version to the next, they will expect that investment to last at least 10-15 years.”
Er, what? How many of your friends are still using 15 year old computers? Even non-geek ones?
sorry Uno, misread you entirely. good post.
right, yeah, way back in the mists of time they bought a dirty hack of CP/M and based an operating system on it. This was a tiny little toy OS. it was hardly what DOS became by version 6.22, though.
Processes run under the context of the user that starts them. That’s the same process isolation you get anywhere else.
There are processes/services that run on the system account which is at least equal to root,no matter who is currently logged on.
Hmm … let’s see. Redmond XPloder is trademarked. So is Lookout extraChange server. So no, I won’t. Sorry.
http://www.penny-arcade.com/view.php3?date=2002-07-22&res=l
It’s all I can say…if you won’t listen.
I’ve read your comments in the past. I’ve read your reply to me this time. I’ll give you full points for consistancy in the message you promote.
If you are honest might I suggest that you have some homework to do? If not, bravo on your trolling skills. That said, I see no need to comment further.
By drsmithy (IP: 61.88.3.—) – Posted on 2004-12-09 22:11:02
Probably because you’ve never actually had to manage a network full of Windows desktops.
umm. thats actually part of my job…
If an idle login screen is using enough resources on your server to affect performance, it’s probably time to upgrade that 486.
..no, you see I don’t have to waist money on a pointless upgrade… as I said – I don’t use windows…
user I can only give my findings as I see them, Microsoft may very well pull off what people say they can’t… that is build a decent OS, now I don’t care what linux or windows people say, reguardless of hardware i’ve put the OS’s on (from 486’s to AMD 64’s) Microsoft has only one thing in it’s favour… and thats pre packaged software… (not there is not lots of *nix software but go in to a computer store…) anyway as for the argument of a server wasting resouces on the gui (hello windows) not only does the “login screen” waste a shitload of resources it’s also a security risk (more dlls used, etc etc, exploitable) also when windows freezes (and it will… altho it takes win 2k3 a little longer) u can’t repair till u reboot… there is no repair shell and if the gui goes just try and ssh (oops not ssh…)er er use terminal services to log in (what u mean u forgot to install it and it cost a shitload per seat — my boss) so more downtime for rebooting and crossing of fingers that it actually boots and then askes why the unexpected crash (or shutdown, reboot) why doesn’t it know? event viewer? forget it
) and as for paying to have something removed that should not be on the HD in the first place (hello virus, adware, spyware) and b4 u ask i know that there are free removers but only if it’s personal use… i can’t imagine 500 pc’s being personal use.
just my fiver’s worth
oh i havn’t mentioned the downside of the *nix installations like erm, erm, i upgraded the kernel and had to reboot, lets see when did i last have to reboot because of a app hang… i’m thinking…. nope still thinking…. ah uptime will tell me… 452 Days 15 Hours 9 Mins ago… oh well…