No, we don't believe three HDs all randomly failed within two days of each other, either.
In my experience, the most likely time for additional drives to fail is during a RAID rebuild, because it's usually the first time for a long while that every part of each disks surface has been touched.
I was rather referring to the desktop/home user segment within the last 15 years or so (you know, the computers that are relavent to the story).
Right, so just x86, PPC and 68k.
And x86-64 is just an extension of x86 that lets you use more memory, it's not really a different architecture. More complex, perhaps, but not a full rewrite.
When you're talking about low level OS code, I suspect it's fairly different.
No peripheral manufacturere bothered with making USB devices. After 2 years on the market, there were only 12 usb devices. Then Apple added USB. 6 months later, there were over 400 usb devices.
And you don't think the near simultaneous release of Windows 98 might have had just the slightest impact ?
Why should the MacBook be any faster then any other DuoCore notebook out there.
It's not. The delta between the fastest PC laptop and the MacBook was *1 second*, representing aproximately a 1% - 1.5% difference. That's a rounding error, not a "win".
(Additionally, the "Windows Media encode" was a somewhat more significant 7% slower on the MacBook Pro.)
Of more interest was the fact that the slower PCs were on the order of *40 seconds* slower - although they also all have basically the same times (1:52, 1:52, 1:55), so whatever it is that is making them slower is the same on all of them (my guess would be the hard disk, but without knowing anything specific about the machines, it's hard to say for sure). Strangely, all their "Windows Media Encode" times are basically the same - so whatever is slowing down the PS benchmarks isn't having the same effect on their media encoding speed.
I just can't see how the same equipment can run better on one system over another.
Things like memory timings and different hard disks can make machines with identical processors have *substantially* different performance characteristics (memory timings much less so than hard disks). There's also things like video cards, but none of the tests here would be meaningfully affected by the machine's video card.
But OSX has as least as much Mach lineage as it has BSD. It's often characterized as a bloated train wreck of an OS, but Apple has been able to get decent performance by controlling the hardware platform (they don't have to attempt to optimize for everything on the market, they can focus on being really good on optimal equipment).
OS X's performance is pretty poor, with everything from raw numbers (process creation times, disk I/O - anandtech did a bunch of benchmarking comparing to Linux) through to subjective "teh snappy" observations of GUI responsiveness.
I used to think this was just the sucky G4 platform holding everything back, but even G5 Macs are comparitively slow running OS X (I have not yet had a chance to sit down for a decent length of time with an x86 Mac, but I doubt it would make much difference, as the G5 platform was quite good). I think they just need to buckle down and do some hardcore low-level systems optimisations, rather than flashy new features (and probably will, for the release after 10.5 - OS X is maturing and new features/fixing obvious user-visible problems is/are becoming less important).
Apple has a tendency to heavily customize their machines, [...]
Not for a very long time. All remotely current Macs are basically just PCs. Even the PPC models are basically just PCs with a different motherboard and CPU. Everything else is off-the-shelf (which is why that whole "Apple uses better hardware" meme is so laughable).
[...] and one of their selling points is a tight coupling between hardware and software (namely, OS X.)
There is no "tight coupling" per se (as OS X x86 running on non-Apple PCs has demonstrated). What there is, is a development process that is able to focus on a very small set of hardware to support and optimise for, and very effectively identify any hardware issues that might exist, then get them fixed (or develop workarounds).
Running these benchmarks also allowed a direct comparison between Apple hardware and other manufacturers' that always used to be cloaked a little by the difference in OSes. Now of course you can argue that the driver situation may have affected our results, but I hope this will be only the first of many data points. It's a start.
What will be more interested is comparing the performance between the OSes themselves, particularly in situations where the OS will actually have a non-trivial impact (ie: multiple processes with heavy I/O and large memory footprints).
But the news in my mind isn't a one-second difference in this or that. It's that Apple's machines run Windows comparably to the best designed-for-Windows machines.
I fail to see why anyone with anything past a basic acquaintance with hardware would find this surprising, when talking about the MacBook Pro.
I could see some surprised regarding the Mini, or iMac, for those who have been around a while, as Apple has a long and glorious history of crippling their lower-end products to "encourage" customers to buy the higher end stuff, but the MacBook Pro *is* a high-end Mac. Apple always throw the best stuff they can into them.
Because lots of people -- namely, many Windows and Mac users -- are scared of free software. They're expecting some kind of "catch" (time-expiring demos, nagware, spyware, etc.) so they stay away because it's free.
It always amuses me that the assumption made why "normal people" don't like free software is because it's free, or because it being free "scares them".
(Particularly since this is the complete opposite to what happens with just about everything else in the world that's free.)
Have y'all considered that maybe "normal people" don't like free software because it's not as good, doesn't do what/anything they need, or they simply don't know about it ?
Microsoft has a long history of manipulating hardware manufacturers. They effectively control what Dell and others put in their computers. They are control freaks and will not be satisfied with a manufacturer they can't influence, such as Apple.
Right, that must be why they don't sell a version of Windows to the general public, because then people would be able to install it onto any hardware they want.
Oh, wait, they *do*. It's *Apple* who tell you what machines you can and can't install their OS on.
According to MS, it's so tightly integrated that it's impossible to remove it from the OS. Since it's not open source, I can't tell you how tightly integrated it is, but based on statements like that, i'd say it's pretty well integrated.
You do realise that by "OS" they mean the *entire package* and not just the kernel, right ?
Rather large chunks of the shell - not to mention numerous third party applications - depend on IE. If it is removed, then those things break. Hence, the OS breaks. Just like if you rip, say, glibc or any of a dozen other libraries out of the typical Linux distro, the OS breaks.
Architecturally, IE is basically identical to KDE/Konquerer.
The point is, Honda makes plain Jane, reliable cars. (Painting them neon green and putting a giant wing on a front wheel drive car doesn't make them non-vanilla)
Say what ?
* NSX
* S2000
* CBR1000RR, VTR1000SP-2 and CBR600RR
* Not to mention their success in Formula 1.
Just because you're only familiar with Ricers and soccer mums doesn't mean that's all Honda makes.
What really struck me about this is that Microsoft can make a horrible design decision, at least from a security point of view, continue making that mistake for 10 years, and it doesn't dent their market share.
What's really struck me over the years is how KDE, GNOME and OS X all went on to copy the same design, yet Windows is the only one that gets criticised for it...
I doubt though that something so integrated into windows explorer can be seperated and reprogrammed into a seperate application within the extra 2 months.
IE is not, and never has been, "integrated into Windows Explorer". It has *always* been a bunch of separate components that Explorer loads when necessary to display certain types of information. It's just like KDE/Konquerer/KHTML.
Riddle me this, Microsoft stated (under oath) IE was a core Windows component and couldn't be removed. So why can they remove it now? The Hal hasn't changed. Win32 hasn't changed?
The *shell* has changed. The shell being a rather important part of an OS.
Konqueror is not a file manager or a web browser, in the strictest sense. Konqueror is just a container that runs KParts. That's all. There is a file management KPart, and a Web Browsing KPart, which is what most people use by far. But really konqueror is just a shell that loads whatever you want it to. The KHTML KPart is no more 'integrated' into KDE or the OS than the PDF KPart is, or the MPlayer KPart.
Congratulations, you have just described how Windows works, as well.
This is so incorrect I don't even know where to begin. SLASHBOTS: how can you mod such blatantly wrong shit up so high?
Because the fact every major platform (KDE, GNOME, OS X) has gone on to copy Windows's "browser component" design causes a segfault in their tiny "Windows sucks, not Windows rules" logic centres.
I know I know, not entirely accurate but that's the closest thing I could come up with to Windows' architecture.
The "closest thing you could up with" would be "IE and Konquerer are basically the same".
IE does not run with elevated privileges. It does not not have secret hooks into the kernel. It does not have magical powers to circumvent OS security. It's just regualr old user-space code. IE does not let a user - or anything they run - do anything they/it would not otherwise be able to do.
Slashdot Mirror
In my experience, the most likely time for additional drives to fail is during a RAID rebuild, because it's usually the first time for a long while that every part of each disks surface has been touched.
Right, so just x86, PPC and 68k.
And x86-64 is just an extension of x86 that lets you use more memory, it's not really a different architecture. More complex, perhaps, but not a full rewrite.
When you're talking about low level OS code, I suspect it's fairly different.
And you don't think the near simultaneous release of Windows 98 might have had just the slightest impact ?
It's not. The delta between the fastest PC laptop and the MacBook was *1 second*, representing aproximately a 1% - 1.5% difference. That's a rounding error, not a "win".
(Additionally, the "Windows Media encode" was a somewhat more significant 7% slower on the MacBook Pro.)
Of more interest was the fact that the slower PCs were on the order of *40 seconds* slower - although they also all have basically the same times (1:52, 1:52, 1:55), so whatever it is that is making them slower is the same on all of them (my guess would be the hard disk, but without knowing anything specific about the machines, it's hard to say for sure). Strangely, all their "Windows Media Encode" times are basically the same - so whatever is slowing down the PS benchmarks isn't having the same effect on their media encoding speed.
I just can't see how the same equipment can run better on one system over another.
Things like memory timings and different hard disks can make machines with identical processors have *substantially* different performance characteristics (memory timings much less so than hard disks). There's also things like video cards, but none of the tests here would be meaningfully affected by the machine's video card.
OS X's performance is pretty poor, with everything from raw numbers (process creation times, disk I/O - anandtech did a bunch of benchmarking comparing to Linux) through to subjective "teh snappy" observations of GUI responsiveness.
I used to think this was just the sucky G4 platform holding everything back, but even G5 Macs are comparitively slow running OS X (I have not yet had a chance to sit down for a decent length of time with an x86 Mac, but I doubt it would make much difference, as the G5 platform was quite good). I think they just need to buckle down and do some hardcore low-level systems optimisations, rather than flashy new features (and probably will, for the release after 10.5 - OS X is maturing and new features/fixing obvious user-visible problems is/are becoming less important).
Apple has a tendency to heavily customize their machines, [...]
Not for a very long time. All remotely current Macs are basically just PCs. Even the PPC models are basically just PCs with a different motherboard and CPU. Everything else is off-the-shelf (which is why that whole "Apple uses better hardware" meme is so laughable).
[...] and one of their selling points is a tight coupling between hardware and software (namely, OS X.)
There is no "tight coupling" per se (as OS X x86 running on non-Apple PCs has demonstrated). What there is, is a development process that is able to focus on a very small set of hardware to support and optimise for, and very effectively identify any hardware issues that might exist, then get them fixed (or develop workarounds).
Running these benchmarks also allowed a direct comparison between Apple hardware and other manufacturers' that always used to be cloaked a little by the difference in OSes. Now of course you can argue that the driver situation may have affected our results, but I hope this will be only the first of many data points. It's a start.
What will be more interested is comparing the performance between the OSes themselves, particularly in situations where the OS will actually have a non-trivial impact (ie: multiple processes with heavy I/O and large memory footprints).
I fail to see why anyone with anything past a basic acquaintance with hardware would find this surprising, when talking about the MacBook Pro.
I could see some surprised regarding the Mini, or iMac, for those who have been around a while, as Apple has a long and glorious history of crippling their lower-end products to "encourage" customers to buy the higher end stuff, but the MacBook Pro *is* a high-end Mac. Apple always throw the best stuff they can into them.
It always amuses me that the assumption made why "normal people" don't like free software is because it's free, or because it being free "scares them".
(Particularly since this is the complete opposite to what happens with just about everything else in the world that's free.)
Have y'all considered that maybe "normal people" don't like free software because it's not as good, doesn't do what/anything they need, or they simply don't know about it ?
Right, that must be why they don't sell a version of Windows to the general public, because then people would be able to install it onto any hardware they want.
Oh, wait, they *do*. It's *Apple* who tell you what machines you can and can't install their OS on.
Of course it's not. That's the kind of thing Apple does.
If it's not in there today I can image a new paragraph in the near future.
For what possible reason ?
Yeah, except for all the software they wrote for x86-64, Itanium, PPC, Alpha, MIPS, Moto 68k...
You do realise that by "OS" they mean the *entire package* and not just the kernel, right ?
Rather large chunks of the shell - not to mention numerous third party applications - depend on IE. If it is removed, then those things break. Hence, the OS breaks. Just like if you rip, say, glibc or any of a dozen other libraries out of the typical Linux distro, the OS breaks.
Architecturally, IE is basically identical to KDE/Konquerer.
Thank you for restoring a tiny bit of faith in Slashdot for me.
False. IE is completely userspace.
Say what ?
* NSX
* S2000
* CBR1000RR, VTR1000SP-2 and CBR600RR
* Not to mention their success in Formula 1.
Just because you're only familiar with Ricers and soccer mums doesn't mean that's all Honda makes.
The situation in Windows is exactly the same.
What's really struck me over the years is how KDE, GNOME and OS X all went on to copy the same design, yet Windows is the only one that gets criticised for it...
IE is not, and never has been, "integrated into Windows Explorer". It has *always* been a bunch of separate components that Explorer loads when necessary to display certain types of information. It's just like KDE/Konquerer/KHTML.
Please expand on what "integrated so tightly with the Operating system" means to you.
The *shell* has changed. The shell being a rather important part of an OS.
Which is basically what IE *is* (strictly speaking it's a COM component, but the principle is the same).
Not in unix, true - but Windows is not unix.
Congratulations, you have just described how Windows works, as well.
Because the fact every major platform (KDE, GNOME, OS X) has gone on to copy Windows's "browser component" design causes a segfault in their tiny "Windows sucks, not Windows rules" logic centres.
The "closest thing you could up with" would be "IE and Konquerer are basically the same".
IE does not run with elevated privileges. It does not not have secret hooks into the kernel. It does not have magical powers to circumvent OS security. It's just regualr old user-space code. IE does not let a user - or anything they run - do anything they/it would not otherwise be able to do.