The only thing the latest Pentium IVs have on the Athlons is the SSE3 extensions, and that is pretty much irrelevant to most code. "It must be true, Intel told us."
Small correction - K8 revision E supports SSE3.
(Venice and San Diego cores in A64, Troy and Italy cores in Opteron).
And these branch predictors come at the expense of higher transistor counts, more heat, and the fact that >98% is not good enough to make up for the fact that your pipeline is *twice* the depth of your competitors'.
The P4 quite possibly has the best branch predictor out there right now, but that doesn't make up for the fact that it *still* underperforms compared to AMD's chips.
Actually, Ubuntu was already committed to 18 months of support on each release. So while it releases every 6 months, you'd be supported by security updates etc, at least two releases back. So you're not abandoned simply because there's a new release every six months, you just simply won't be running the latest and greatest.
Not only that, but the majority of AMD's outsourced jobs don't go to places like India, but rather to Germany, another first-world nation to workers that are making wages similar to, if not better than, those that their American counterparts. Sure, I'd prefer that money stayed at home, but in all fairness, I can't lump sending that money to another country where workers are generally treated a little bit better than the US as far as benefits go in the same category as sending programming jobs to India to save money. Lou Dobbs is a fairly smart guy, but he's got a huge nationalist streak, and it shows in reports like this.
1. With the next generation of consoles becoming nothing more than computers, what becomes the purpose of having two separate machines? Or perhaps the real point is, why use your computer for gaming?
Personally, I'd argue that this is an incorrect premise. The next-gen consoles are *not* general-purpose computers, but rather, extremely powerful media DSPs. The multicore, in-order execution of the Cell and the Xenon are meant to eat through datasets very quickly, but aren't going to be particularly powerful for general purpose use.
Both these CPUs are going to require quite a bit of rethinking of design by developers as traditional engine designs just aren't going to perform that well. Design is going to have the shift toward highly threaded engines which are designed around the idea of feeding datasets into each of the many cores as quickly as possible. The XBox 360 in particular was designed around the idea of dynamically generating as much content as possible, rather than using stored content as in the past, but the Cell's design lends toward the same type of approach.
I'm not sure whether or not this is why developers are reporting the machines to be relatively underpowered or not, but I'd certainly suspect it. To be quite frank, our current engine designs will *not* run well on this type of an architechture. In general computing terms, it wouldn't surprise me that the 360 is only twice as powerful. As a media DSP, however, Xenon should run circles around the P6 based CPU of the XBox.
In short, these next-gen consoles are based around a very specific set of requirements, and I wouldn't expect them to replace your desktop PC any time particularly soon. Set-top box, sure. But they're not general purpose computers.
That's exactly my experience. My family owns a PC shop, and we stopped buying WD drives around the time of their 30-60GBs because we were getting way too many failures - many DOAs and, worse still, drives that were failing a few weeks after being sold.
We've been selling pretty much exclusively Samsung and Seagate since then. The other *huge* often unmentioned advantage is that they're both much quieter than WD and Maxtor equivalents - Samsung being a little quieter in my experience.
Really, outside of the Raptor line, I see no compelling reason to buy a WD drive. I can definitely agree with the sentiment that I have *never* heard anyone say they hated Seagate drives, especially if you talk to the SCSI freaks out there.:-)
Are you really this unaware of how the GPU market works? There will be budget versions of the nVidia cards. Period.
In fact, nVidia is the choice of many enthusiasts right now specifically because they pulled through with a solidly performing mid-range card in the 6600 where ATI failed to do so. ATI's competitor to the 6600GT was supposed to be the X700XT - IE: the card that was paper launched and *NEVER* made it to market. So while ATI had a slight lead on the high end this time around, nVidia was the way to go for those of us who aren't comfortable spending $400 on a video card.
Whether ATI will repeat that mistake this generation remains to be seen, but nVidia getting a two month certainly won't help. While I doubt it'll be a GeForce FX fiasco, it's certainly going to be an uphill battle for them.
I have to say I've had the same observation about Slackware to Ubuntu convertees. I think a large part of this stems from the fact that, for a while, Slackware+Dropline Gnome was one of the most straightforward, easy to use Linux desktop environments around to a lot of people.
Ubuntu arrived on the scene at almost the exact same time that Dropline was starting to stagnate a bit - Todd had pretty much burned out on the project and started a transition from a one-man metadistro to the project being community maintained. So while Dropline had always had the latest Gnome packages inside of the week in the past, they were very, very late getting Gnome's October 2.8 release, and when HAL was added in, it was more than a bit flakey.
At this same time, Ubuntu 4.10 Warty was released, which was a stable, easy to install distro with the same type of focus on usability that Dropline had always had. For many, it was their first taste of apt on a system that was easy to get running, and seemed to be headed in the right direction.
Personally, I made the move in late October/early November, and haven't looked back. While I still love Slackware, its underlying system wasn't the friendliest in the world for a usability-oriented desktop system. It's a distribution that's really fleshed out much of the potential I saw in Dropline, and Canonical is moving everything in the right direction to keep this happening.
Although I also have to say that it's not just Slackware users pulled in by this - it's no coincidence that it's been the top distro on Distrowatch for quite some time now. I think the most compelling case I've seen personally is a good friend of mine that's a total Mac fanatic. He's a geek, and enjoys computers, but in the past, he's hated anything Linux simply because it was too much hassle for him. That said, he's fallen absolutely in love with Ubuntu, and has been installing it on almost any piece of x86 hardware he touches lately. That's quite an endorsement to me.:-)
Quite frankly, Dvorak isn't somebody I'd trust to confirm or deny anything. The man makes the same predictions over and over every year, and they're often so broad that they almost have to be right. I've come to find he doesn't know what he's talking about as often as not.
I got as far in that blog entry as him claiming that the Intel deal "happened" in early 2003, and Apple has simply been in the process of "developing the middleware" to handle the transition. For those who don't know, the G5 was introduced at WWDC in the summer of 2003.
So Dvorak is trying to say that despite preparing to transition to PPC970, which everyone was expecting to have longer legs than it's shown, Apple was already preparing to switch to Intel chips less than two years later, and only taking that long because they needed to "develop the middleware?"
Color me skeptical, and color Dvorak full of crap.
Personally, I'd rather get a single, widescreen display. I caught Dell's 2005FPW on sale for $400 a couple of weeks ago, and it's been outstanding, moving up from a dying 17" CRT @ 1600x1200.
I'd used widescreen displays previously on notebooks, but it was nice to finally get that kind of screen ratio on my desktop. The biggest thing about widescreen is breaking users of the habit many have of maximizing and subsequently minimizing *every* window they use and instead simply sizing the window down and leaving everything open. A widescreen allows you to work in a manner similar to a dual monitor setup, but without forcing your brain to constantly context-switch between the two displays. It's really a great compromise, IMO.
Except a 4000+ and a 3200+ are the exact same chip. The 4000+ has just been certified to run at a higher clock. Now, I suppose it's possible that if both were rated at the same voltage, you *might* be able to argue that you could drop both the clock and voltage lower on the 4000+ while remaining stable, thus reducing heat output.
Personally, if that were what you were shooting for, I'd go for an A64 mobile, which are essentially just binned to run at lower voltages anyway. I'd have to agree with other posters that I don't quite see the point of this test.
Actually, they're not 100% compatible. I'm not sure about the latest chips, but the first round of EMT64 chips Intel put out (Nocona Xeons) had serious compatibility issues with some AMD64 code. They wouldn't boot the Win x64 betas that were available at the time (though MS got them working with the next beta release) and Redhat had to do some serious kernel level hacking to get them working properly. The kicker of it was that part of RH's hack essentially was to do software paging for memory above 4GB - more or less negating what Intel claims the whole point of having a 64 bit chip is.
Again, I'm not saying that there are necessarily lingering compatibility issues or the like, simply that I highly doubt that Intel's x86-64 implementation is anywhere near as elegant as AMD's.
That depends largely on who's sending the message, just like IM. Most phones these days have extremely good predictive spelling, which actually makes it easier to just spell out a word than abbreviate it more often than not.
Calling SMS a 'code' is a bit off-target, IMO. SMS simply stands for "simple message system." Calling it a code is the Morse equivalent of calling the telegraph a code. It's the medium, not the data.
That was exactly my thought when I read this. Even if Debian is planning on moving to a much faster release schedule, this goes directly against the extremely stable argument I often get in Debian's favor.
I hate to hold Ubuntu up as an example every time Debian comes up, but they've at least got the right idea: A new release every 6 months, supported for 18 months. They also have plans for an 'Enterprise' release every 12-24 months, sporting Debian-like stability testing.
Personally, I'd argue that even 18 months is a bit on the short side for a production environment, but it's certainly better than 12. This isn't even taking into account that there's also a stable release pattern, and since 3 releases should be supported at any given time, there's nothing to stop you from picking one in the 'middle' if you prefer tried and true to the leading edge.
I was confused by the fact that you were replying to a post on the merits of Linux-style security then, my bad.
I do stand by the fact that I see nothing particularly special about OS X style permission escalation within the Unix world. In fact, it essentially just puts a pretty face on sudo. And I'm well aware that Mandrake has used PAM and the like for years now (MDK was actually my first distro, back in the 6.5 release). I pointed out Ubuntu in particular because it goes a step farther (as does OS X) by disabling the root account by default. This prevents users from running as root entirely, thus minimizing the "I drug my System folder to the trash" problems.
My point was primarily that OS X style priveledge assignment works well because it allows a user to run as a limited account by default without it being a complete pain, while still leaving the bar to screwing up the system high enough that it generally won't happen in normal usage. I got confused by the "OS X's security seems to work better" nature of your post since you were replying to someone making a similar point about Linux. Again, my bad.:-)
You may want to check out Ubuntu, if you haven't. It uses pretty much the exact same user authentication scheme as OS X. Root account disabled by default, escalated privs only when necessary. Updates are handled by an applet, all the admin tools will prompt for a root pass when needed, and most OSS software can be installed from Synaptic, the GUI package manager.
You're actually the first person I've heard complain about nforce2 video having problems. I've built quite a few systems based on nForce2 IGP chipsets, and haven't ever had a problem. Out of curiousity, which boards were you using? The one place I have had major headaches in the past is with nForce2-based MSI boards and drivers, but all of those were non-IGP.
Upgrade your video card every 6 months? Hardly. My current gaming PC is well over 2 years old at this point, and while I've added storage and optical drives along the way, the core system is still the same as the day I bought it - an Athlon XP 2600+ with a gig of RAM and a Radeon 9700.
While it wasn't a budget box at the time, it wasn't extraordinarily expensive either. Something to the effect of about $700. And I've yet to find a game that won't run acceptably well on it, up through all the latest - Doom 3, HL2, Far Cry, etc. Personally, I'm not even *considering* an upgrade until the A64 X2s are priced reasonably enough for me to afford one.
I'm certainly not arguing that raw hardware costs are lower for a PC - they're not. However, the idea that it gets incredibly expensive because you "have" to upgrade every 6 months is ludicrous. I haven't spent any more on the gaming aspects of my PC than I would have buying a couple of consoles (with no games, extra controllers, etc) at launch. 2+ years down the road and still kicking it? I can't complain about that.
Granted, the fact that I'm a geek and use my system for far more than just gaming, it's easier for me to justify, but all the same, I just fail to see where the "I have to spend $300 upgrading every 6 months just to keep up" mentality comes from. It just doesn't line up with the reality I've come to know in this field. Buy solid, middle of the line parts that are at whatever the price/performance sweetspot at the time is, and you'll be fine for quite some time.
(though ubuntu doen't yet feature a graphical fronted for this method)
Ubuntu uses.debs for prebuilt packages not in the repositories, and while I'm not in front of my Ubuntu box at this moment to check, I'm fairly sure Synaptic will handle those just fine.
I beg to differ. Ubuntu, which takes flak for still using Debian's text based installer, has a much more straight-forward install than any version of Windows. Nevermind Mandriva, Fedora Core, Xandros, SuSE or any number of other distributions with fully graphic installers that many would consider even more friednly.
While your average user probably wouldn't install Linux on their own, I imagine it's more a matter of motivation than anything. I have no doubt that most people certainly could get Linux on their machine with most modern, user-oriented distributions if they were to put forth the effort, and they most certainly could do so easier than with Windows, whose installer is loaded with confusing and conflicting options.
Seriously. Look at the installers side-by-side. I've certainly spent enough time looking at them working in IT, and it's very clear to me which of them needs work.
This is so incredibly offbase I'm not even sure where to start, outside of saying that I use OS X, Windows, and Linux all on a practically daily basis.
That said, a user can trash an OS X install in manner approximately equal to that of their ability to trash a Linux install. A normal user account has access to their home directory, and can essentially do what they want, up to and including trashing everything in it.
Some distros (namely Ubuntu) have taken up OS X style password prompting with root disabled by default, which I definitely think is rather important, even if you downplay it. It's one of the things about OS X that was well thought out, and it's good to see Linux distros adopting a similar system. The reason so many Windows users run as admin is two fold - for one some things just plain don't work. But the other is even when things work, it's a pain when you do need admin privs. By making escalating priveleges no more difficult than a simple prompt, users are encouraged to run limited accounts.
Not to say it's an infallible system. Remember the fake Office 'trial' for OS X that was going around on P2P a few months back? It'd prompt a user for their admin password, then go and helpfully start trashing files. Moral of the story? Any system is susceptible to user stupidity in the end, it's just a matter of how easy it is for them to trash it.
Honest question: Have you evaluated Ubuntu for this purpose? The reason I ask is because for the administration tasks a non-technical user might need to worry about (read: security updates and the like) it's a bit more user friendly, while maintaining a Debian base.
Not meaning to start a (meta)distro flamewar here, just genuinely curious since Knoppix HD installs always seemed somewhat like a hack to me.
Slashdot Mirror
Small correction - K8 revision E supports SSE3.
(Venice and San Diego cores in A64, Troy and Italy cores in Opteron).
And these branch predictors come at the expense of higher transistor counts, more heat, and the fact that >98% is not good enough to make up for the fact that your pipeline is *twice* the depth of your competitors'.
The P4 quite possibly has the best branch predictor out there right now, but that doesn't make up for the fact that it *still* underperforms compared to AMD's chips.
Actually, Ubuntu was already committed to 18 months of support on each release. So while it releases every 6 months, you'd be supported by security updates etc, at least two releases back. So you're not abandoned simply because there's a new release every six months, you just simply won't be running the latest and greatest.
Not only that, but the majority of AMD's outsourced jobs don't go to places like India, but rather to Germany, another first-world nation to workers that are making wages similar to, if not better than, those that their American counterparts. Sure, I'd prefer that money stayed at home, but in all fairness, I can't lump sending that money to another country where workers are generally treated a little bit better than the US as far as benefits go in the same category as sending programming jobs to India to save money. Lou Dobbs is a fairly smart guy, but he's got a huge nationalist streak, and it shows in reports like this.
Personally, I'd argue that this is an incorrect premise. The next-gen consoles are *not* general-purpose computers, but rather, extremely powerful media DSPs. The multicore, in-order execution of the Cell and the Xenon are meant to eat through datasets very quickly, but aren't going to be particularly powerful for general purpose use.
Both these CPUs are going to require quite a bit of rethinking of design by developers as traditional engine designs just aren't going to perform that well. Design is going to have the shift toward highly threaded engines which are designed around the idea of feeding datasets into each of the many cores as quickly as possible. The XBox 360 in particular was designed around the idea of dynamically generating as much content as possible, rather than using stored content as in the past, but the Cell's design lends toward the same type of approach.
I'm not sure whether or not this is why developers are reporting the machines to be relatively underpowered or not, but I'd certainly suspect it. To be quite frank, our current engine designs will *not* run well on this type of an architechture. In general computing terms, it wouldn't surprise me that the 360 is only twice as powerful. As a media DSP, however, Xenon should run circles around the P6 based CPU of the XBox.
In short, these next-gen consoles are based around a very specific set of requirements, and I wouldn't expect them to replace your desktop PC any time particularly soon. Set-top box, sure. But they're not general purpose computers.
Zombie dogs?
YES! IRL Quake is one step closer!!
That's exactly my experience. My family owns a PC shop, and we stopped buying WD drives around the time of their 30-60GBs because we were getting way too many failures - many DOAs and, worse still, drives that were failing a few weeks after being sold.
:-)
We've been selling pretty much exclusively Samsung and Seagate since then. The other *huge* often unmentioned advantage is that they're both much quieter than WD and Maxtor equivalents - Samsung being a little quieter in my experience.
Really, outside of the Raptor line, I see no compelling reason to buy a WD drive. I can definitely agree with the sentiment that I have *never* heard anyone say they hated Seagate drives, especially if you talk to the SCSI freaks out there.
No, he colored over the inner ring of the internet with a Sharpie.
Are you really this unaware of how the GPU market works? There will be budget versions of the nVidia cards. Period.
In fact, nVidia is the choice of many enthusiasts right now specifically because they pulled through with a solidly performing mid-range card in the 6600 where ATI failed to do so. ATI's competitor to the 6600GT was supposed to be the X700XT - IE: the card that was paper launched and *NEVER* made it to market. So while ATI had a slight lead on the high end this time around, nVidia was the way to go for those of us who aren't comfortable spending $400 on a video card.
Whether ATI will repeat that mistake this generation remains to be seen, but nVidia getting a two month certainly won't help. While I doubt it'll be a GeForce FX fiasco, it's certainly going to be an uphill battle for them.
I have to say I've had the same observation about Slackware to Ubuntu convertees. I think a large part of this stems from the fact that, for a while, Slackware+Dropline Gnome was one of the most straightforward, easy to use Linux desktop environments around to a lot of people.
:-)
Ubuntu arrived on the scene at almost the exact same time that Dropline was starting to stagnate a bit - Todd had pretty much burned out on the project and started a transition from a one-man metadistro to the project being community maintained. So while Dropline had always had the latest Gnome packages inside of the week in the past, they were very, very late getting Gnome's October 2.8 release, and when HAL was added in, it was more than a bit flakey.
At this same time, Ubuntu 4.10 Warty was released, which was a stable, easy to install distro with the same type of focus on usability that Dropline had always had. For many, it was their first taste of apt on a system that was easy to get running, and seemed to be headed in the right direction.
Personally, I made the move in late October/early November, and haven't looked back. While I still love Slackware, its underlying system wasn't the friendliest in the world for a usability-oriented desktop system. It's a distribution that's really fleshed out much of the potential I saw in Dropline, and Canonical is moving everything in the right direction to keep this happening.
Although I also have to say that it's not just Slackware users pulled in by this - it's no coincidence that it's been the top distro on Distrowatch for quite some time now. I think the most compelling case I've seen personally is a good friend of mine that's a total Mac fanatic. He's a geek, and enjoys computers, but in the past, he's hated anything Linux simply because it was too much hassle for him. That said, he's fallen absolutely in love with Ubuntu, and has been installing it on almost any piece of x86 hardware he touches lately. That's quite an endorsement to me.
System --> Administrative Tools --> Synaptic Package Manager.
:-)
Fully GUI based package manager complete with search function and automatic upgrade functionality. HTH.
Quite frankly, Dvorak isn't somebody I'd trust to confirm or deny anything. The man makes the same predictions over and over every year, and they're often so broad that they almost have to be right. I've come to find he doesn't know what he's talking about as often as not.
I got as far in that blog entry as him claiming that the Intel deal "happened" in early 2003, and Apple has simply been in the process of "developing the middleware" to handle the transition. For those who don't know, the G5 was introduced at WWDC in the summer of 2003.
So Dvorak is trying to say that despite preparing to transition to PPC970, which everyone was expecting to have longer legs than it's shown, Apple was already preparing to switch to Intel chips less than two years later, and only taking that long because they needed to "develop the middleware?"
Color me skeptical, and color Dvorak full of crap.
Personally, I'd rather get a single, widescreen display. I caught Dell's 2005FPW on sale for $400 a couple of weeks ago, and it's been outstanding, moving up from a dying 17" CRT @ 1600x1200.
I'd used widescreen displays previously on notebooks, but it was nice to finally get that kind of screen ratio on my desktop. The biggest thing about widescreen is breaking users of the habit many have of maximizing and subsequently minimizing *every* window they use and instead simply sizing the window down and leaving everything open. A widescreen allows you to work in a manner similar to a dual monitor setup, but without forcing your brain to constantly context-switch between the two displays. It's really a great compromise, IMO.
Except a 4000+ and a 3200+ are the exact same chip. The 4000+ has just been certified to run at a higher clock. Now, I suppose it's possible that if both were rated at the same voltage, you *might* be able to argue that you could drop both the clock and voltage lower on the 4000+ while remaining stable, thus reducing heat output.
Personally, if that were what you were shooting for, I'd go for an A64 mobile, which are essentially just binned to run at lower voltages anyway. I'd have to agree with other posters that I don't quite see the point of this test.
Actually, they're not 100% compatible. I'm not sure about the latest chips, but the first round of EMT64 chips Intel put out (Nocona Xeons) had serious compatibility issues with some AMD64 code. They wouldn't boot the Win x64 betas that were available at the time (though MS got them working with the next beta release) and Redhat had to do some serious kernel level hacking to get them working properly. The kicker of it was that part of RH's hack essentially was to do software paging for memory above 4GB - more or less negating what Intel claims the whole point of having a 64 bit chip is.
Again, I'm not saying that there are necessarily lingering compatibility issues or the like, simply that I highly doubt that Intel's x86-64 implementation is anywhere near as elegant as AMD's.
That depends largely on who's sending the message, just like IM. Most phones these days have extremely good predictive spelling, which actually makes it easier to just spell out a word than abbreviate it more often than not.
Calling SMS a 'code' is a bit off-target, IMO. SMS simply stands for "simple message system." Calling it a code is the Morse equivalent of calling the telegraph a code. It's the medium, not the data.
That was exactly my thought when I read this. Even if Debian is planning on moving to a much faster release schedule, this goes directly against the extremely stable argument I often get in Debian's favor.
I hate to hold Ubuntu up as an example every time Debian comes up, but they've at least got the right idea: A new release every 6 months, supported for 18 months. They also have plans for an 'Enterprise' release every 12-24 months, sporting Debian-like stability testing.
Personally, I'd argue that even 18 months is a bit on the short side for a production environment, but it's certainly better than 12. This isn't even taking into account that there's also a stable release pattern, and since 3 releases should be supported at any given time, there's nothing to stop you from picking one in the 'middle' if you prefer tried and true to the leading edge.
I was confused by the fact that you were replying to a post on the merits of Linux-style security then, my bad.
:-)
I do stand by the fact that I see nothing particularly special about OS X style permission escalation within the Unix world. In fact, it essentially just puts a pretty face on sudo. And I'm well aware that Mandrake has used PAM and the like for years now (MDK was actually my first distro, back in the 6.5 release). I pointed out Ubuntu in particular because it goes a step farther (as does OS X) by disabling the root account by default. This prevents users from running as root entirely, thus minimizing the "I drug my System folder to the trash" problems.
My point was primarily that OS X style priveledge assignment works well because it allows a user to run as a limited account by default without it being a complete pain, while still leaving the bar to screwing up the system high enough that it generally won't happen in normal usage. I got confused by the "OS X's security seems to work better" nature of your post since you were replying to someone making a similar point about Linux. Again, my bad.
You may want to check out Ubuntu, if you haven't. It uses pretty much the exact same user authentication scheme as OS X. Root account disabled by default, escalated privs only when necessary. Updates are handled by an applet, all the admin tools will prompt for a root pass when needed, and most OSS software can be installed from Synaptic, the GUI package manager.
You're actually the first person I've heard complain about nforce2 video having problems. I've built quite a few systems based on nForce2 IGP chipsets, and haven't ever had a problem. Out of curiousity, which boards were you using? The one place I have had major headaches in the past is with nForce2-based MSI boards and drivers, but all of those were non-IGP.
Upgrade your video card every 6 months? Hardly. My current gaming PC is well over 2 years old at this point, and while I've added storage and optical drives along the way, the core system is still the same as the day I bought it - an Athlon XP 2600+ with a gig of RAM and a Radeon 9700.
While it wasn't a budget box at the time, it wasn't extraordinarily expensive either. Something to the effect of about $700. And I've yet to find a game that won't run acceptably well on it, up through all the latest - Doom 3, HL2, Far Cry, etc. Personally, I'm not even *considering* an upgrade until the A64 X2s are priced reasonably enough for me to afford one.
I'm certainly not arguing that raw hardware costs are lower for a PC - they're not. However, the idea that it gets incredibly expensive because you "have" to upgrade every 6 months is ludicrous. I haven't spent any more on the gaming aspects of my PC than I would have buying a couple of consoles (with no games, extra controllers, etc) at launch. 2+ years down the road and still kicking it? I can't complain about that.
Granted, the fact that I'm a geek and use my system for far more than just gaming, it's easier for me to justify, but all the same, I just fail to see where the "I have to spend $300 upgrading every 6 months just to keep up" mentality comes from. It just doesn't line up with the reality I've come to know in this field. Buy solid, middle of the line parts that are at whatever the price/performance sweetspot at the time is, and you'll be fine for quite some time.
Ubuntu uses
I beg to differ. Ubuntu, which takes flak for still using Debian's text based installer, has a much more straight-forward install than any version of Windows. Nevermind Mandriva, Fedora Core, Xandros, SuSE or any number of other distributions with fully graphic installers that many would consider even more friednly.
While your average user probably wouldn't install Linux on their own, I imagine it's more a matter of motivation than anything. I have no doubt that most people certainly could get Linux on their machine with most modern, user-oriented distributions if they were to put forth the effort, and they most certainly could do so easier than with Windows, whose installer is loaded with confusing and conflicting options.
Seriously. Look at the installers side-by-side. I've certainly spent enough time looking at them working in IT, and it's very clear to me which of them needs work.
This is so incredibly offbase I'm not even sure where to start, outside of saying that I use OS X, Windows, and Linux all on a practically daily basis.
That said, a user can trash an OS X install in manner approximately equal to that of their ability to trash a Linux install. A normal user account has access to their home directory, and can essentially do what they want, up to and including trashing everything in it.
Some distros (namely Ubuntu) have taken up OS X style password prompting with root disabled by default, which I definitely think is rather important, even if you downplay it. It's one of the things about OS X that was well thought out, and it's good to see Linux distros adopting a similar system. The reason so many Windows users run as admin is two fold - for one some things just plain don't work. But the other is even when things work, it's a pain when you do need admin privs. By making escalating priveleges no more difficult than a simple prompt, users are encouraged to run limited accounts.
Not to say it's an infallible system. Remember the fake Office 'trial' for OS X that was going around on P2P a few months back? It'd prompt a user for their admin password, then go and helpfully start trashing files. Moral of the story? Any system is susceptible to user stupidity in the end, it's just a matter of how easy it is for them to trash it.
Honest question: Have you evaluated Ubuntu for this purpose? The reason I ask is because for the administration tasks a non-technical user might need to worry about (read: security updates and the like) it's a bit more user friendly, while maintaining a Debian base.
Not meaning to start a (meta)distro flamewar here, just genuinely curious since Knoppix HD installs always seemed somewhat like a hack to me.