I think Apple is realizing that selling a complete closed system ala IBM, SUN, SGI to the general public is not going to take them to the next level in the consumer market.
But unless Steve drops a REALLY BIG "and one more thing" in WWDC06 or MWSF07, Apple's still going to be selling a complete closed system ala IBM, SUN, or SGI. It's certainly possible he'll do this... I never believed Steve would give up on his "No Ugly Monitors on Nice Macs" position (let alone dumping the PPC right when the PPC was about to take a BIG performance jump across the board), but right now Apple is saying that you'll only be able to run OS X on Apple's own hardware.
So? The Pentium Mobile isn't a P4, it's a PIII with a faster bus, and that doesn't seem to be hurting it any. Intel talks about it having "P4 technology", but there really isn't much of the P4 in there other than the 533 Mhz FSB.
Apple has played the Freescale/Motorola game before, haven't they?
They still are. And the only problem with the Freescale/Motorola game is they were slow coming to the table with a new processor using the G4 core. And... I wouldn't bet on Apple not being part of the problem there: Apple was real reluctant to make any big changes in their Powermacs... they basically had two (maybe 3, depending on how you count the "Sawtooth" G4) families of motherboards with minor tweaks from the Beige G3 up to the G5: the "G3 ZIF" models (Beige G3, B&W G3, and "Yikes" G4) and the "G4 ZIF" models ("AGP" G4 through to the end of the line).
If Motorola/Freescale had come out with something like the e600 package a few years back, would Apple have been willing to dump the "ZIF" concept and redesign the whole motherboard around it?
The e700? It's not anything Apple needs right now. The 970 is still perfectly fine for the high end desktop, and the e600 for the laptop and low-end desktop. There's no reason they need to just "play the IBM game" or "play the Freescale game".
The e600 is not scaling in processor speed like they'd promised
Nobody is scaling up in processor speed like they promised. Not even Intel.
But here's the kicker : Apple wants to be selling them now
Apple can't sell Pentium M Powerbooks now, either. If they can sell Pentium M Powerbooks sooner than they could sell e600s I'd be stunned. And if a 1.5 GHz e600 didn't smoke a 2 GHz Pentium M for real world use, I'd be surprised.
A dual-core 1.5 Ghz e600?
I'd take that over a 3 GHz G5, simply because I live in Houston and I have to pay 3 times for every watt of electricity that goes into my computer: once for the computer, and then twice as much again for the air-conditioning it takes to keep my office livable. When I turned off my 1.7 GHz P4 running free UNIX because my G4/466 running OSX was good enough I took a performance hit, but so did my A/C bill.
they're not looking to be much faster in clock speed than what's in a PowerBook right now.
But they'll be MUCH faster in actual performance. The G4 is a very good core, and if it wasn't for the slow bus holding it back I wouldn't have bet on the original 1.8 GHz G5s actually being faster than their contemporary G4 models. With two cores, one that can be spun up on demand?
Meanwhile, what's a common Intel laptop? Pentium M laptops at 2GHz and up are all over the place.
None of them are running OS X, and none of them will be running OS X until the second half of 2006... or maybe even the first quarter of 2007: Apple's not promising they'll have everything Intelified until 4Q07.
it is all about being able to deliver on promises, and in quantity
Intel's had to renege on a few big ones lately, too, which is why they're going back to the PIII core and pumping it up with a faster bus (like Freescale is doing for the G4) in the first place.
And Apple may have overstated the Megahertz Myth, but not by much. They were basically correct, and I wish they'd remember that. Maybe they will, once they've finally killed off "Classic" Mac OS and completed the ten-year transition to NeXTstep.
Maybe the potential for people pirating OS X and running it on Thinkpads will convince Apple to come out with a Powerbook that's got a better keyboard and pointing device.
That's what's kept me from buying a 'book... I can't stand typing on that weird flat keyboard after using a good laptop.
What's the difference between installing OS X under a hypervisor and installing it on the raw silicon and iron? If you can get it to boot on drivers for the hypervisor, you can get it to boot on drivers for any other hardware. If the Hypervisor can spoof the copy protection, you can patch OS X to spoof the copy protection the same way.
but switching architectures gives Apple the opportunity to grow their market share through piracy
I've thought about this too. It would be really really hard to make OS X non-crackable without destroying a huge part of the value of the OS, and without pretty much abandoning their commitment to open source. And, of course, Apple would benefit from the resulting stealth market share over the long term... so long as it didn't cut too heavily into their sales.
I've also worried about this. Because if Apple really did apply some kind of strong rights management to OS X half the things that I find valuable in the system, including its comparative reliability and robustness (compared to Windows at least) and the ease of digging in and fixing things when it does stumble, would vanish.
But Steve Jobs has indicated that he does understand why DRM is at best a stopgap, so you could be right. I hope you are.
Remember Apple announcing that they were going to drag ISVs kicking and screaming into a new API, called "Yellow Box", and existing Mac apps were going to run in the "Blue Box" emulation environment?
Remember Adobe and the rest digging in their heels?
Remember Apple coming up with a stopgap called "Carbon"?
Remember Yellow Box becoming "Cocoa" and Blue Box becoming "Classic"?
Remember Apple cancelling the last "OS-9 bootable Mac". And backing down? I think that happened a couple of times, actually.
Well, that happened again, last year, and it's stayed gone. Classic Mac OS is no longer supported on any shipping Macs. Carbon apps are going to have to go through a much larger upgrade process than Cocoa to run on OSX on Intel.
I think that if the ISVs hadn't screamed, Macs would have gone to Intel some time between 2000 or 2002. "Marklar" wasn't "Just in case", it was "When we can get away with it".
The 3 GHz G5 and G5 on Powerbook complaints? They're just excuses. IBM and Freescale haven't actually dropped the ball nearly as badly as people claim... and IBM's actually done better than Intel in clock speed improvements. No, this isn't because IBM did anything wrong. It's not a new transition. It's the last step of the transition that started when Steve came back to Apple and brought NeXTstep and its CPU-agnostic architecture with him.
Making Spotlight use its own DB is fine - if all you care about is using Spotlight. But if the DB were built into the filesystem (or at least made an OS-level component) then other applications could make use of the search capabilities as well.
What's an "OS-level component"? Spotlight's got a framework that any application can call: that's how Spotlight works in Finder and Mail.
If some 'big company' then tries to patent it, there's obvious prior art.
That sounds good but in practice it doesn't work that way. For example:
Palm is no longer using Graffiti (Graffiti 2 is a variant of a previously competing input method called Jot) because of a Xerox patent on single-stroke character input. There's an open-source character recogniser that's closer to Graffiti than Xerox' Unistroke system, that predated either. It wasn't considered prior art.
there are no entry barriers that would slow a competitor down from reaping the rewards from your software-implemented invention.
You're still thinking "patents exist to reward inventors" rather than "patents exist to promote invention". Patents exist because it takes time and money to build a physical device, production line, or other hardware to turn a patent into a product, and by granting a temporary monopoly on the patented concept you give the inventor time to profit from that investment. For software, it takes less effort to prepare a concept for production than it does to prepare it for filing a patent. So what is the point of granting it a patent, if just preparing it (let alone filing it) has already cost more than the actions it's intended to promote.
people are buying more laptops than desktops and IBM is not making powerful laptop PowerPC-based chips.
IBM isn't, but IBM isn't the one who makes low power PPCs. I can't conceive of why anyone would even be talking about G5s in laptops... that's not IBM's job. That's Freescale's. And Freescale have some really tasty new chips coming down the pipe.
And how is that easier than a DB-oriented filesystem where any set of search keywords can be attached to a file?
Well, it already exists, it doesn't break any current software, it's got a 35 year track record of actually working, and it doesn't force you to redesign everything when you have to deal with multiple different operating systems. Oh, yeh, these are just practical considerations... but they're kind of important.
And on the other hand, you don't actually need a DB-oriented file system to attach search keywords to a file. You just need a user interface that lets you declare associations between keywords and files. Whether those associations are explicitly in the filesystem, as in BeFS, or whether they can be added in a separate database, as in Spotlight, is an implementation detail.
I see no reason why those involving physics or mechanical engineering are more deserving than those involving information processing.
Because the word "deserving" in that sentence is the result of a misunderstanding. The purpose of patents is not to reward inventors, it is to encourage the development of the arts and sciences by rewarding inventors. Software and other "pure process" patents are not necessary because there other mechanisms for achieving that goal, and because on balance software patents do more to discourage than promote invention.
What's broken about it is that a single hierarchical classification scheme may not always be appropriate for a given body of data.
The current file system isn't just a hierarchy. It provides UNIX hard links, so there is no single location a file has to be in. If I create a hard link to a file it looks like two files, but it's really one.
This is not the same as soft links or aliases, both names for the file are equally valid... you can delete either and the other is still there.
The only problem is that Finder doesn't let you do this. You can create an alias, but it's more combersome than it should be and it's not a hard link. You should be able to drag a file into another folder with some command key pressed and it would create a hard link for you.
if each file were in a database, with search keywords, I could find anything I wanted just as easily as anything else - there's no predetermined hierarchy that makes it easier to find some things than others.
But once you found it, you can show it in the finder and there's a folder right there with the rest of the information associated with the project.
It's not "broken", and Spotlight would work better as an augmentation tool for Finder, not a replacement.
Hands up everyone with a 64-bit app on Tiger. Not just "I built $thing for 64-bit OS X", but "This app requires 64-bit OS X to work" or "This app's a dog if it's not compiled 64-bit". If there's more than a couple in the world, I'll be very much surprised. If there's even one that won't compile right up on Linux-64 I'll be amazed.
OK, now hands up everyone with a 64-bit app on Windows, same caveat. There may be a few more, because Win32 has more of an address space bottleneck than other 32-bit environments.
And finally... 64-bit Linux, with the same restriction. This is where I expect to see the most hands.
Because, you know, we've been using Alpha for 10 years now, and there's very little code that really needs or even benefits from it. If you're not sure, without trying it, that your code's going to crater going "back" to 32-bit then your code was almost certainly not really using 64-bit mode.
This would make it physically impossible for non-apple intel processors to boot OS X.
How will that stop someone from loading the install CD onto disk, replacing the boot with the one from Darwin x86 (which they already have), and burning a new install CD?
The video card is still the most likely place I can see a lock being hidden.
When the terms of the settlement were announced several people brought up the way they could be used against open-source software and Samba was frequently mentioned.
That's one reason I registered my own domain so early on.
Then some dope at work got their laptop infected, and all of a sudden I was getting spam (my address was in their address book).
Oh yeh, that's a REALLY effective technique... for the bad guys. The spammers don't even need to cooperate with the virus authors (though I don't doubt the stories that they do), they just need to seed a dropbox address in a spam somewhere and wait for viruses to hit it.
There's really no good protection for that other than things like a dated address... mail to anything@slashdot.2004.taronga.com, for example, now routes to 127.0.0.1:)
I get a similar amount of spam delivery attempts. What realy impresses me are the Nigerians. I get 419ers in clumps, and many of them have different typos... which implies they're typing these things in (presumably in some internet cafe in Lagos) over and over and over again...
Slashdot Mirror
I think Apple is realizing that selling a complete closed system ala IBM, SUN, SGI to the general public is not going to take them to the next level in the consumer market.
But unless Steve drops a REALLY BIG "and one more thing" in WWDC06 or MWSF07, Apple's still going to be selling a complete closed system ala IBM, SUN, or SGI. It's certainly possible he'll do this... I never believed Steve would give up on his "No Ugly Monitors on Nice Macs" position (let alone dumping the PPC right when the PPC was about to take a BIG performance jump across the board), but right now Apple is saying that you'll only be able to run OS X on Apple's own hardware.
Not only are these new Freescale chips not G5s...
So? The Pentium Mobile isn't a P4, it's a PIII with a faster bus, and that doesn't seem to be hurting it any. Intel talks about it having "P4 technology", but there really isn't much of the P4 in there other than the 533 Mhz FSB.
Apple has played the Freescale/Motorola game before, haven't they?
They still are. And the only problem with the Freescale/Motorola game is they were slow coming to the table with a new processor using the G4 core. And... I wouldn't bet on Apple not being part of the problem there: Apple was real reluctant to make any big changes in their Powermacs... they basically had two (maybe 3, depending on how you count the "Sawtooth" G4) families of motherboards with minor tweaks from the Beige G3 up to the G5: the "G3 ZIF" models (Beige G3, B&W G3, and "Yikes" G4) and the "G4 ZIF" models ("AGP" G4 through to the end of the line).
If Motorola/Freescale had come out with something like the e600 package a few years back, would Apple have been willing to dump the "ZIF" concept and redesign the whole motherboard around it?
The e700? It's not anything Apple needs right now. The 970 is still perfectly fine for the high end desktop, and the e600 for the laptop and low-end desktop. There's no reason they need to just "play the IBM game" or "play the Freescale game".
The e600 is not scaling in processor speed like they'd promised
Nobody is scaling up in processor speed like they promised. Not even Intel.
But here's the kicker : Apple wants to be selling them now
Apple can't sell Pentium M Powerbooks now, either. If they can sell Pentium M Powerbooks sooner than they could sell e600s I'd be stunned. And if a 1.5 GHz e600 didn't smoke a 2 GHz Pentium M for real world use, I'd be surprised.
A dual-core 1.5 Ghz e600?
I'd take that over a 3 GHz G5, simply because I live in Houston and I have to pay 3 times for every watt of electricity that goes into my computer: once for the computer, and then twice as much again for the air-conditioning it takes to keep my office livable. When I turned off my 1.7 GHz P4 running free UNIX because my G4/466 running OSX was good enough I took a performance hit, but so did my A/C bill.
they're not looking to be much faster in clock speed than what's in a PowerBook right now.
But they'll be MUCH faster in actual performance. The G4 is a very good core, and if it wasn't for the slow bus holding it back I wouldn't have bet on the original 1.8 GHz G5s actually being faster than their contemporary G4 models. With two cores, one that can be spun up on demand?
Meanwhile, what's a common Intel laptop? Pentium M laptops at 2GHz and up are all over the place.
None of them are running OS X, and none of them will be running OS X until the second half of 2006... or maybe even the first quarter of 2007: Apple's not promising they'll have everything Intelified until 4Q07.
it is all about being able to deliver on promises, and in quantity
Intel's had to renege on a few big ones lately, too, which is why they're going back to the PIII core and pumping it up with a faster bus (like Freescale is doing for the G4) in the first place.
And Apple may have overstated the Megahertz Myth, but not by much. They were basically correct, and I wish they'd remember that. Maybe they will, once they've finally killed off "Classic" Mac OS and completed the ten-year transition to NeXTstep.
Maybe the potential for people pirating OS X and running it on Thinkpads will convince Apple to come out with a Powerbook that's got a better keyboard and pointing device.
That's what's kept me from buying a 'book... I can't stand typing on that weird flat keyboard after using a good laptop.
What's the difference between installing OS X under a hypervisor and installing it on the raw silicon and iron? If you can get it to boot on drivers for the hypervisor, you can get it to boot on drivers for any other hardware. If the Hypervisor can spoof the copy protection, you can patch OS X to spoof the copy protection the same way.
but switching architectures gives Apple the opportunity to grow their market share through piracy
I've thought about this too. It would be really really hard to make OS X non-crackable without destroying a huge part of the value of the OS, and without pretty much abandoning their commitment to open source. And, of course, Apple would benefit from the resulting stealth market share over the long term... so long as it didn't cut too heavily into their sales.
I've also worried about this. Because if Apple really did apply some kind of strong rights management to OS X half the things that I find valuable in the system, including its comparative reliability and robustness (compared to Windows at least) and the ease of digging in and fixing things when it does stumble, would vanish.
But Steve Jobs has indicated that he does understand why DRM is at best a stopgap, so you could be right. I hope you are.
Most people don't care what's inside the watch and most don't care what's inside the computer.
So why don't most people buy cheap $10.00 watches at the convenience store?
Remember Apple announcing that they were going to drag ISVs kicking and screaming into a new API, called "Yellow Box", and existing Mac apps were going to run in the "Blue Box" emulation environment?
Remember Adobe and the rest digging in their heels?
Remember Apple coming up with a stopgap called "Carbon"?
Remember Yellow Box becoming "Cocoa" and Blue Box becoming "Classic"?
Remember Apple cancelling the last "OS-9 bootable Mac". And backing down? I think that happened a couple of times, actually.
Well, that happened again, last year, and it's stayed gone. Classic Mac OS is no longer supported on any shipping Macs. Carbon apps are going to have to go through a much larger upgrade process than Cocoa to run on OSX on Intel.
I think that if the ISVs hadn't screamed, Macs would have gone to Intel some time between 2000 or 2002. "Marklar" wasn't "Just in case", it was "When we can get away with it".
The 3 GHz G5 and G5 on Powerbook complaints? They're just excuses. IBM and Freescale haven't actually dropped the ball nearly as badly as people claim... and IBM's actually done better than Intel in clock speed improvements. No, this isn't because IBM did anything wrong. It's not a new transition. It's the last step of the transition that started when Steve came back to Apple and brought NeXTstep and its CPU-agnostic architecture with him.
Making Spotlight use its own DB is fine - if all you care about is using Spotlight. But if the DB were built into the filesystem (or at least made an OS-level component) then other applications could make use of the search capabilities as well.
What's an "OS-level component"? Spotlight's got a framework that any application can call: that's how Spotlight works in Finder and Mail.
That's about as "OS-level" as things get in OS X.
If some 'big company' then tries to patent it, there's obvious prior art.
That sounds good but in practice it doesn't work that way. For example:
Palm is no longer using Graffiti (Graffiti 2 is a variant of a previously competing input method called Jot) because of a Xerox patent on single-stroke character input. There's an open-source character recogniser that's closer to Graffiti than Xerox' Unistroke system, that predated either. It wasn't considered prior art.
there are no entry barriers that would slow a competitor down from reaping the rewards from your software-implemented invention.
You're still thinking "patents exist to reward inventors" rather than "patents exist to promote invention". Patents exist because it takes time and money to build a physical device, production line, or other hardware to turn a patent into a product, and by granting a temporary monopoly on the patented concept you give the inventor time to profit from that investment. For software, it takes less effort to prepare a concept for production than it does to prepare it for filing a patent. So what is the point of granting it a patent, if just preparing it (let alone filing it) has already cost more than the actions it's intended to promote.
people are buying more laptops than desktops and IBM is not making powerful laptop PowerPC-based chips.
IBM isn't, but IBM isn't the one who makes low power PPCs. I can't conceive of why anyone would even be talking about G5s in laptops... that's not IBM's job. That's Freescale's. And Freescale have some really tasty new chips coming down the pipe.
Too bad the website was designed by Catbert.
Did I miss something or is here no "next" link on any of the gallery pages?
And how is that easier than a DB-oriented filesystem where any set of search keywords can be attached to a file?
Well, it already exists, it doesn't break any current software, it's got a 35 year track record of actually working, and it doesn't force you to redesign everything when you have to deal with multiple different operating systems. Oh, yeh, these are just practical considerations... but they're kind of important.
And on the other hand, you don't actually need a DB-oriented file system to attach search keywords to a file. You just need a user interface that lets you declare associations between keywords and files. Whether those associations are explicitly in the filesystem, as in BeFS, or whether they can be added in a separate database, as in Spotlight, is an implementation detail.
I see no reason why those involving physics or mechanical engineering are more deserving than those involving information processing.
Because the word "deserving" in that sentence is the result of a misunderstanding. The purpose of patents is not to reward inventors, it is to encourage the development of the arts and sciences by rewarding inventors. Software and other "pure process" patents are not necessary because there other mechanisms for achieving that goal, and because on balance software patents do more to discourage than promote invention.
Then said "young inventor" needs to keep his/her invention (and other supporting materials) away from prying eyes.
This breaks open source and open systems.
What's broken about it is that a single hierarchical classification scheme may not always be appropriate for a given body of data.
The current file system isn't just a hierarchy. It provides UNIX hard links, so there is no single location a file has to be in. If I create a hard link to a file it looks like two files, but it's really one.
This is not the same as soft links or aliases, both names for the file are equally valid... you can delete either and the other is still there.
The only problem is that Finder doesn't let you do this. You can create an alias, but it's more combersome than it should be and it's not a hard link. You should be able to drag a file into another folder with some command key pressed and it would create a hard link for you.
if each file were in a database, with search keywords, I could find anything I wanted just as easily as anything else - there's no predetermined hierarchy that makes it easier to find some things than others.
But once you found it, you can show it in the finder and there's a folder right there with the rest of the information associated with the project.
It's not "broken", and Spotlight would work better as an augmentation tool for Finder, not a replacement.
Looks more like Perl crossed with the Plan 9 shell "rc".
I look forward to a real shell being a standard, shipping part of the OS!
They should just ship Interix with the OS instead of creating yet another new and incompatible interface.
If your Mac Tiger app is 64 bits, you're screwed.
Hands up everyone with a 64-bit app on Tiger. Not just "I built $thing for 64-bit OS X", but "This app requires 64-bit OS X to work" or "This app's a dog if it's not compiled 64-bit". If there's more than a couple in the world, I'll be very much surprised. If there's even one that won't compile right up on Linux-64 I'll be amazed.
OK, now hands up everyone with a 64-bit app on Windows, same caveat. There may be a few more, because Win32 has more of an address space bottleneck than other 32-bit environments.
And finally... 64-bit Linux, with the same restriction. This is where I expect to see the most hands.
Because, you know, we've been using Alpha for 10 years now, and there's very little code that really needs or even benefits from it. If you're not sure, without trying it, that your code's going to crater going "back" to 32-bit then your code was almost certainly not really using 64-bit mode.
300 megabytes of spam, not 300 million messages.
Probably less than 100,000 per month.
This would make it physically impossible for non-apple intel processors to boot OS X.
How will that stop someone from loading the install CD onto disk, replacing the boot with the one from Darwin x86 (which they already have), and burning a new install CD?
The video card is still the most likely place I can see a lock being hidden.
When the terms of the settlement were announced several people brought up the way they could be used against open-source software and Samba was frequently mentioned.
Why is anyone surprised?
Or if you don't have a choice.
:)
Granted.
That's one reason I registered my own domain so early on.
Then some dope at work got their laptop infected, and all of a sudden I was getting spam (my address was in their address book).
Oh yeh, that's a REALLY effective technique... for the bad guys. The spammers don't even need to cooperate with the virus authors (though I don't doubt the stories that they do), they just need to seed a dropbox address in a spam somewhere and wait for viruses to hit it.
There's really no good protection for that other than things like a dated address... mail to anything@slashdot.2004.taronga.com, for example, now routes to 127.0.0.1
the mouse is faster than arrow keys and drop menus
Well, ah, yeh, I suppose the mouse is faster for a mouse-based user interface. But it's not faster for a command-based user interface.
I get a similar amount of spam delivery attempts. What realy impresses me are the Nigerians. I get 419ers in clumps, and many of them have different typos... which implies they're typing these things in (presumably in some internet cafe in Lagos) over and over and over again...