Saying that you keep a Windows system clean is no different than saying you run a Linux desktop. Neither is out of reach for a power user; both are not the norm.
I agree, however, the implication I was responding to was that such a feat was impossible, not unlikely.
Added to that, in a managed environment, keeping Windows machines clean is *not* especially difficult and certainly shouldn't require periodic reinstalls/reimaging unless the users are actively trying to exploit them (eg: school computer lab).
Umanaged home desktops are a completely different scenario with regards to keeping machines "clean" and, IMHO, cannot reasonably be compared.
Viruses/malware are not a joke. Your anecdotal evidence proves absolutely nothing; statistically, the vast majority of computer users are infested, badly. More computer users are infested with viruses/malware than vote in the presidental election. More computer users are infested than support ANY given political issue, *including* the concept of first amendment rights.
Indeed - and that doesn't even highlight the *big* problem, which is that most of them are infested due to their own actions and no amount of "OS security" will ever help them.
The malware problem is *not* going to improve any time soon. In fact, outside of some pretty hardcore international law enforcement co-ordination, I doubt it's _ever_ going to improve. Technology has a pretty poor record at solving social problems.
At a recent auto expo, Ford CEO William Clay Ford Jr. spoke of how fellow company officer Derrick Kuzak was asked to rid his car of all the annoying squeeks, quirks, and failing parts that had made the product highly unreliable over the few years he has owned it.
However, Mr Kuzak had never performed any basic maintenance tasks on his vehicle, nor taken it for regular servicing, always left it parked outdoors in conditions ranging from sub-zero cold to fry-an-egg heat and sandstorms, never locked the doors (often leaving the key in the ignition), used incorrect sized tyres, was unable to operate the clutch, regularly drove the car (a small hatchback, approximately twenty years old) over kerbs, through hip-deep water and down heavily-rutted offroad trails, and didn't even hold a driver's license.
But seriously, If you must stick to Windows, the only way to insure your safety is to make an image of your clean system, and periodically restore from that.
I guess the systems I've kept uninfested for years without reinstalling are just figments of my imagination ?
Interesting, when talking about the design of windows, we're not allowed to mention all the win32 backwards compatability goodnes that was ported from the 9x branch?
Windows NT was released 2 years before Windows 95. Its design phase started 7 years before the release of Window 95. Heck, Windows *3.0* wasn't even released when Windows NT was being designed.
[Relevant] Things weren't ported from Windows 9x to NT. If they weren't independently implemented on both platforms, they were ported from NT to Windows 95.
Riiiiiiiight, in your windows world, windows was completely redesigned from the ground up with NT, all of the APIs were chucked out, the A:, b:, c: naming conventions were thrown away, the UI changed completely, filetypes were no longer based on the extension, but on the file's header. In other words, Microsoft completely threw out backwards compatability.
The win32 API was created first for Windows NT and then on Windows 95 (and kind of on Windows 3.x, in the form of "Win32s"). IIRC, Win32 for Windows 9x and win32 for Windows NT had separate codebases. Every other example you've listed there is a matter of shell and UI semantics and has nothing to do with OS design.
Your argument is a strawman. "Backwards compatibility" does not require a common heritage. OS X happily runs a significant proportion of MacOS Classic software, despite them being completely different OSes. FreeBSD will run most Linux binaries - sometimes better than Linux does - despite them having nothing in common (from an OS design and implementation perspective). Just because Windows NT runs DOS, Windows 3.x and Windows 9x software, does *not* mean it shares code, or even design concepts, with any of them.
"Windows" (well, strictly speaking, at the time, OS/2) *was* completely "redesigned" with Windows NT. That was the whole fucking point of it !
Do you honestly believe you can talk about windows 'design' (and again, I state, that's a kind word for windows) without considering its past? (back through 9x & dos?)
Absolutely. The design of Windows NT and the "design" of Windows 3.x, 9x and DOS are completely and utterly different. Outside of high-level APIs and UI, you'd be hard pressed to find anything about them that was similar.
I apologise for calling you a bigot - you're clearly a shill.
Yes, yes. And my mother wears combat boots and smells of elderberries. I've heard it all before.
Oh, and I never said windows was stealing from unix - I don't believe in terminology like that - I said it was becoming more unix-like, this means unless you're playing silly games, that windows is introducing features and workflows that are common (not unique) to Unix-like operating systems.
Then why isn't it becoming more some-other-OS-like, rather than more unix-like ?
Not to mention, you *still* haven't listed all these ways Windows is becoming "more unix-like". I certainly hope you're thinking of something more interesting than graphical sudo prompts.
I still don't get why everyone is into vitualization of servers. Just one thought; if you have to take down the server inquestion for something like a server relocation; you take down 5 (virtual servers). What happens why you fry a motherboard? Not one but 5 servers now go down.
It only makes sense if you have - or expect to have - requirements for a large and/or dynamic number of machines.
For example, if you have 10 different production machines, for 10 distinct tasks, but want to provide redundancy for them (in the form of standby machines) then you could either go out and buy ten additional machines, or you could buy 1 - 3 machines and have the ten standby servers as VMs (the chances of more than a single primary machine failing simultaneously is pretty rare, multiple simultaneous failures even more so).
Another example might be where load varies throughout the day to different areas of your infrastructure - so instead of always having to have the physical machines to cope with your peak load, you can take additional VMs on and offline as needed to cope with varying load.
A third example is where you want to have a good, scalable, partitioned architecture from the start (eg: by separating functions out into independent machines) but don't have a suitable hardware budget (or current requirements) to justify it. By using VMs, you can create your multi-machine architecture on a single physical machine and the subsequent migration to multiple physical machines (as requirements increase and/or budget allows) becomes relatively trivial.
Finally, there are situations where physical rack space is extremely limited, but you still want to have multiple "machines". Since you can fit a lot of power into only a few RUs these days, it could be quite feasible to have a couple of multicore 2U servers running a dozen VMs only taking up 4U, rather than a dozen real machines taking up 12U (or a blade chassis taking up ~7U).
With all that said, the incredibly low (and dropping) cost of relatively powerful servers has, IMHO, put a serious dent in the usefulness of VMs in production environments.
"Longhorn is going to stop being a whole new thing and more of an XP with a lot of good new stuff," said one developer close to Microsoft, who requested anonymity.
Your sources need to be more credible.
Well, not really, following your "reasoning" regarding NeXTSTEP and OS X Vista's filesystem has been being worked on since at least 2000. Keep reading, you'll see why...
Your comparison is broken.
My "reasoning" regarding NeXTSTEP and Vista is not only sound, but solidly backed by known facts. OS X is a major revision of NeXT, with similar levels of modifications made as those going from Windows 2003 -< Vista.
Sounds pretty important. But, instead they're basing the file system off the 2003 Server file system.
WinFS is not - and never was - a filesystem. You might want to get the basic facts right before you start trying to get too advanced. However...
Which was "...first introduced to technical beta testers in mid-2000, it was known by its codename, "Whistler Server"..." which slightly increases your development timeline.
Let's be fair, if you're going to insist that OS X is simply NeXTSTEP, which of course, it's got it's heritage in, then let's look at the "full development" cycle of Vista. The file system started being distributed to developers in 2000. Looks more like seven years after all.
No, it doesn't, because as I have repeatedly said, pretty much all Longhorn development was thrown out ca. 2003 and the project started over from the Windows 2003 codebase.
Your assertion that "the full development cycle of Vista" should be measured from the first alpha code that would eventually become the codebase that Windows Vista would be branched from (ie: two degrees of separation), is ridiculous on its face. You may as well say OS X is ten years late because NeXT appeared in the late '80s. Do you measure the development time of OS X 10.4 from the release of OS 10.0, or 10.3 ? Do you measure the development time of Linux 2.6.16 from 2.6.15 or 1.3 ?
Apple also didn't hype their production of OS X in 1997 when they purchased NeXT, so you can't really make the comparison that you are without looking at the marketing machinations of both companies. Microsoft likes to announce things way in advance to get people talking, while Apple generally likes to keep things tightly under wraps until they make their announcements. The two companies work differently than each other, certainly, but you can't simply ignore how they hype their products as that shows quite a bit about how they do business.
You *really* seem to be having difficult separating "Vista the product", from "Vista the codebase". Since this is a discussion about *software development*, "Vista the codebase" is what matters. "Vista the codebase" branched from Windows 2003 ca. 2003. It's been in development for around 3 - 4 years. I have no interest in what emanates from various companies' PR and marketing departments in a discussion about a development lifecycle.
I still contend that Microsoft's development schedule from first announcement to actual launch, which is still to be seen, is the true measurement.
And from the perspective of measuring the actual software development time of the codebase that will be Windows Vista your contention is wrong.
Now, you can certainly argue that the *product* Windows Vista is very late. You can certainly argue the project for a successor to Windows XP has been badly managed. But you can't argue that the codebase of Windows Vista has been around for more than 3 - 4 years, unless you're also going to apply the same principle (and subsequent criticism) consistently to every other platform as well.
Or maybe you do ? I wonder, when each minor (or even major, for that matter) revision of Linux comes out, do you apply the same logic to call its development timeline "embarassing", since it started back in 1991 ? Do you call every release of OS X "embarrassing", because NeXTSTEP started back in 1986 ?
(As always, the logical gymnastics (and just sheer effort) that some people go to just to get another version of "Microsoft sucks" really does amaze me. Why bother with the fancy verbage ? Just say it.)
However, you are mistaken as to when Microsoft started talking about Vista/Longhorn.
I didn't say anything about when Microsoft "started talking about Vista/Longhorn".
I said Windows Vista (ne: Longhorn) was the product that was going to supercede Windows XP. Windows XP was released in 2001. Based on that - without knowing anything else - the *earliest* anyone would expecting a successor would be ca. 2003.
I said the codebase from which Vista has been developed has only been being worked on since about 2003. This is pretty well known, as there was quite a bit of fanfare at the time about Microsoft "rebooting" the Longhorn project and basing it off the Windows 2003 codebase.
With a supposed 2007 launch, that is still getting pushed back, it'll only be just under six years from first mention to launch and only four years since the first reported launch date.
Measuring from "first mention" - if you're talking about software development - is, well, meaningless. Measuring from the first projected release date is marginally more reasoanble, but taking the first date that was ever talked about is silly, considering every major (and even most minor) software projects run well over time. I can't imagine there was anyone outside of Microsoft's PR department who thought "Longhorn" would see the light of day before mid-late-2004.
Also with a completely stripped down feature set since most of what they had originally touted it to have won't be present.
The only feature I can think of off the top of my head that Vista will be missing is WinFS, and while it does class as "major", I can't imagine many actual users will miss it. Saying it has a "completely stripped down feature set" is ridiculous.
For the largest software company in the world with the most monetary resources available to them than any other software company, I find that timeframe amazing, don't you?
Not really. But a) I haven't got a chip on my shoulder, b) I'm primarily interested in the the software development cycle, not marketing fluff and c) I'm well aware that throwing more people at a problem - particularly a software development problem - rarely helps it proceed more quickly.
When you consider that Vista pretty much started over in 2003 and that it is a significant improvement on nearly all aspects of Windows - including major changes in things like the display system, driver model, memory management and other low-level components - then the ~4 years the code will have actually been in development, isn't unreasonable or unexpected. It took Microsoft about four years to go from NT4 to NT5 (Windows 2000), which was a smaller magnitude of change. It took Apple 4 - 5 years to turn NeXTSTEP into OS X 10.2 (arguably the first production quality, feature-complete release) - that's a change on about the same magnitude.
So, no, I don't find Vista's timeframe particularly "amazing". As a *product*, it's extremely late. As a *software development project*, it's pretty much bang on time.
Good morning, Pot, pleased to meet you. My name is Kettle.
There's always room for improvement of course but the "dramatic" difference you're claiming is simply wrong.
To go back to my multi-monitor example, setting up four screens in Windows with a single dual-head card and a couple of PCI cheapies was a matter of a few mouse clicks (I'll be very generous and say it took ten minutes, including downloading and installing the drivers). Reconfiguring it (eg: when I add and remove monitors) takes seconds.
Getting that same setup even minimially configured in Linux (FC5, Ubuntu, RHEL) - it never even got close to working as well - was a half-day marathon, and I'm _very_ familiar with Linux. Modifications to the setup (when I was brave enough to try them) were similarly frustrating affairs and usually resulted in breaking the whole setup multiple times.
The problem with Linux on the desktop is it's ok if you want a centrally managed and/or unchanging "kiosk" or "appliance" style setup and it's great if you want to spend days/weeks/months/years tweaking, customising and hacking around with it, but it sucks if you want a powerful, consistent, transparent interface that just works *without* having to spend half your time making it work just the way you want.
In most situations the difference in maintenance effort isn't much.
I have no "maintenance effort" for Windows. Or for OS X. They just work and I just use them. To be honest, I'd prefer to use OS X, since I spend most of my time interacting with other unix machines and in many ways its nicer, but it's just too slow for heavy interactive multitasking. Windows does ~90% of the job at ~130% of the performance, so I use it instead.
So, write a polite complain to the manufacturer of your machine for not providing linux drivers on CD.
I imagine manufacturers would be more inclined to make the effort delivering "drivers on CD" if they didn't have to worry about every different minor kernel revision breaking them.
So what's some "innovation" - as per the dictonary definition - you can find in shipping OSes from the last, say, twenty years ?
Might as well conclude physics and chemistry have nearly reached their endings, as well.
I'm pretty sure Physics is still making new discoveries on a regular basis.
I should also point out that I never said OS development "has reached its ending", I said there hasn't been much innovation that's appeared in the last ~20 years.
There's plenty of innovation left to be done in OSes. There's lots of innovation in OSes recently.
Like what ?
*shrug* You've got to justify a claim like this;-)
It all depends on what you mean by "innovation". If you mean "new features unlike anything ever seen or conceived before", then there's pretty much nothing.
That is totally untrue. When Unix hit the scene, it was advanced, but seen as only a solution for supercomputers and possibly mainframes and minicomputers. It wasn't seen as a general computing solution
When UNIX hit the scene, it was seen as a hobbyist toy. It was a very long time before anyone saw it as mature enough for "real" use.
That, along with the rest of your brief history, still misses the point - none of these OSes was doing anything outlandishly new, different, or unseen. They were all either reimplemented or making obvious, incremental improvements on, features and concepts already seen or described elsewhere.
You are comparing a snapshot in time from twenty years ago to the modern day and going "see, innovation happened", while completely ignoring the twenty years of incremental development in the interim.
The problem is that the word "innovation" is so overused as to have become meaningless. Worse, double standards are used when applying it. For example, you might say Windows is not "innovative", but by that measure you could not then turn around and say OS X or Linux *are* "innovative".
I am not denying there's a lot of cool, new stuff still to be done in the field of operating systems. I am merely pointing out that most/all of it comes from obvious incremental developments of existing technology and concepts.
The situation doesn't need to "improve" any further.
Yes, it does. Setting up multiple monitors in Windows (and OS X) is *trivially* easy. Doing it in Linux - assuming you can make it work at all - even *with* well-supported hardware, is nowhere near as simple or flexible.
If you think that installing Windows on PC hardware is as simple as "hitting [Enter] half a dozen times", you haven't installed Windows much.
If all you're after is a normal install on an otherwise unused PC, that's about all it takes.
Maybe you're thinking of a system restore from manufacturer's disks, but a real from-scratch Windows install on recent hardware requires that you track down drivers for many different pieces of hardware, followed by hours of re-installing applications to get back to a fairly complete, working system.
That's hardly an apples-to-apples comparison. Shall we discuss what an incredibly massive PITA installing Linux can be if your hardware is not supported, or (as is disturbingly common) only "kind of" supported ?
If you're going to compare an ideal scenario Linux install then the only valid Windows install to talk about is a similarly ideal scenario - and in that, an installation is simple.
Linux is enormously widely used at universities, in research labs, in retail desktop applications, and on corporate desktops (and that's not even counting all the embedded consumer applications). The presence of OS X in all those areas is at best modest.
IME, Universities are dominated by Windows, and MacOS, with Linux's (/unix's) presence outside maybe the Math, CS and Engineering departments to be practically nonexstant (and even within them, relatively small - *BSDs tend to be nearly as popular). Retail applications (by which I assume you mean POS) tend to be DOS or Windows. Corporate desktops are overwhelmingly Windows. The idea that Linux machines are more common than Macs amongst home users doesn't even pass the laugh test.
So, I think the assertion that OS X is more widely used on the desktop than Linux is implausible; if you want to make that claim, why don't you supply some evidence?
I didn't say OS X, I said MacOS, including Classic. I have no evidence and, quite frankly, little interest in chasing any up for a discussion like this. I am basing it on my experience in educational and private enterprise environments and the simple fact that MacOS and Windows have been established forces on the desktop for 15-odd years, whereas Linux has been a reasonable choice for only ~4 and offers few compelling reasons to migrate to if you already have established Windows or MacOS infrastructure.
Outside of the server room, I can't think of anywhere I've seen Linux as anything more than a esoteric oddity. Macs - from the environments I've seen - are downright common compared to Linux desktops.
Maybe the distributions are just different in Australia. But if there are significant Linux desktop deployments out there, I'd like to know where they're hiding.
People who have only been observing the industry for a handul of years shouldn't be talking about "tradition". The Athlon was the first product AMD made that could even offer a match to intel's CPUs, let alone exceed them.
I'm sure there's more than enough people here who remember how intel poorly comparbadly the K6 ran Doom, for example.
Not to mention the atrocious record of motherboard chipsets for >K6 AMD processors that, alone, contributed more to slowing their uptake by the market than any other factor (it astounds me that VIA has managed to stay in business).
There is no "lack of Linux adoption"; at this point, Linux is the most common OS after Windows, with OS X trailing a distant third on servers and a closer third on desktops.
I think the claim that Linux has more presence on the desktop that MacOS [X], requires some evidence.
The reason it's not there yet is probably because there hasn't been a big need for it in the past: most vendors didn't make drivers available, and the few people for whom this mattered spent the 30 minutes to figure it out.
Well, speaking as someone who can - and has - "figured it out", the experience was enough to keep me away from Linux as a desktop until the situation has *dramatically* improved. It's the tedium and difficulty of these sorts of tasks - which should be trivially easy - that really needs addressing in the Linux desktop.
My job is adminning unix machines (amongst others). I waste *more* than enough time in my day making servers work (or keep working), I have zero interest in performing that same struggle on the machine that should be a transparent tool. Linux on the desktop is simply too much work, as far as I'm concerned.
Installing it by hand is a major undertaking that involves significant text-mode interactions.
Maybe if your idea of a "major undertaking" is hitting [Enter] half a dozen times. In which case, I'd hate to think how you judge something like the OpenBSD install (or the prior version of Ubuntu).
If Tom's review is a good list of the new features, well... I'm significantly underwhelmed.
It's not. In particular, the meat and potatoes of the under-the-hood changes are pretty much ignored.
Microsoft is entirely capable of producing a next-generation OS with lots of new, innovative technology. Vist isn't it.
There's not a lot of "innovation" left to be done in OSes. It's not like much "innovation" (if you want to go by a strict dictionary definition) has appeared in OSes for 20 - 30 years.
Seven years ago, Windows 2000 wasn't even released.
While Vista the *product* is superceding Windows XP, releasedonly in 2001, the Vista *codebase* has only been in development for about 3 years. They basically started it over again from scratch (as a development of the Windows 2003 branch) in 2003.
The software our company writes uses the registry to store settings. However, the customers that buy our software like to lock down their users to where they have to 'write' access to ANYTHING, especially the registry.
Denying users write access to their own Registry hives is just silly. Indeed, I can't imagine you could even run Windows in such a configuration.
Unless you developers are trying to write to non-user parts of the Registry. In which case they are incompetent and should be fired before they can cause any more damage.
Of course, our other product writes to text files...and we are constantly having to tell people to give write access to those text files.
Why ? Where are you putting them ? Users have write permissions already to files they create in their home directories - where else are you trying to create text files ?
And finally, another product writes to files that are stored in the users space. (Flavor of the day is "C:\Documents and Settings\username\Local Settings\application\" Have fun walking a non-techy user through checking that. (Especially since it's typically hidden by default.)
I hope (but doubt, based on the information given thus far) your developers are aware they shouldn't be storing anything that's meant to be persistent in that directory ? It isn't part of a user's "persistent" profile and, for example, won't be copied if you're using Roaming Profiles.
I guess there's no way to win...but we've definitely 'lost' the most when using the registry.
Follow Microsoft's guidelines on where to store different types of data. Then, it's not your fault if some over-zealous administrator tries to lock their machines down too hard.
No, it took five years (really, three, since they basically started over in 2003) to create a significant upgrade to XP.
You shouldn't judge the improvements in Vista from a bunch of screenshots any more than you should judge the improvements between two Linux distros three years apart based on a few screenshots.
Slashdot Mirror
I agree, however, the implication I was responding to was that such a feat was impossible, not unlikely.
Added to that, in a managed environment, keeping Windows machines clean is *not* especially difficult and certainly shouldn't require periodic reinstalls/reimaging unless the users are actively trying to exploit them (eg: school computer lab).
Umanaged home desktops are a completely different scenario with regards to keeping machines "clean" and, IMHO, cannot reasonably be compared.
Viruses/malware are not a joke. Your anecdotal evidence proves absolutely nothing; statistically, the vast majority of computer users are infested, badly. More computer users are infested with viruses/malware than vote in the presidental election. More computer users are infested than support ANY given political issue, *including* the concept of first amendment rights.
Indeed - and that doesn't even highlight the *big* problem, which is that most of them are infested due to their own actions and no amount of "OS security" will ever help them.
The malware problem is *not* going to improve any time soon. In fact, outside of some pretty hardcore international law enforcement co-ordination, I doubt it's _ever_ going to improve. Technology has a pretty poor record at solving social problems.
However, Mr Kuzak had never performed any basic maintenance tasks on his vehicle, nor taken it for regular servicing, always left it parked outdoors in conditions ranging from sub-zero cold to fry-an-egg heat and sandstorms, never locked the doors (often leaving the key in the ignition), used incorrect sized tyres, was unable to operate the clutch, regularly drove the car (a small hatchback, approximately twenty years old) over kerbs, through hip-deep water and down heavily-rutted offroad trails, and didn't even hold a driver's license.
I guess the systems I've kept uninfested for years without reinstalling are just figments of my imagination ?
Windows NT was released 2 years before Windows 95. Its design phase started 7 years before the release of Window 95. Heck, Windows *3.0* wasn't even released when Windows NT was being designed.
[Relevant] Things weren't ported from Windows 9x to NT. If they weren't independently implemented on both platforms, they were ported from NT to Windows 95.
Riiiiiiiight, in your windows world, windows was completely redesigned from the ground up with NT, all of the APIs were chucked out, the A:, b:, c: naming conventions were thrown away, the UI changed completely, filetypes were no longer based on the extension, but on the file's header. In other words, Microsoft completely threw out backwards compatability.
The win32 API was created first for Windows NT and then on Windows 95 (and kind of on Windows 3.x, in the form of "Win32s"). IIRC, Win32 for Windows 9x and win32 for Windows NT had separate codebases. Every other example you've listed there is a matter of shell and UI semantics and has nothing to do with OS design.
Your argument is a strawman. "Backwards compatibility" does not require a common heritage. OS X happily runs a significant proportion of MacOS Classic software, despite them being completely different OSes. FreeBSD will run most Linux binaries - sometimes better than Linux does - despite them having nothing in common (from an OS design and implementation perspective). Just because Windows NT runs DOS, Windows 3.x and Windows 9x software, does *not* mean it shares code, or even design concepts, with any of them.
"Windows" (well, strictly speaking, at the time, OS/2) *was* completely "redesigned" with Windows NT. That was the whole fucking point of it !
Do you honestly believe you can talk about windows 'design' (and again, I state, that's a kind word for windows) without considering its past? (back through 9x & dos?)
Absolutely. The design of Windows NT and the "design" of Windows 3.x, 9x and DOS are completely and utterly different. Outside of high-level APIs and UI, you'd be hard pressed to find anything about them that was similar.
I apologise for calling you a bigot - you're clearly a shill.
Yes, yes. And my mother wears combat boots and smells of elderberries. I've heard it all before.
Oh, and I never said windows was stealing from unix - I don't believe in terminology like that - I said it was becoming more unix-like, this means unless you're playing silly games, that windows is introducing features and workflows that are common (not unique) to Unix-like operating systems.
Then why isn't it becoming more some-other-OS-like, rather than more unix-like ?
Not to mention, you *still* haven't listed all these ways Windows is becoming "more unix-like". I certainly hope you're thinking of something more interesting than graphical sudo prompts.
It only makes sense if you have - or expect to have - requirements for a large and/or dynamic number of machines.
For example, if you have 10 different production machines, for 10 distinct tasks, but want to provide redundancy for them (in the form of standby machines) then you could either go out and buy ten additional machines, or you could buy 1 - 3 machines and have the ten standby servers as VMs (the chances of more than a single primary machine failing simultaneously is pretty rare, multiple simultaneous failures even more so).
Another example might be where load varies throughout the day to different areas of your infrastructure - so instead of always having to have the physical machines to cope with your peak load, you can take additional VMs on and offline as needed to cope with varying load.
A third example is where you want to have a good, scalable, partitioned architecture from the start (eg: by separating functions out into independent machines) but don't have a suitable hardware budget (or current requirements) to justify it. By using VMs, you can create your multi-machine architecture on a single physical machine and the subsequent migration to multiple physical machines (as requirements increase and/or budget allows) becomes relatively trivial.
Finally, there are situations where physical rack space is extremely limited, but you still want to have multiple "machines". Since you can fit a lot of power into only a few RUs these days, it could be quite feasible to have a couple of multicore 2U servers running a dozen VMs only taking up 4U, rather than a dozen real machines taking up 12U (or a blade chassis taking up ~7U).
With all that said, the incredibly low (and dropping) cost of relatively powerful servers has, IMHO, put a serious dent in the usefulness of VMs in production environments.
Windows ("IE") has an equivalent called the "Certificate Store". I can't remember if it first appeared in Windows 2000 or Windows XP, however.
Firefox is just an application, so a comparison to OS-level functionality is nonsensical.
Your sources need to be more credible.
Well, not really, following your "reasoning" regarding NeXTSTEP and OS X Vista's filesystem has been being worked on since at least 2000. Keep reading, you'll see why...
Your comparison is broken.
My "reasoning" regarding NeXTSTEP and Vista is not only sound, but solidly backed by known facts. OS X is a major revision of NeXT, with similar levels of modifications made as those going from Windows 2003 -< Vista.
Sounds pretty important. But, instead they're basing the file system off the 2003 Server file system.
WinFS is not - and never was - a filesystem. You might want to get the basic facts right before you start trying to get too advanced. However...
Which was "...first introduced to technical beta testers in mid-2000, it was known by its codename, "Whistler Server"..." which slightly increases your development timeline.
Let's be fair, if you're going to insist that OS X is simply NeXTSTEP, which of course, it's got it's heritage in, then let's look at the "full development" cycle of Vista. The file system started being distributed to developers in 2000. Looks more like seven years after all.
No, it doesn't, because as I have repeatedly said, pretty much all Longhorn development was thrown out ca. 2003 and the project started over from the Windows 2003 codebase.
Your assertion that "the full development cycle of Vista" should be measured from the first alpha code that would eventually become the codebase that Windows Vista would be branched from (ie: two degrees of separation), is ridiculous on its face. You may as well say OS X is ten years late because NeXT appeared in the late '80s. Do you measure the development time of OS X 10.4 from the release of OS 10.0, or 10.3 ? Do you measure the development time of Linux 2.6.16 from 2.6.15 or 1.3 ?
Apple also didn't hype their production of OS X in 1997 when they purchased NeXT, so you can't really make the comparison that you are without looking at the marketing machinations of both companies. Microsoft likes to announce things way in advance to get people talking, while Apple generally likes to keep things tightly under wraps until they make their announcements. The two companies work differently than each other, certainly, but you can't simply ignore how they hype their products as that shows quite a bit about how they do business.
You *really* seem to be having difficult separating "Vista the product", from "Vista the codebase". Since this is a discussion about *software development*, "Vista the codebase" is what matters. "Vista the codebase" branched from Windows 2003 ca. 2003. It's been in development for around 3 - 4 years. I have no interest in what emanates from various companies' PR and marketing departments in a discussion about a development lifecycle.
I still contend that Microsoft's development schedule from first announcement to actual launch, which is still to be seen, is the true measurement.
And from the perspective of measuring the actual software development time of the codebase that will be Windows Vista your contention is wrong.
Now, you can certainly argue that the *product* Windows Vista is very late. You can certainly argue the project for a successor to Windows XP has been badly managed. But you can't argue that the codebase of Windows Vista has been around for more than 3 - 4 years, unless you're also going to apply the same principle (and subsequent criticism) consistently to every other platform as well.
Or maybe you do ? I wonder, when each minor (or even major, for that matter) revision of Linux comes out, do you apply the same logic to call its development timeline "embarassing", since it started back in 1991 ? Do you call every release of OS X "embarrassing", because NeXTSTEP started back in 1986 ?
(As always, the logical gymnastics (and just sheer effort) that some people go to just to get another version of "Microsoft sucks" really does amaze me. Why bother with the fancy verbage ? Just say it.)
I didn't say anything about when Microsoft "started talking about Vista/Longhorn".
I said Windows Vista (ne: Longhorn) was the product that was going to supercede Windows XP. Windows XP was released in 2001. Based on that - without knowing anything else - the *earliest* anyone would expecting a successor would be ca. 2003.
I said the codebase from which Vista has been developed has only been being worked on since about 2003. This is pretty well known, as there was quite a bit of fanfare at the time about Microsoft "rebooting" the Longhorn project and basing it off the Windows 2003 codebase.
With a supposed 2007 launch, that is still getting pushed back, it'll only be just under six years from first mention to launch and only four years since the first reported launch date.
Measuring from "first mention" - if you're talking about software development - is, well, meaningless. Measuring from the first projected release date is marginally more reasoanble, but taking the first date that was ever talked about is silly, considering every major (and even most minor) software projects run well over time. I can't imagine there was anyone outside of Microsoft's PR department who thought "Longhorn" would see the light of day before mid-late-2004.
Also with a completely stripped down feature set since most of what they had originally touted it to have won't be present.
The only feature I can think of off the top of my head that Vista will be missing is WinFS, and while it does class as "major", I can't imagine many actual users will miss it. Saying it has a "completely stripped down feature set" is ridiculous.
For the largest software company in the world with the most monetary resources available to them than any other software company, I find that timeframe amazing, don't you?
Not really. But a) I haven't got a chip on my shoulder, b) I'm primarily interested in the the software development cycle, not marketing fluff and c) I'm well aware that throwing more people at a problem - particularly a software development problem - rarely helps it proceed more quickly.
When you consider that Vista pretty much started over in 2003 and that it is a significant improvement on nearly all aspects of Windows - including major changes in things like the display system, driver model, memory management and other low-level components - then the ~4 years the code will have actually been in development, isn't unreasonable or unexpected. It took Microsoft about four years to go from NT4 to NT5 (Windows 2000), which was a smaller magnitude of change. It took Apple 4 - 5 years to turn NeXTSTEP into OS X 10.2 (arguably the first production quality, feature-complete release) - that's a change on about the same magnitude.
So, no, I don't find Vista's timeframe particularly "amazing". As a *product*, it's extremely late. As a *software development project*, it's pretty much bang on time.
Good morning, Pot, pleased to meet you. My name is Kettle.
There's always room for improvement of course but the "dramatic" difference you're claiming is simply wrong.
To go back to my multi-monitor example, setting up four screens in Windows with a single dual-head card and a couple of PCI cheapies was a matter of a few mouse clicks (I'll be very generous and say it took ten minutes, including downloading and installing the drivers). Reconfiguring it (eg: when I add and remove monitors) takes seconds.
Getting that same setup even minimially configured in Linux (FC5, Ubuntu, RHEL) - it never even got close to working as well - was a half-day marathon, and I'm _very_ familiar with Linux. Modifications to the setup (when I was brave enough to try them) were similarly frustrating affairs and usually resulted in breaking the whole setup multiple times.
The problem with Linux on the desktop is it's ok if you want a centrally managed and/or unchanging "kiosk" or "appliance" style setup and it's great if you want to spend days/weeks/months/years tweaking, customising and hacking around with it, but it sucks if you want a powerful, consistent, transparent interface that just works *without* having to spend half your time making it work just the way you want.
In most situations the difference in maintenance effort isn't much.
I have no "maintenance effort" for Windows. Or for OS X. They just work and I just use them. To be honest, I'd prefer to use OS X, since I spend most of my time interacting with other unix machines and in many ways its nicer, but it's just too slow for heavy interactive multitasking. Windows does ~90% of the job at ~130% of the performance, so I use it instead.
I imagine manufacturers would be more inclined to make the effort delivering "drivers on CD" if they didn't have to worry about every different minor kernel revision breaking them.
So what's some "innovation" - as per the dictonary definition - you can find in shipping OSes from the last, say, twenty years ?
Might as well conclude physics and chemistry have nearly reached their endings, as well.
I'm pretty sure Physics is still making new discoveries on a regular basis.
I should also point out that I never said OS development "has reached its ending", I said there hasn't been much innovation that's appeared in the last ~20 years.
There's plenty of innovation left to be done in OSes. There's lots of innovation in OSes recently.
Like what ?
*shrug* You've got to justify a claim like this ;-)
It all depends on what you mean by "innovation". If you mean "new features unlike anything ever seen or conceived before", then there's pretty much nothing.
When UNIX hit the scene, it was seen as a hobbyist toy. It was a very long time before anyone saw it as mature enough for "real" use.
That, along with the rest of your brief history, still misses the point - none of these OSes was doing anything outlandishly new, different, or unseen. They were all either reimplemented or making obvious, incremental improvements on, features and concepts already seen or described elsewhere.
You are comparing a snapshot in time from twenty years ago to the modern day and going "see, innovation happened", while completely ignoring the twenty years of incremental development in the interim.
The problem is that the word "innovation" is so overused as to have become meaningless. Worse, double standards are used when applying it. For example, you might say Windows is not "innovative", but by that measure you could not then turn around and say OS X or Linux *are* "innovative".
I am not denying there's a lot of cool, new stuff still to be done in the field of operating systems. I am merely pointing out that most/all of it comes from obvious incremental developments of existing technology and concepts.
Yes, it does. Setting up multiple monitors in Windows (and OS X) is *trivially* easy. Doing it in Linux - assuming you can make it work at all - even *with* well-supported hardware, is nowhere near as simple or flexible.
If you think that installing Windows on PC hardware is as simple as "hitting [Enter] half a dozen times", you haven't installed Windows much.
If all you're after is a normal install on an otherwise unused PC, that's about all it takes.
Maybe you're thinking of a system restore from manufacturer's disks, but a real from-scratch Windows install on recent hardware requires that you track down drivers for many different pieces of hardware, followed by hours of re-installing applications to get back to a fairly complete, working system.
That's hardly an apples-to-apples comparison. Shall we discuss what an incredibly massive PITA installing Linux can be if your hardware is not supported, or (as is disturbingly common) only "kind of" supported ?
If you're going to compare an ideal scenario Linux install then the only valid Windows install to talk about is a similarly ideal scenario - and in that, an installation is simple.
IME, Universities are dominated by Windows, and MacOS, with Linux's (/unix's) presence outside maybe the Math, CS and Engineering departments to be practically nonexstant (and even within them, relatively small - *BSDs tend to be nearly as popular). Retail applications (by which I assume you mean POS) tend to be DOS or Windows. Corporate desktops are overwhelmingly Windows. The idea that Linux machines are more common than Macs amongst home users doesn't even pass the laugh test.
So, I think the assertion that OS X is more widely used on the desktop than Linux is implausible; if you want to make that claim, why don't you supply some evidence?
I didn't say OS X, I said MacOS, including Classic. I have no evidence and, quite frankly, little interest in chasing any up for a discussion like this. I am basing it on my experience in educational and private enterprise environments and the simple fact that MacOS and Windows have been established forces on the desktop for 15-odd years, whereas Linux has been a reasonable choice for only ~4 and offers few compelling reasons to migrate to if you already have established Windows or MacOS infrastructure.
Outside of the server room, I can't think of anywhere I've seen Linux as anything more than a esoteric oddity. Macs - from the environments I've seen - are downright common compared to Linux desktops.
Maybe the distributions are just different in Australia. But if there are significant Linux desktop deployments out there, I'd like to know where they're hiding.
Ugh. Must use preview. That should say:
I'm sure there's more than enough people here who remember how comparably poorly the K6 ran Doom, for example.
I'm sure there's more than enough people here who remember how intel poorly comparbadly the K6 ran Doom, for example.
Not to mention the atrocious record of motherboard chipsets for >K6 AMD processors that, alone, contributed more to slowing their uptake by the market than any other factor (it astounds me that VIA has managed to stay in business).
Or they could go and claim their refund.
I think the claim that Linux has more presence on the desktop that MacOS [X], requires some evidence.
The reason it's not there yet is probably because there hasn't been a big need for it in the past: most vendors didn't make drivers available, and the few people for whom this mattered spent the 30 minutes to figure it out.
Well, speaking as someone who can - and has - "figured it out", the experience was enough to keep me away from Linux as a desktop until the situation has *dramatically* improved. It's the tedium and difficulty of these sorts of tasks - which should be trivially easy - that really needs addressing in the Linux desktop.
My job is adminning unix machines (amongst others). I waste *more* than enough time in my day making servers work (or keep working), I have zero interest in performing that same struggle on the machine that should be a transparent tool. Linux on the desktop is simply too much work, as far as I'm concerned.
Installing it by hand is a major undertaking that involves significant text-mode interactions.
Maybe if your idea of a "major undertaking" is hitting [Enter] half a dozen times. In which case, I'd hate to think how you judge something like the OpenBSD install (or the prior version of Ubuntu).
It's not. In particular, the meat and potatoes of the under-the-hood changes are pretty much ignored.
Microsoft is entirely capable of producing a next-generation OS with lots of new, innovative technology. Vist isn't it.
There's not a lot of "innovation" left to be done in OSes. It's not like much "innovation" (if you want to go by a strict dictionary definition) has appeared in OSes for 20 - 30 years.
Can't see how it will make any difference. NT's low-level features have never been lacking in terms of security (or much else).
Seven years ago, Windows 2000 wasn't even released.
While Vista the *product* is superceding Windows XP, releasedonly in 2001, the Vista *codebase* has only been in development for about 3 years. They basically started it over again from scratch (as a development of the Windows 2003 branch) in 2003.
Denying users write access to their own Registry hives is just silly. Indeed, I can't imagine you could even run Windows in such a configuration.
Unless you developers are trying to write to non-user parts of the Registry. In which case they are incompetent and should be fired before they can cause any more damage.
Of course, our other product writes to text files...and we are constantly having to tell people to give write access to those text files.
Why ? Where are you putting them ? Users have write permissions already to files they create in their home directories - where else are you trying to create text files ?
And finally, another product writes to files that are stored in the users space. (Flavor of the day is "C:\Documents and Settings\username\Local Settings\application\" Have fun walking a non-techy user through checking that. (Especially since it's typically hidden by default.)
I hope (but doubt, based on the information given thus far) your developers are aware they shouldn't be storing anything that's meant to be persistent in that directory ? It isn't part of a user's "persistent" profile and, for example, won't be copied if you're using Roaming Profiles.
I guess there's no way to win...but we've definitely 'lost' the most when using the registry.
Follow Microsoft's guidelines on where to store different types of data. Then, it's not your fault if some over-zealous administrator tries to lock their machines down too hard.
No, it took five years (really, three, since they basically started over in 2003) to create a significant upgrade to XP.
You shouldn't judge the improvements in Vista from a bunch of screenshots any more than you should judge the improvements between two Linux distros three years apart based on a few screenshots.
The low-level changes and improvements made in Vista are *significant*. They alone easily justify the major version bump (NT 5.x -> NT 6.x).