I would not put it past him to revive something very much like it.</strong><hr></blockquote>
Your first line I assumed was sarcasm since Steve quickly axed the Newton (and the eMate which ran the same OS) but you second line sounded serious. I find it hard to believe he would revive something like the eMate. Was I just missing a little more sarcasm?
[quote]Your first line I assumed was sarcasm since Steve quickly axed the Newton (and the eMate which ran the same OS) but you second line sounded serious. I find it hard to believe he would revive something like the eMate. Was I just missing a little more sarcasm?<hr></blockquote>
Nope. Amorph is right-- Steve did like the eMate, which is why his rolling Newton, Inc. back into Apple and then killing the entire platform wholesale was a bit of a surprise. And to think Newton, Inc. was about to post it's first profitable quarter ever...
<strong>Who wouldn't buy a new PowerMac with an IBM 970 running OS 10.2? It will still run 64 bit applications, even if the OS is 32 bits, and we know that OS X will soon be upgraded to 64 bits too.</strong><hr></blockquote>
This is incorrect -- if the OS isn't "64-bit" then it cannot present any 64-bit APIs or even a 64-bit address space to any 64-bit applications. The OS support for 64-bitness must be in place before any 64-bit applications can be created for it. The cart before the horse, and all that.
For what its worth I think we'll see 10.3 (64-bit) and the 970 hardware appear at the same time. There might be a few limitations on what 64-bit apps can have access to, but they'll deliver. Apple has actually had a pretty good software release track record since 10.1 arrived... at least compared to their history (and most of the rest of the industry).
<strong>Your theory hinges on the difficulty of moving OSX to 64bit.</strong><hr></blockquote>
Actually, I'd originally stated that I know many people have said that bringing OS X to 64 bits won't be that hard.
But even if all you had to do was change five #define lines and recompile, you can bet that there'd still be at least three months of internal QA work, and then a beta test phase, to make sure those little changes didn't result in a subtle bug or two or cause some instability.
Plus, as a marketing and as a philosophical milestone, going to 64-bit support is enough of a big step, regardless of the technical difficulty (or lack thereof) of doing so, that I think the change would warrant a major version change -- to 10.3. This would then tie 64-bitness together with who knows what other features are planned for 10.3.
64-bitness might be easy, but better Windows file sharing, fixing Finder bugs (still a few of those in Jaguar!), or the tele-olfactory version of iChat might be running behind schedule. That would mean no 64-bit support until the whole enchilada is ready to go.
This is why I postulate the idea of a 10.2.x release that does nothing more than tweak the OS to recognize the 970 and makes sure it runs in its 32-bit compatibility mode.
This might be better in the Looprumors 970 thread, particularly because this thread seems to be wondering toward another newton returns thread (and Fran isn't contributing? :eek: )
But I'll put it here anyway:
Just heard a rumor that 10.3 isn't scheduled until fall. This does make some sense with 10.2 coming out last fall, but it doesn't seem to jive with when 970 shipments are to begin to Apple.
Just throwing that out for whatever it's worth to you.
(One disclaimer: this is a rumor I heard, and thus comes from a much different source than previous info that I have recieved, some of which I've shared with you folks.)
Oh, and Apple's had a heck of a long time to plan for the move to the 970 and the 64bitness that comes with, so I would sure hope that they have done all they can to make it easy for themselves when they finally flip the 64bit switch in gcc (or whatever).
Where are all the lonely people that normally dig through Mac OS's to find interesting strings and such? If there is support in 10.2 already, has Apple really hidden any such references so well that people with time on their hands can't find them?
This is incorrect -- if the OS isn't "64-bit" then it cannot present any 64-bit APIs or even a 64-bit address space to any 64-bit applications. The OS support for 64-bitness must be in place before any 64-bit applications can be created for it. The cart before the horse, and all that.
For what its worth I think we'll see 10.3 (64-bit) and the 970 hardware appear at the same time. . .
</strong><hr></blockquote>
Ah, in that case I change my mind, and agree with your last comment above. I don't believe Apple will ship new PowerMacs unless they run 64 bit applications. I think Apple will make a big deal of 64 bits and the future of computing, and will want 64 bits working right from the start. They likely will have one or more 64 bit applications ready to go, just to make their point.
Thanks for the correction. I assumed too much from IBM's statement that the 970 would work with a 32 bit OS, with just a few minor changes.
IIRC the IBM PowerSeries chips were designed to run different OSes that IBM uses with the setting of one-bit software switches. I don't see why this can't be done with OS X, which is Unix Based. Therefore they could set a switch in the 970 to run 32-bit or 64-bit OS X if they wanted to. This would also give IBM the ability to run their OSes on the 970, which I am sure they are planning on.
<strong>IIRC the IBM PowerSeries chips were designed to run different OSes that IBM uses with the setting of one-bit software switches. I don't see why this can't be done with OS X, which is Unix Based. Therefore they could set a switch in the 970 to run 32-bit or 64-bit OS X if they wanted to. This would also give IBM the ability to run their OSes on the 970, which I am sure they are planning on.</strong><hr></blockquote>
Yes, this is correct -- but unless your OS is 64-bit aware you cannot run 64-bit applications in it. 64-bit applications require a 64-bit OS. If they ship a 10.2.x update with support for the 970 (which is easily doable) then only 32-bit applications will be able to run on it. When the OS is 64-bit it will look at the application, decide if it is a 32-bit or 64-bit app and setup the mode bit plus the address space accordingly. A 32-bit OS doesn't know how to do that... if it did it would be a 64-bit OS!
For what its worth, MacOS X 10.2 shipped last August and its not unreasonable for Apple to plan on a .1 increment roughly annually. The 970 hardware looks like it'll be ready in the Aug-Sept timeframe. This is an interesting co-incidence, is it not? Last time I checked August was a summer month so shipping in August would fullfill the "before fall" rumour. Applying Occam's Razor we should therefore conclude that the most likely course of events will be a preview at WWDC to allow developers to start preparing, followed by an August introduction.
Yes, understood about the 32/64-bit switch and I was real tempted to get the DP 1.42, has anyone seen speed comparison between the original DP 1 GHz (Jan 2002) and the 1.42?. I'd still consider if it has a 50% speed increase over the DP-gigger
Yes, understood about the 32/64-bit switch and I was real tempted to get the DP 1.42, has anyone seen speed comparison between the original DP 1 GHz (Jan 2002) and the 1.42?. I'd still consider if it has a 50% speed increase over the DP-gigger
Does it really matter? If you need a machine now, buy it now. You really aren't going to notice a +/- 10% speed difference anyhow unless you have hard real-time constraints (in which case you probably don't want a Mac or a Windows box).
Does it really matter? If you need a machine now, buy it now. You really aren't going to notice a +/- 10% speed difference anyhow unless you have hard real-time constraints (in which case you probably don't want a Mac or a Windows box).
Nope it doesn't really matter, its just my mental aberration (justification) to not buy a new Mac until I can gain a 50% increase in speed over my old one. My DP gigger suits me fine for my level of work (in most cases except limited use of 3-d rendering in POV-ray and 700 MB tiff file work in AI/PS)
Nope it doesn't really matter, its just my mental aberration (justification) to not buy a new Mac until I can gain a 50% increase in speed over my old one. My DP gigger suits me fine for my level of work (in most cases except limited use of 3-d rendering in POV-ray and 700 MB tiff file work in AI/PS)
If those exeptions are the only where your current 1gigger isn't fast enough for you, I'd say wait for the good stuff, the 1,42Ghz DP one is still not going to be fast enough for you, since 40-50% is pretty unnoticeable on that kind of work IMO. Unless you have no problems throwing lots of cash on a small step-up in speed, I don't think there's a very good point in upgrading to the current top-end.
Yes, this is correct -- but unless your OS is 64-bit aware you cannot run 64-bit applications in it. 64-bit applications require a 64-bit OS.
I'd like to contest your conclusion that a 64-bit app would need a 64-bit OS.
As we've been told the 970 will be able to run 32-bit software without skipping a beat. This implies that we have both 64-bit and 32-bit instruction sets.
Therefore, the operating system API can be 32-bit even on a 64-bit processor. If Apple decides to make OS X 970 aware that does not imply that awareness will force all 64-bit (or 32-bit) instructions to be ignored (it's not an either/or proposition).
What we could have is a 64-bit enabled processor with 32-bit APIs running applications that can be either 32-bit or 64-bit, BUT have both type call on the 32-bit APIs (existing OS X 10.2.x).
When the OS is ready for full 64-bit operation those applications that are taking advantage of the 64-bit instruction set could be re-linked with 64-bit API libraries (OS X 10.3).
The 32-bit applications would have to call the 32-bit APIs for the rest of their product lives anyway, so a 32-bit API must be in place even if the operating system is pure 64-bit code.
Those that would like to adopt the 64-bit instruction set early would have to chose between developing their own 64-bit to 32-bit API translation layer or pick up the Apple supplied version if the OS is not fully 64-bit at the time of the 970 release.
I would be very surprised if Apple does not already have an API strategy already in place.
None the less as long as the 970 has the 64-bitness bit enabled by the OS developers could use 64-bit instructions to develop custom code (alas rendering, audio processing, etc.) that runs on a 32-bit OS.
Please let me know where I may have made a mistake in my reasoning.
I'd like to contest your conclusion that a 64-bit app would need a 64-bit OS.
As we've been told the 970 will be able to run 32-bit software without skipping a beat. This implies that we have both 64-bit and 32-bit instruction sets.
Therefore, the operating system API can be 32-bit even on a 64-bit processor. If Apple decides to make OS X 970 aware that does not imply that awareness will force all 64-bit (or 32-bit) instructions to be ignored (it's not an either/or proposition).
What we could have is a 64-bit enabled processor with 32-bit APIs running applications that can be either 32-bit or 64-bit, BUT have both type call on the 32-bit APIs (existing OS X 10.2.x).
When the OS is ready for full 64-bit operation those applications that are taking advantage of the 64-bit instruction set could be re-linked with 64-bit API libraries (OS X 10.3).
The 32-bit applications would have to call the 32-bit APIs for the rest of their product lives anyway, so a 32-bit API must be in place even if the operating system is pure 64-bit code.
Those that would like to adopt the 64-bit instruction set early would have to chose between developing their own 64-bit to 32-bit API translation layer or pick up the Apple supplied version if the OS is not fully 64-bit at the time of the 970 release.
I would be very surprised if Apple does not already have an API strategy already in place.
None the less as long as the 970 has the 64-bitness bit enabled by the OS developers could use 64-bit instructions to develop custom code (alas rendering, audio processing, etc.) that runs on a 32-bit OS.
Please let me know where I may have made a mistake in my reasoning.
The main point of using 64 bits is the ability to address memory using 64 bit pointers. If your OS (which does all the memory management) only uses 32 bits for memory addresses there is no way it can understand the pointers a 64 bit programme will want to pass to it.
The ability to use the 64 bit instructions (basically 64 bit integers) is not, of itself, enough, and indeed, very few programmes would benefit from using them.
Therefore, the operating system API can be 32-bit even on a 64-bit processor. If Apple decides to make OS X 970 aware that does not imply that awareness will force all 64-bit (or 32-bit) instructions to be ignored (it's not an either/or proposition).
Yes, but, you are linking against OS provided libraries. When you call malloc, you will get a 32-bit pointer back. As mentioned, the OS in doing memory management, CPU management, etc. is still treating everything in 32-bit (in your scenario).
Though I would suspect 10.3 will be our 64-bit OS, from the sounds of IBM's quote, there could be a 10.2.x that was only for the 970 and supported 64-bit.
OK, I understand all the objections. This to me simply means that a 970 Mac MUST have a 64-bit OS on practical grounds.
This conclusion leads us to the questions:
Will Apple have the OS X 10.3 (64-bit) code ready at the introduction of the 970 based Mac?
Will the 970 be available in short order from IBM?
The problem is that BOTH the OS and CPU must be ready for the new Mac to appear (not to mention a number things people are asking for on these forums: FW2, USB2, Serial ATA, etc. plus things we don't even think of).
We can only hope that Apple has everything properly lined up for July. I sure hope so, because I will not be able to take any more insta-pundit comments such as the latest Dvorak column.
OK, I understand all the objections. This to me simply means that a 970 Mac MUST have a 64-bit OS on practical grounds.
This conclusion leads us to the questions:
Will Apple have the OS X 10.3 (64-bit) code ready at the introduction of the 970 based Mac?
Will the 970 be available in short order from IBM?
The problem is that BOTH the OS and CPU must be ready for the new Mac to appear (not to mention a number things people are asking for on these forums: FW2, USB2, Serial ATA, etc. plus things we don't even think of).
We can only hope that Apple has everything properly lined up for July. I sure hope so, because I will not be able to take any more insta-pundit comments such as the latest Dvorak column.
And what about PCI-Express or other fast bus to replace the present PCI/AGP busses? I had heard that Apple was interested in the fiber channel some time back, but not much recently. PCI Express is probably not to be had in PCs until Q1 or Q2 next year unless Dell is pulling a fast one. Did you know that the AGP 8X spec allows for *two* AGP 8X slots??? That would allow some interesting possibilities (even if you only install AGP 4X cards in the slot).
Comments
<strong>
Jobs apparently loved the eMate.
I would not put it past him to revive something very much like it.</strong><hr></blockquote>
Your first line I assumed was sarcasm since Steve quickly axed the Newton (and the eMate which ran the same OS) but you second line sounded serious. I find it hard to believe he would revive something like the eMate. Was I just missing a little more sarcasm?
Nope. Amorph is right-- Steve did like the eMate, which is why his rolling Newton, Inc. back into Apple and then killing the entire platform wholesale was a bit of a surprise. And to think Newton, Inc. was about to post it's first profitable quarter ever...
<strong>Who wouldn't buy a new PowerMac with an IBM 970 running OS 10.2? It will still run 64 bit applications, even if the OS is 32 bits, and we know that OS X will soon be upgraded to 64 bits too.</strong><hr></blockquote>
This is incorrect -- if the OS isn't "64-bit" then it cannot present any 64-bit APIs or even a 64-bit address space to any 64-bit applications. The OS support for 64-bitness must be in place before any 64-bit applications can be created for it. The cart before the horse, and all that.
For what its worth I think we'll see 10.3 (64-bit) and the 970 hardware appear at the same time. There might be a few limitations on what 64-bit apps can have access to, but they'll deliver. Apple has actually had a pretty good software release track record since 10.1 arrived... at least compared to their history (and most of the rest of the industry).
<strong>Your theory hinges on the difficulty of moving OSX to 64bit.</strong><hr></blockquote>
Actually, I'd originally stated that I know many people have said that bringing OS X to 64 bits won't be that hard.
But even if all you had to do was change five #define lines and recompile, you can bet that there'd still be at least three months of internal QA work, and then a beta test phase, to make sure those little changes didn't result in a subtle bug or two or cause some instability.
Plus, as a marketing and as a philosophical milestone, going to 64-bit support is enough of a big step, regardless of the technical difficulty (or lack thereof) of doing so, that I think the change would warrant a major version change -- to 10.3. This would then tie 64-bitness together with who knows what other features are planned for 10.3.
64-bitness might be easy, but better Windows file sharing, fixing Finder bugs (still a few of those in Jaguar!), or the tele-olfactory version of iChat might be running behind schedule. That would mean no 64-bit support until the whole enchilada is ready to go.
This is why I postulate the idea of a 10.2.x release that does nothing more than tweak the OS to recognize the 970 and makes sure it runs in its 32-bit compatibility mode.
But I'll put it here anyway:
Just heard a rumor that 10.3 isn't scheduled until fall. This does make some sense with 10.2 coming out last fall, but it doesn't seem to jive with when 970 shipments are to begin to Apple.
Just throwing that out for whatever it's worth to you.
(One disclaimer: this is a rumor I heard, and thus comes from a much different source than previous info that I have recieved, some of which I've shared with you folks.)
Oh, and Apple's had a heck of a long time to plan for the move to the 970 and the 64bitness that comes with, so I would sure hope that they have done all they can to make it easy for themselves when they finally flip the 64bit switch in gcc (or whatever).
Where are all the lonely people that normally dig through Mac OS's to find interesting strings and such? If there is support in 10.2 already, has Apple really hidden any such references so well that people with time on their hands can't find them?
Or has no one looked? I doubt that.
----------------------------------
970 bound in Nashville
<strong>
This is incorrect -- if the OS isn't "64-bit" then it cannot present any 64-bit APIs or even a 64-bit address space to any 64-bit applications. The OS support for 64-bitness must be in place before any 64-bit applications can be created for it. The cart before the horse, and all that.
For what its worth I think we'll see 10.3 (64-bit) and the 970 hardware appear at the same time. . .
</strong><hr></blockquote>
Ah, in that case I change my mind, and agree with your last comment above. I don't believe Apple will ship new PowerMacs unless they run 64 bit applications. I think Apple will make a big deal of 64 bits and the future of computing, and will want 64 bits working right from the start. They likely will have one or more 64 bit applications ready to go, just to make their point.
Thanks for the correction. I assumed too much from IBM's statement that the 970 would work with a 32 bit OS, with just a few minor changes.
<strong>IIRC the IBM PowerSeries chips were designed to run different OSes that IBM uses with the setting of one-bit software switches. I don't see why this can't be done with OS X, which is Unix Based. Therefore they could set a switch in the 970 to run 32-bit or 64-bit OS X if they wanted to. This would also give IBM the ability to run their OSes on the 970, which I am sure they are planning on.</strong><hr></blockquote>
Yes, this is correct -- but unless your OS is 64-bit aware you cannot run 64-bit applications in it. 64-bit applications require a 64-bit OS. If they ship a 10.2.x update with support for the 970 (which is easily doable) then only 32-bit applications will be able to run on it. When the OS is 64-bit it will look at the application, decide if it is a 32-bit or 64-bit app and setup the mode bit plus the address space accordingly. A 32-bit OS doesn't know how to do that... if it did it would be a 64-bit OS!
For what its worth, MacOS X 10.2 shipped last August and its not unreasonable for Apple to plan on a .1 increment roughly annually. The 970 hardware looks like it'll be ready in the Aug-Sept timeframe. This is an interesting co-incidence, is it not? Last time I checked August was a summer month so shipping in August would fullfill the "before fall" rumour. Applying Occam's Razor we should therefore conclude that the most likely course of events will be a preview at WWDC to allow developers to start preparing, followed by an August introduction.
Originally posted by Bigc
Yes, understood about the 32/64-bit switch and I was real tempted to get the DP 1.42, has anyone seen speed comparison between the original DP 1 GHz (Jan 2002) and the 1.42?. I'd still consider if it has a 50% speed increase over the DP-gigger
Does it really matter? If you need a machine now, buy it now. You really aren't going to notice a +/- 10% speed difference anyhow unless you have hard real-time constraints (in which case you probably don't want a Mac or a Windows box).
Originally posted by Programmer
Does it really matter? If you need a machine now, buy it now. You really aren't going to notice a +/- 10% speed difference anyhow unless you have hard real-time constraints (in which case you probably don't want a Mac or a Windows box).
Nope it doesn't really matter, its just my mental aberration (justification) to not buy a new Mac until I can gain a 50% increase in speed over my old one. My DP gigger suits me fine for my level of work (in most cases except limited use of 3-d rendering in POV-ray and 700 MB tiff file work in AI/PS)
Originally posted by Bigc
Nope it doesn't really matter, its just my mental aberration (justification) to not buy a new Mac until I can gain a 50% increase in speed over my old one. My DP gigger suits me fine for my level of work (in most cases except limited use of 3-d rendering in POV-ray and 700 MB tiff file work in AI/PS)
If those exeptions are the only where your current 1gigger isn't fast enough for you, I'd say wait for the good stuff, the 1,42Ghz DP one is still not going to be fast enough for you, since 40-50% is pretty unnoticeable on that kind of work IMO. Unless you have no problems throwing lots of cash on a small step-up in speed, I don't think there's a very good point in upgrading to the current top-end.
Originally posted by Programmer
Yes, this is correct -- but unless your OS is 64-bit aware you cannot run 64-bit applications in it. 64-bit applications require a 64-bit OS.
I'd like to contest your conclusion that a 64-bit app would need a 64-bit OS.
As we've been told the 970 will be able to run 32-bit software without skipping a beat. This implies that we have both 64-bit and 32-bit instruction sets.
Therefore, the operating system API can be 32-bit even on a 64-bit processor. If Apple decides to make OS X 970 aware that does not imply that awareness will force all 64-bit (or 32-bit) instructions to be ignored (it's not an either/or proposition).
What we could have is a 64-bit enabled processor with 32-bit APIs running applications that can be either 32-bit or 64-bit, BUT have both type call on the 32-bit APIs (existing OS X 10.2.x).
When the OS is ready for full 64-bit operation those applications that are taking advantage of the 64-bit instruction set could be re-linked with 64-bit API libraries (OS X 10.3).
The 32-bit applications would have to call the 32-bit APIs for the rest of their product lives anyway, so a 32-bit API must be in place even if the operating system is pure 64-bit code.
Those that would like to adopt the 64-bit instruction set early would have to chose between developing their own 64-bit to 32-bit API translation layer or pick up the Apple supplied version if the OS is not fully 64-bit at the time of the 970 release.
I would be very surprised if Apple does not already have an API strategy already in place.
None the less as long as the 970 has the 64-bitness bit enabled by the OS developers could use 64-bit instructions to develop custom code (alas rendering, audio processing, etc.) that runs on a 32-bit OS.
Please let me know where I may have made a mistake in my reasoning.
Originally posted by iCode
I'd like to contest your conclusion that a 64-bit app would need a 64-bit OS.
As we've been told the 970 will be able to run 32-bit software without skipping a beat. This implies that we have both 64-bit and 32-bit instruction sets.
Therefore, the operating system API can be 32-bit even on a 64-bit processor. If Apple decides to make OS X 970 aware that does not imply that awareness will force all 64-bit (or 32-bit) instructions to be ignored (it's not an either/or proposition).
What we could have is a 64-bit enabled processor with 32-bit APIs running applications that can be either 32-bit or 64-bit, BUT have both type call on the 32-bit APIs (existing OS X 10.2.x).
When the OS is ready for full 64-bit operation those applications that are taking advantage of the 64-bit instruction set could be re-linked with 64-bit API libraries (OS X 10.3).
The 32-bit applications would have to call the 32-bit APIs for the rest of their product lives anyway, so a 32-bit API must be in place even if the operating system is pure 64-bit code.
Those that would like to adopt the 64-bit instruction set early would have to chose between developing their own 64-bit to 32-bit API translation layer or pick up the Apple supplied version if the OS is not fully 64-bit at the time of the 970 release.
I would be very surprised if Apple does not already have an API strategy already in place.
None the less as long as the 970 has the 64-bitness bit enabled by the OS developers could use 64-bit instructions to develop custom code (alas rendering, audio processing, etc.) that runs on a 32-bit OS.
Please let me know where I may have made a mistake in my reasoning.
The main point of using 64 bits is the ability to address memory using 64 bit pointers. If your OS (which does all the memory management) only uses 32 bits for memory addresses there is no way it can understand the pointers a 64 bit programme will want to pass to it.
The ability to use the 64 bit instructions (basically 64 bit integers) is not, of itself, enough, and indeed, very few programmes would benefit from using them.
michael
Originally posted by iCode
Therefore, the operating system API can be 32-bit even on a 64-bit processor. If Apple decides to make OS X 970 aware that does not imply that awareness will force all 64-bit (or 32-bit) instructions to be ignored (it's not an either/or proposition).
Yes, but, you are linking against OS provided libraries. When you call malloc, you will get a 32-bit pointer back. As mentioned, the OS in doing memory management, CPU management, etc. is still treating everything in 32-bit (in your scenario).
Though I would suspect 10.3 will be our 64-bit OS, from the sounds of IBM's quote, there could be a 10.2.x that was only for the 970 and supported 64-bit.
This conclusion leads us to the questions:
Will Apple have the OS X 10.3 (64-bit) code ready at the introduction of the 970 based Mac?
Will the 970 be available in short order from IBM?
The problem is that BOTH the OS and CPU must be ready for the new Mac to appear (not to mention a number things people are asking for on these forums: FW2, USB2, Serial ATA, etc. plus things we don't even think of).
We can only hope that Apple has everything properly lined up for July. I sure hope so, because I will not be able to take any more insta-pundit comments such as the latest Dvorak column.
Originally posted by iCode
OK, I understand all the objections. This to me simply means that a 970 Mac MUST have a 64-bit OS on practical grounds.
This conclusion leads us to the questions:
Will Apple have the OS X 10.3 (64-bit) code ready at the introduction of the 970 based Mac?
Will the 970 be available in short order from IBM?
The problem is that BOTH the OS and CPU must be ready for the new Mac to appear (not to mention a number things people are asking for on these forums: FW2, USB2, Serial ATA, etc. plus things we don't even think of).
We can only hope that Apple has everything properly lined up for July. I sure hope so, because I will not be able to take any more insta-pundit comments such as the latest Dvorak column.
And what about PCI-Express or other fast bus to replace the present PCI/AGP busses? I had heard that Apple was interested in the fiber channel some time back, but not much recently. PCI Express is probably not to be had in PCs until Q1 or Q2 next year unless Dell is pulling a fast one. Did you know that the AGP 8X spec allows for *two* AGP 8X slots??? That would allow some interesting possibilities (even if you only install AGP 4X cards in the slot).