XServe is still PPC, so XServe's kernel (XNU PPC) is still avaliable. Don't put the cart before the horse
It doesn't matter if the XServe kernel is available or not. Mach based kernels are not going to ever be speed demons: Mach's the posterboy for "how not to implement a microkernel" and so Mac OS X has simply got a lot more overhead than FreeBSD or Linux.
If you really want to get high performance out of an Apple server... regardless of whether it's a G5 or a G5 or Intel Core... you're not going to be tweaking XNU... you're going to be replacing the kernel completely with something a lot lighter.
Anyone who thinks that Mac OS X is capable of competing with FreeBSD where FreeBSD excels is either suffering from concussion or has never actually used Mac OS X and FreeBSD as servers.
The point is that by going with the BSD license, the freebsd team has made sure there is a propriety OS which will always be better than theirs.
FreeBSD is still far superior to Mac OS X as a server. OS X is much slower and bulkier, it has no standard UNIX tape support, it doesn't implement softupdates or jails or pretty much any of the other features of FreeBSD that make it a great server OS. It's only on the desktop that Mac OS X shines.
And FreeBSD is a dull gem indeed as a desktop OS. If you're looking for an all-out best-of-the-breed open source desktop, even a total BSD fan like me isn't going to recommend FreeBSD. Hell, even when you included Mac OS and Linux I was still telling most people "well, you're probably best off with Windows, unfortunately" right up until OS X 10.2 Jaguar came out.
So rather than saying "the freebsd team has made sure there is a propriety OS which will always be better than theirs", try "the freebsd team has made sure there is a proprietary UNIX which is good enough as a server, and better than Windows as a desktop". Which is a win-win situation for this long term FreeBSD developer, thank you very much.:)
Then why the big hub-bub about including ATI and NVidea drivers is various linux distros?
I didn't say "integrating proprietary code into a GPLed kernel is not an issue".
I said "Bundling GPLed code with a non-GPLed kernel or OS is not an issue".
The GPL doesn't cover "simple bundling". It even permits linking between GPLed and proprietary code so long as you're using open APIs that have appropriately licensed versions, so that the proprietary code is a derivitive work of the open API as opposed to a GPLed one. So gcc bundled with Mac OS X using UNIX APIs to call the OS doesn't require the OS to be GPLed.
The ATI and nVidia drivers, though, need to use Linux-only APIs, and that makes them derivitive works... so they need to be GPLed.
I think the MPAA had as much to do with this as Maxxus.
I expected this as soon as I heard the Intel hardware DRM support was included. As hard as strong DRM is, strong DRM on top of an open source kernel where an attacker can plant a "tap" anywhere under the user-kernel boundary out of sight of the application is virtually impossible.
So... the iTunes Video Store would ever be able to ship first-tier movies until Apple had some kind of equivalent of Microsoft's "trusted path", and that ain't going to happen with an open source kernel. Expect it in Leopard, for Intel Macs only.
Is this new Darwin version COMPLETELY new, with no trace of GPL-protected code?
If you look at the Darwin source tree you will see that there are a number of components with "Other" licenses. Any GPL or LGPL code will be in these trees. A quick look through the first few "Other" entries finds Chess, JBoss, and MySQL, and of course gcc and its component parts are GPLed.
But just because something's GPLed doesn't mean it can't be used by or distributed with a proprietary kernel or OS. After all, Microsoft is inlcuding Interix in Windows Vista, and Interix ships with many of the same GPLed components as Darwin and Mac OS X such as gcc...
If FreeBSD was GPL Darwin wouldn't have existed. Apple already had a licenced commercial UNIX in NeXTstep, and they had other options like BeOS (which was already running on the PowerMac and supported Mac OS 8 hosted under the amusingly named Sheepshaver). There's approximately zero percent chance that they would have gone with a GPL kernel, whether BSD or Linux.
So what you're really saying, based on all your comments attached to this story, is that OS X was never really suited to high-performance/server/scientific computing?
Anyone who's used an XServe knows that. It's not and never has been a high performance beast, it's a decent small server that's targeted towards Windows-class admins with fewer Windows-class annoyances.
Yeh, without BSD licensing they'd probably have had to go with something like BeOS for the kernel, or just keep using the NeXTstep kernel they already had an irrevocable license from AT&T for. It's not like the only alternatives were ever FreeBSD or Linux.
My friend once calculated that it wasn't worth bill gates' time to bend over to pick up a $100, since the time he'd waste doing that wouldn't exceed what he'd make in that time just 'doing his job'
Of course he'd still be making all that money 'just doing his job' whether he picked up the C-note or not.
(Re: That's just a fancy version of "the word on the street".)
No, consumer reports is a based on a statistically representative survey.
I used to be a Consumers Union member, and took part in those surveys. I know exactly what they are. And how reliable they are. For things like washers and dryers and refrigerators they're great, but for cars and computers they're all over the place because there's a lot of people who make a significant emotional commitment to these kinds of purchases and that has a huge effect on the CR ratings.
The bottom line for CR is that it's a self-selected survey of what "the word on the street" is. It's not based on return rates, or repair rates, because Consumer Reports doesn't have access to that kind of manufacturer data. The difference between a CR rating and googling blogs for comments is quantitative, not qualitative. They get *more* responses from users, and they do a better job of compiling the results, but they're still just telling you what the company's reputation for quality is... they're not measuring the company's actual product quality, and even if they were the average product quality doesn't tell you whether a particular product is a "lemon" or a "road apple".
Where else can you get a Core Duo laptop from a premium brand for $1099?
The Core Duo is Intel's response to the dual-core Athlons. Dual core or single core, Athlon or Intel Core or even mobile Pentium, a comparable laptop is one with comparable features and performance regardless of the chipset.
You can get second mouse button functionality by touching two fingers on the track pad and clicking.
*boggle*
That sounds actually less convenient than control-click.
That's like the stupid "Mighty Mouse". Yes, you can right-click... but I am unable to actually right-click one without stopping and making sure my index finger isn't resting on the mouse anywhere so the capacitance sensor isn't confused.
It also means I still can't chord left-and-right buttons for the "middle click" in X11 applications.
Apple has no freaking clue, and the ridiculous levels they're going to to avoid just putting another five cent microswitch in the laptop are embarassing.
Telling everybody there was no need for stuff like SELinux because correctly configured apps and OS's will work.
First, that begs the question: What is SELinux but another attempt at creating a "correctly configured OS"? Having a bunch of fine-graned permissions and even mandatory access controls (though I'm far from convinced that MAC is actually suitable for this job... have you used a system with real orange-book MAC implemented?) that have to be "correctly configured" to allow users to continue to work and actually improve security doesn't get you anything by itself.
Second, what I'm saying is that the value of time spent on using the tools already available is far far greater than the value of making it slightly harder to hide an exploit from an average user, and is far more likely to be accepted by the average user than a bunch of low level restrictions. There is an immediate need for, and an immediate security win from, decreasing the surface area available for attack in the first place. That's what we need right now, not a bunch of extra security levels that will take years to be properly configured so they can be turned on by default.
MAC provides a good low layer protection to keep a vulnerability in some software from becoming root compromise.
In a single user system the difference between a root compromise and a non-root compromise is a lot less relevant than the difference between no compromise and a remote user compromise.
And... how would you apply MAC to prevent a root compromise?
Simply put: the word on the street isn't necessarily a reliable indicator of reality.
Indeed.
Apple's got a ridiculously high customer satisfaction rating from consumer reports.
That's just a fancy version of "the word on the street".
Again, I did not say "apples suck because they have a rep for problems". I said "you can't use their reputation as evidence that this is a good product" because every company has models that are lemons (or, in the case of Apple, "road apples"). You have to look at the particular model, and this particular model is not a superior laptop for a bargain price, it's a typical laptop for a higher than comparable price.
But every survey conducted objectively (Consumer Reports, PC Magazine, etc) that looks at large numbers of customers says unconditionally that Apple is one of the best, if not THE best, manufacturers of computer hardware.
That's what "reputation" means... a lot of people think good things about it. Reputation is a great tool when deciding what products you're going to look into.
My point isn't that the OP was wrong about their reputation for having great hardware. It's that the fact that they have a reputation for great hardware doesn't mean that any particular machine is great. Objectively, many of the machines that got great reports in magazines turned out to be duds. The first generation Powermacs, for example, are now widely considered "Road Apples" in retrospect. The later consumer-grade Powermacs like the 6400 also turned out to be dead-end machines. The first revision of the Beige G3s, the Rev A-D iMacs, the "Yikes" G4, and so on.
For that matter, the "Cube" is still widely considered a great computer by many people, but objectively the cooling problems and resulting poor expansion capability make it a poor buy... and if you were unlucky enough to get the first-generation monitor with it you ended up losing out both ways, because they don't seem to have came out with any video cards after the Rage that could power an analog ADC monitor.
I believe that the first generation low-end Intel macs with the Intel graphics chips are going to end up in the same "Road Apple" category as the Powermac 6100 and 7100 or the "Yikes" G4. I may be wrong, but the "overall reputation" of the Apple product line... or even the reputation of specific products... is neither proof nor evidence of that.
Apple is selling a product that has the reputation of Thinkpads and VAIOs, in a price range right around that of Dell and HP...
Apple is selling a product that has the reputation of Gateways and Packard Bells, if you listen to people who've had hinge/connector/power/heat problems in the past.
Reputation is a useful guide for deciding what to evaluate, but you have to evaluate what they're actually selling... not what the rumor mill says about them.
The $500 mini was a pretty good deal, for a year ago.
Then, Apple's requirements changed with the introduction of Tiger and the logical upgrade would have been a Radeon 9600 (like the contemporary eMac) or nVidia g05200 (like the contemporary iBook). Without that, it was an OK deal... but it was rational to wait for them to upgrade it as a replacement for the eMac, and match the iBook with one or the other of the fully supported GPUs.
Then came the intel announcement.
So... anyway...
One would expect that a new version of the mini would actually be an upgrade over the original, and at a comparable or lower price.
If the new mini was $500 with the Core Solo and an ATI Radeon 9600 or nVidia go5200, that would be an OK deal. Or, if they were going to shove an intel GPU in it, then a Core Duo for $500 would be an OK deal: it's doing the missing 3d openGL calls in software and when that's going on it uses up pretty much a whole core,... and while it's faster overall that's balanced by the fact that there's less usable RAM.
But paying half as much again and then having to buy extra RAM?
I don't think so. No, that's not what I call "an OK deal" at all.
"Choose iLife. Choose a blog. Choose a podcast. Choose a Macbook. Choose a fucking big hard disk, choose washing machines, cars, cable modems and electrical tin openers. Choose good health, low cholesterol, and a 21" LCD. Choose fixed interest mortage repayments. Choose an ISP. Choose your friends. Online. Choose Livejournal and.Mac. Choose an office suite from Microsoft with a range of fucking security holes. Choose WoW and wondering who the fuck you are on a Sunday morning. Choose sitting on that Aeron chair playing mind-numbing, spirit-crushing MMORPGs, stuffing fucking junk food into your mouth. Choose rotting away at the end of it all, pishing your last in a miserable home, nothing more than an embarrassment to the selfish, fucked up brats you spawned to replace yourself."
I remember trying a T23 keyboard and not being that impressed with it.
I love my T23.
The keys have a reasonable amount of travel given how flat it is, are nicely shaped, and have a positive feel. I'd prefer a keyboard like the mid-90s Toshibas, but nobody makes full-sized keys for laptops any more because they'd add more than half a cm to the thickness of the laptop. Not that they couldn't afford it.:P
I have no idea about the T30, and it's not really the sag in the keyboard that bothers me.
The Powerbook keys are flat, have almost no travel, don't give you any real feedback, and aren't stable... they don't sag but they slip a little sideways. I have this creepy feeling that I'm typing on a dead alligator and the scales are about to come loose.
The part that says "Apple's shipping Windows XP on these babies", because if they're shipping them with Mac OS X they need a real GPU. Mac OS X graphics are built on top of OpenGL, and use the GPU heavily.
The part that says "consumers don't play games".
Seriously. A businessman's less likely to miss the GPU than a "consumer".
If you want whiz-bang GPU, they've got a nice shiny MacBook Pro to sell you.
I don't want a "whiz-bang GPU". Any current ATI or nVidia would do nicely, thanks.
Cause 10.4.6 runs just fine on my 6 year old, 500 MHz G3 laptop with 512 megs of RAM.
Yeh, but on an Intel chip you're losing a bunch of that RAM to Rosetta.
And on the Mac mini and this new laptop you're losing 64M to the GPU, and even more to the software OpenGL Apple has to use to fill in for the shortcomings of the Intel graphics chip.
So it's not really 512MB... it's 448M to start, and you're losing maybe 100M of that to emulation...
Am I the only one who thinks 1280 is ridiculous for a 13.3" screen width?
It's not great, but it beats the pants off the crappy 1024 they used to have.
1152 is the absolute minimal acceptable screen width with modern applications, at least until they get the GUI to really take advantage of OpenGL's 3d support and automatically rescales text and icons.
Slashdot Mirror
XServe is still PPC, so XServe's kernel (XNU PPC) is still avaliable. Don't put the cart before the horse
It doesn't matter if the XServe kernel is available or not. Mach based kernels are not going to ever be speed demons: Mach's the posterboy for "how not to implement a microkernel" and so Mac OS X has simply got a lot more overhead than FreeBSD or Linux.
If you really want to get high performance out of an Apple server... regardless of whether it's a G5 or a G5 or Intel Core... you're not going to be tweaking XNU... you're going to be replacing the kernel completely with something a lot lighter.
He may have hit something with his head.
Anyone who thinks that Mac OS X is capable of competing with FreeBSD where FreeBSD excels is either suffering from concussion or has never actually used Mac OS X and FreeBSD as servers.
The point is that by going with the BSD license, the freebsd team has made sure there is a propriety OS which will always be better than theirs.
:)
FreeBSD is still far superior to Mac OS X as a server. OS X is much slower and bulkier, it has no standard UNIX tape support, it doesn't implement softupdates or jails or pretty much any of the other features of FreeBSD that make it a great server OS. It's only on the desktop that Mac OS X shines.
And FreeBSD is a dull gem indeed as a desktop OS. If you're looking for an all-out best-of-the-breed open source desktop, even a total BSD fan like me isn't going to recommend FreeBSD. Hell, even when you included Mac OS and Linux I was still telling most people "well, you're probably best off with Windows, unfortunately" right up until OS X 10.2 Jaguar came out.
So rather than saying "the freebsd team has made sure there is a propriety OS which will always be better than theirs", try "the freebsd team has made sure there is a proprietary UNIX which is good enough as a server, and better than Windows as a desktop". Which is a win-win situation for this long term FreeBSD developer, thank you very much.
Then why the big hub-bub about including ATI and NVidea drivers is various linux distros?
I didn't say "integrating proprietary code into a GPLed kernel is not an issue".
I said "Bundling GPLed code with a non-GPLed kernel or OS is not an issue".
The GPL doesn't cover "simple bundling". It even permits linking between GPLed and proprietary code so long as you're using open APIs that have appropriately licensed versions, so that the proprietary code is a derivitive work of the open API as opposed to a GPLed one. So gcc bundled with Mac OS X using UNIX APIs to call the OS doesn't require the OS to be GPLed.
The ATI and nVidia drivers, though, need to use Linux-only APIs, and that makes them derivitive works... so they need to be GPLed.
I think the MPAA had as much to do with this as Maxxus.
I expected this as soon as I heard the Intel hardware DRM support was included. As hard as strong DRM is, strong DRM on top of an open source kernel where an attacker can plant a "tap" anywhere under the user-kernel boundary out of sight of the application is virtually impossible.
So... the iTunes Video Store would ever be able to ship first-tier movies until Apple had some kind of equivalent of Microsoft's "trusted path", and that ain't going to happen with an open source kernel. Expect it in Leopard, for Intel Macs only.
Is this new Darwin version COMPLETELY new, with no trace of GPL-protected code?
If you look at the Darwin source tree you will see that there are a number of components with "Other" licenses. Any GPL or LGPL code will be in these trees. A quick look through the first few "Other" entries finds Chess, JBoss, and MySQL, and of course gcc and its component parts are GPLed.
But just because something's GPLed doesn't mean it can't be used by or distributed with a proprietary kernel or OS. After all, Microsoft is inlcuding Interix in Windows Vista, and Interix ships with many of the same GPLed components as Darwin and Mac OS X such as gcc...
If FreeBSD was GPL Darwin wouldn't have existed. Apple already had a licenced commercial UNIX in NeXTstep, and they had other options like BeOS (which was already running on the PowerMac and supported Mac OS 8 hosted under the amusingly named Sheepshaver). There's approximately zero percent chance that they would have gone with a GPL kernel, whether BSD or Linux.
So what you're really saying, based on all your comments attached to this story, is that OS X was never really suited to high-performance/server/scientific computing?
Anyone who's used an XServe knows that. It's not and never has been a high performance beast, it's a decent small server that's targeted towards Windows-class admins with fewer Windows-class annoyances.
Yeh, without BSD licensing they'd probably have had to go with something like BeOS for the kernel, or just keep using the NeXTstep kernel they already had an irrevocable license from AT&T for. It's not like the only alternatives were ever FreeBSD or Linux.
My friend once calculated that it wasn't worth bill gates' time to bend over to pick up a $100, since the time he'd waste doing that wouldn't exceed what he'd make in that time just 'doing his job'
Of course he'd still be making all that money 'just doing his job' whether he picked up the C-note or not.
(Re: That's just a fancy version of "the word on the street".)
No, consumer reports is a based on a statistically representative survey.
I used to be a Consumers Union member, and took part in those surveys. I know exactly what they are. And how reliable they are. For things like washers and dryers and refrigerators they're great, but for cars and computers they're all over the place because there's a lot of people who make a significant emotional commitment to these kinds of purchases and that has a huge effect on the CR ratings.
The bottom line for CR is that it's a self-selected survey of what "the word on the street" is. It's not based on return rates, or repair rates, because Consumer Reports doesn't have access to that kind of manufacturer data. The difference between a CR rating and googling blogs for comments is quantitative, not qualitative. They get *more* responses from users, and they do a better job of compiling the results, but they're still just telling you what the company's reputation for quality is... they're not measuring the company's actual product quality, and even if they were the average product quality doesn't tell you whether a particular product is a "lemon" or a "road apple".
Where else can you get a Core Duo laptop from a premium brand for $1099?
The Core Duo is Intel's response to the dual-core Athlons. Dual core or single core, Athlon or Intel Core or even mobile Pentium, a comparable laptop is one with comparable features and performance regardless of the chipset.
You can get second mouse button functionality by touching two fingers on the track pad and clicking.
*boggle*
That sounds actually less convenient than control-click.
That's like the stupid "Mighty Mouse". Yes, you can right-click... but I am unable to actually right-click one without stopping and making sure my index finger isn't resting on the mouse anywhere so the capacitance sensor isn't confused.
It also means I still can't chord left-and-right buttons for the "middle click" in X11 applications.
Apple has no freaking clue, and the ridiculous levels they're going to to avoid just putting another five cent microswitch in the laptop are embarassing.
Telling everybody there was no need for stuff like SELinux because correctly configured apps and OS's will work.
First, that begs the question: What is SELinux but another attempt at creating a "correctly configured OS"? Having a bunch of fine-graned permissions and even mandatory access controls (though I'm far from convinced that MAC is actually suitable for this job... have you used a system with real orange-book MAC implemented?) that have to be "correctly configured" to allow users to continue to work and actually improve security doesn't get you anything by itself.
Second, what I'm saying is that the value of time spent on using the tools already available is far far greater than the value of making it slightly harder to hide an exploit from an average user, and is far more likely to be accepted by the average user than a bunch of low level restrictions. There is an immediate need for, and an immediate security win from, decreasing the surface area available for attack in the first place. That's what we need right now, not a bunch of extra security levels that will take years to be properly configured so they can be turned on by default.
MAC provides a good low layer protection to keep a vulnerability in some software from becoming root compromise.
In a single user system the difference between a root compromise and a non-root compromise is a lot less relevant than the difference between no compromise and a remote user compromise.
And... how would you apply MAC to prevent a root compromise?
Simply put: the word on the street isn't necessarily a reliable indicator of reality.
Indeed.
Apple's got a ridiculously high customer satisfaction rating from consumer reports.
That's just a fancy version of "the word on the street".
Again, I did not say "apples suck because they have a rep for problems". I said "you can't use their reputation as evidence that this is a good product" because every company has models that are lemons (or, in the case of Apple, "road apples"). You have to look at the particular model, and this particular model is not a superior laptop for a bargain price, it's a typical laptop for a higher than comparable price.
But every survey conducted objectively (Consumer Reports, PC Magazine, etc) that looks at large numbers of customers says unconditionally that Apple is one of the best, if not THE best, manufacturers of computer hardware.
That's what "reputation" means... a lot of people think good things about it. Reputation is a great tool when deciding what products you're going to look into.
My point isn't that the OP was wrong about their reputation for having great hardware. It's that the fact that they have a reputation for great hardware doesn't mean that any particular machine is great. Objectively, many of the machines that got great reports in magazines turned out to be duds. The first generation Powermacs, for example, are now widely considered "Road Apples" in retrospect. The later consumer-grade Powermacs like the 6400 also turned out to be dead-end machines. The first revision of the Beige G3s, the Rev A-D iMacs, the "Yikes" G4, and so on.
For that matter, the "Cube" is still widely considered a great computer by many people, but objectively the cooling problems and resulting poor expansion capability make it a poor buy... and if you were unlucky enough to get the first-generation monitor with it you ended up losing out both ways, because they don't seem to have came out with any video cards after the Rage that could power an analog ADC monitor.
I believe that the first generation low-end Intel macs with the Intel graphics chips are going to end up in the same "Road Apple" category as the Powermac 6100 and 7100 or the "Yikes" G4. I may be wrong, but the "overall reputation" of the Apple product line... or even the reputation of specific products... is neither proof nor evidence of that.
Apple is selling a product that has the reputation of Thinkpads and VAIOs, in a price range right around that of Dell and HP...
Apple is selling a product that has the reputation of Gateways and Packard Bells, if you listen to people who've had hinge/connector/power/heat problems in the past.
Reputation is a useful guide for deciding what to evaluate, but you have to evaluate what they're actually selling... not what the rumor mill says about them.
What does the new keyboard feel like?
Slippery/slidy with poor feedback like the Powerbook/Macbook Pro?
Wobbly with poor feedback like the iBook?
Still, if the $800 mini is an ok deal, then [...]
... and while it's faster overall that's balanced by the fact that there's less usable RAM.
The $500 mini was a pretty good deal, for a year ago.
Then, Apple's requirements changed with the introduction of Tiger and the logical upgrade would have been a Radeon 9600 (like the contemporary eMac) or nVidia g05200 (like the contemporary iBook). Without that, it was an OK deal... but it was rational to wait for them to upgrade it as a replacement for the eMac, and match the iBook with one or the other of the fully supported GPUs.
Then came the intel announcement.
So... anyway...
One would expect that a new version of the mini would actually be an upgrade over the original, and at a comparable or lower price.
If the new mini was $500 with the Core Solo and an ATI Radeon 9600 or nVidia go5200, that would be an OK deal. Or, if they were going to shove an intel GPU in it, then a Core Duo for $500 would be an OK deal: it's doing the missing 3d openGL calls in software and when that's going on it uses up pretty much a whole core,
But paying half as much again and then having to buy extra RAM?
I don't think so. No, that's not what I call "an OK deal" at all.
"Choose iLife. Choose a blog. Choose a podcast. Choose a Macbook. Choose a fucking big hard disk, choose washing machines, cars, cable modems and electrical tin openers. Choose good health, low cholesterol, and a 21" LCD. Choose fixed interest mortage repayments. Choose an ISP. Choose your friends. Online. Choose Livejournal and .Mac. Choose an office suite from Microsoft with a range of fucking security holes. Choose WoW and wondering who the fuck you are on a Sunday morning. Choose sitting on that Aeron chair playing mind-numbing, spirit-crushing MMORPGs, stuffing fucking junk food into your mouth. Choose rotting away at the end of it all, pishing your last in a miserable home, nothing more than an embarrassment to the selfish, fucked up brats you spawned to replace yourself."
Core Solo
Core Duo
Core 2 Duo
Core 2 Extreme
Good god almighty, they really want to bury the Pentium name.
But an x86 by any other name will smell the same.
I remember trying a T23 keyboard and not being that impressed with it.
:P
I love my T23.
The keys have a reasonable amount of travel given how flat it is, are nicely shaped, and have a positive feel. I'd prefer a keyboard like the mid-90s Toshibas, but nobody makes full-sized keys for laptops any more because they'd add more than half a cm to the thickness of the laptop. Not that they couldn't afford it.
I have no idea about the T30, and it's not really the sag in the keyboard that bothers me.
The Powerbook keys are flat, have almost no travel, don't give you any real feedback, and aren't stable... they don't sag but they slip a little sideways. I have this creepy feeling that I'm typing on a dead alligator and the scales are about to come loose.
Rosetta doesn't run all the time. If you aren't using any PPC apps then its not running.
But when (not if) you do, it is.
And there's still the GPU screw-up. Only 448M of RAM.
What part of "consumer level" do you not grok?
The part that says "Apple's shipping Windows XP on these babies", because if they're shipping them with Mac OS X they need a real GPU. Mac OS X graphics are built on top of OpenGL, and use the GPU heavily.
The part that says "consumers don't play games".
Seriously. A businessman's less likely to miss the GPU than a "consumer".
If you want whiz-bang GPU, they've got a nice shiny MacBook Pro to sell you.
I don't want a "whiz-bang GPU". Any current ATI or nVidia would do nicely, thanks.
Cause 10.4.6 runs just fine on my 6 year old, 500 MHz G3 laptop with 512 megs of RAM.
Yeh, but on an Intel chip you're losing a bunch of that RAM to Rosetta.
And on the Mac mini and this new laptop you're losing 64M to the GPU, and even more to the software OpenGL Apple has to use to fill in for the shortcomings of the Intel graphics chip.
So it's not really 512MB... it's 448M to start, and you're losing maybe 100M of that to emulation...
Am I the only one who thinks 1280 is ridiculous for a 13.3" screen width?
It's not great, but it beats the pants off the crappy 1024 they used to have.
1152 is the absolute minimal acceptable screen width with modern applications, at least until they get the GUI to really take advantage of OpenGL's 3d support and automatically rescales text and icons.
Gone are the days of 14" at 1024.
Yeah!
Why?
Because 1024 is nasty.