Yeah, Apple was a pretty small customer for Freescale (and a very demanding one, as well). I think GM was buying at least 10x the number of chips Apple was, at the peak of Apple's consumption.
It seems like the editors mostly agree with the Slashdot crowd, regarding games which should have been on the list but weren't.
On the following pages, we'll first present eight "honorable mentions" - games that, while certainly innovative and important, did not receive enough votes to make it into the top echelon.
Following this, we'll present the top five first-person shooters voted for by our readers, in reverse order, ending with the overall receipient of Gamasutra's first Quantum Leap Award, which received the largest amount of votes from game professionals.
So, they tallied up the votes, then decided that the results were totally stupid, and put some "honorable mentions" at the front of the list, including: Doom System Shock (the first one) Marathon Deus Ex Alien vs Predator Descent Tribes
I think I like the "honorable mentions" list better than the actual top-5 list. I wonder if part of the problem was people not voting for Doom because it was "too obvious", and the wanted to pick something different.
From the linked page - 1154000 ± 17% 4.3731% 1. Main Page 37500 ± 95% 0.1421% 2. Wikipedia 35500 ± 97% 0.1345% 3. Pluto 23000 ± 121% 0.0872% 4. Hurricane Katrina
Does the ± mean something different in Germany? Otherwise, these statistics don't seem to make any sense - how can you have (+121%, -121%) error bounds on a measurement?
There were somewhere between 73830 and -4830 views of "Hurricane Katrina"? How can you have negative page views?
The key point is that anyone at Apple who has a good reason to be running early builds of Leopard already has access to it. Because early OS builds are buggy, people who don't need to run them as part of their job probably shouldn't run them at all. Retail Store employees have no plausible reason to be running pre-release versions of an OS that isn't shipping for 6 months to a year.
Anybody at Apple who's "working on" Leopard in any sense already has access to regular weekly builds of the whole OS. They couldn't get their job done otherwise. In general, this software distribution is limited to people who actually have a NEED to run pre-release software. Giving out early releases of OS software to anybody at the company that's merely CURIOUS about it would be counter-productive.
Early builds of any software are going to be buggy, more or less by definition. Some of those bugs are going to be really bad - erasing disks, corrupting documents, leaving your Mac unbootable, etc. The loss of productivity isn't going to be worth it for "regular users" until the OS is pretty far along.
Let's make sure we're on the same page. In order to perform the "meet in the middle" attack, you first need to have a known Plaintext and Ciphertext pair, P and C. In our example, we'll assume these are 64 bit blocks, or 8 bytes. We're also assuming that the keys are 64 bits.
For each possible 64-bit key K, you need to generate and store Ek(P) in a table. It's the size of that table I was trying to calculate. So that's 8 bytes * 2^64 keys, which is 147,573,952,589,676,412,928. That sure looks like 147 Billion Billion to me.
In any case, even using your figure of 2.2 Billion GB, you've still got the number of drives very wrong. When you divide 2.2 Billion by 750, you get 2.9 million drives required. First of all, that's probably larger than the total production run of those drives. Second, it's much more like $1.1 Billion, rather than $1.2 Million, in cost.
I don't think hard drives are going to be the way to go. A decent tape library can hold 500 Terabytes or so, so you'd only need to buy (and power, and house) a mere 300,000 of those to hold the whole table.
From the Wikipedia article: If the keysize is n, this attack uses only 2^(n + 1) encryptions (and O(2^n) space) in contrast to the naive attack, which needs 2^2n encryptions (but only O(1) space).
So for a 64-bit key and 64-bit block size, you'd need about 2^64 * 8 bytes of storage, which is 148 Billion Gigabytes. I doubt you could find any off-the-shelf storage solution you could use for that.
Do you happen to have a reference you can cite for this? I've done controller board swaps in the past, and been successful in recovering data, but perhaps more modern controller boards are more finicky. I can certainly believe that some amount of the bad sector mapping is done by the controller board, but my impression was that most of that information was encoded in the servo tracks on the platter itself.
What makes you think Mac OS X 10.5 is going to have a 64-bit kernel? Again, the currently shipping version of Mac OS X supports 64-bit applications just fine with a 32-bit* kernel and drivers.
-Mark
* Okay, so there's a *very small* part of the kernel that's aware of the expanded address space. But the kernel itself runs in a 32-bit memory space, in 32-bit mode.
The PowerPC version of Mac OS X 10.4 has 32 bit drivers and kernel, and supports running both 32 and 64-bit applications seamlessly. I expect that theat level of support was a minimum requirement for Intel 64-bit support on Leopard. I don't know enough about the addressing mode differences in the Intel world to know if the same strategy would work there.
Oh, okay. It wasn't clear to me from the original message whether the information was Mac-market specific. That's interesting. Maybe the Transgaming folks are on the right track, then. I'll still be bummed when Intel-only Mac games start appearing, though.
Of course, by the time that actually happens, I'm quite likely going to have an Intel Mac anyway - they're awfully tempting.
If Apple is as successful with the Intel machines as they hope, the day when Intel-based Macs outnumber PowerPC Macs will no doubt arrive within a few years. However, there are still a lot of PowerPC Macs out there, and will be for some time. There are still Macs out there that aren't even running Mac OS X yet...
I think the folks porting games to Mac OS X will still be around for a while.
I did think about the "early adopter==gamer" thing, but I wasn't sure whether that translated to the Mac market. It's certainly true that in the PC world, "serious" gamers are constantly updating their machines, so there's obviously a strong correlation between hardware purchases and game purchases. Among the Macintosh users I know, the rate of upgrades is considerably slower, and there really aren't any of the crazy gamer types that drive the PC hardware business (at least there aren't yet).
Do you happen to know if anybody's done research about the validity of that model in the Mac market, specifically? All I've got is anecdotal evidence.
Okay, I understand that this guy's trying to sell the solution that his company produces. But it's pretty weird to say that these sorts of API translation technologies will be "the way" to bring games to the Mac when Intel-based Macs are a tiny minority of the total Macintosh user base.
I understand it's a lot less effort for the game developer to utilize something like this technology rather than porting the game to native MacOS X. But to the extent that game publishers claim that the Mac market is "too small" to justify porting games, I can't see how a small fraction of that too-small market is going to look any better.
I'm sure they'll claim that this is a zero-effort solution to supporting th Mac, and it's therefore 100% upside to add this in and get a few hundred sales to Intel-based Mac users. I'm sceptical that's really going to work out.
Sure, you can play a game of Titan against computer opponents in half an hour, but that's not a realistic comparison with playing the actual boradgame, because of the following differences:
1. The computer players in Colossus are pretty dumb. 2. Dice rolling and counting is immensely faster with the computer doing it. 3. Same thing for setting up the battle maps. 4. Perhaps most importantly, you can play very agressively when you're the only human player. If you lose, you don't have to wait 5 hours for the game to finish before you can play again.
It's not just me - read any review of Titan you can find on the web. Lots of people complain about it taking 6-12 hours to play with more than two players.
Yeah, that's right - worse than Monopoly, even. I just can't understand why some people seem to love it so much. Some of my friends play every now and then, and I've joined them a few times. Fundamentally, I suppose I'm unwilling to devote 8 to 12 HOURS of my time to a single game of Titan, when I could play multiple rounds of any other game in the same amount of time.
I suppose if there were some kind of payoff to be had after all that investment, it'd be one thing, but the climactic final battle between armies is just like every smaller skirmish before it, except with more dice.
Amusingly, the back of the rulebook has a short history of the game, explaining all the changes that were made to the rules to make the game SHORTER. I imagine that the first game played by the original rules is actually STILL GOING ON somewhere, even today.
Only folks attending WWDC have the XCode 3.0 pre-release. They're all covered by a Non-Disclosure Agreement, so they can't (legally) say anything about it. Some information will doubtless leak out, but you won't see anything like a full review of it until it ships (presumably with Leopard).
Consider that Intel's current product line has single-core and dual-core processors, and that they're likely to ramp that up to 8-16 cores in a few years (they've said as much, in public statements).
When Core Quad and Core Octo (I made those names up) chips are available, you might see single-chip systems with 8 cores, and dual-chip systems with 4 cores if they're still using Core Duo for some systems.
It's much less confusing to count up the total number of processor cores, and just give people that number. That way, customers don't need a lot of technical know-how to know if the "dual-chip" system is faster than the "single-chip" system, or vice-versa.
The dual-core Xeons (and other dual-core chips) have two (more-or-less) independent processors packaged onto a single piece of Silicon. Apple calls these "quad processor" systems, because as far as the applications can tell, there are four independent processors to work with. The OS may do some things differently in the scheduling department based on which processors are physically located close together, and then again it may not.
Two dual-core processors are essentially equivalent to 4 single-core processors in terms of throughput, so it's a reasonable way to refer to things.
I gotta say that Windows' System Restore and Time Machine are VERY different in intent. System Restore only tracks a subset of changes, and is intended to rescue your system from an unbootable/unusable state after installing some piece of software that completely screws up the system.
System Restore doesn't do anything to preserve your documents, which is more-or-less the whole point of Time Machine. You CAN use Time Machine to roll back the system software, but the intent is mostly to save you from yourself, when you save the wrong version of a file or accidentally delete something.
The article is scant on details
on
Growing Insulin
·
· Score: 4, Interesting
Not much technical detail at all in the article, but from previous articles on splicing human genes into animals and plants, I gather that the real advantage of this technique is that, if it's done right, the plant or animal concentrates the product in one place for you (the seeds, in this case).
Compared to a big churning vat of E. Coli, their food, and their waste products, it's easier to use something like an impeller or centrifuge to separate out bulk quantities of insulin-laced oil from a plant, and then purify it from there.
Also, those same fermenters that are growing insulin currently are also a great breeding ground for other bacteria, molds, and yeasts in the environment. I'm given to understand that keeping unwanted organisms out of the vats is part of what makes the process difficult and expensive. A single bacterium or wild yeast spore that gets into a fermenter can ruin the whole batch.
The macroscopic plants are a much more robust system - they can still get sick of course, but it's fairly easy to keep plants healthy. And even given that you can't spray these plants willy-nilly with pesticides, organic farmers get pretty good yields on most products with just natural controls.
I like having the Windows apps in a penalty box. A couple more months in there to think about what choices they might have made differently, and they'll then be ready to come back out and rejoin civilized society.
I think when this monster comes screaming in at several times the speed of sound, scattering thousands of pounds of explosives behind it in an orgy of death and destruction, "Silent" is probably not the adjective that the survivors (if any) will use to describe it.
Two quick points. 1. That's CNN's article title, so that's hardly the Slashdot editors' fault. 2. It's unmanned. Unmanned combat aircraft are used almost exclusively for assassinations, at least currently.
Slashdot Mirror
Yeah, Apple was a pretty small customer for Freescale (and a very demanding one, as well). I think GM was buying at least 10x the number of chips Apple was, at the peak of Apple's consumption.
Those were the days. "Did I just press 'R', and lose two hours of work? Rats."
So, they tallied up the votes, then decided that the results were totally stupid, and put some "honorable mentions" at the front of the list, including:
Doom
System Shock (the first one)
Marathon
Deus Ex
Alien vs Predator
Descent
Tribes
I think I like the "honorable mentions" list better than the actual top-5 list. I wonder if part of the problem was people not voting for Doom because it was "too obvious", and the wanted to pick something different.
From the linked page -
1154000 ± 17% 4.3731% 1. Main Page
37500 ± 95% 0.1421% 2. Wikipedia
35500 ± 97% 0.1345% 3. Pluto
23000 ± 121% 0.0872% 4. Hurricane Katrina
Does the ± mean something different in Germany? Otherwise, these statistics don't seem to make any sense - how can you have (+121%, -121%) error bounds on a measurement?
There were somewhere between 73830 and -4830 views of "Hurricane Katrina"? How can you have negative page views?
-Mark
I wrote a reply to a similar question earlier:c id=15972803
http://apple.slashdot.org/comments.pl?sid=194869&
The key point is that anyone at Apple who has a good reason to be running early builds of Leopard already has access to it. Because early OS builds are buggy, people who don't need to run them as part of their job probably shouldn't run them at all. Retail Store employees have no plausible reason to be running pre-release versions of an OS that isn't shipping for 6 months to a year.
Anybody at Apple who's "working on" Leopard in any sense already has access to regular weekly builds of the whole OS. They couldn't get their job done otherwise. In general, this software distribution is limited to people who actually have a NEED to run pre-release software. Giving out early releases of OS software to anybody at the company that's merely CURIOUS about it would be counter-productive.
Early builds of any software are going to be buggy, more or less by definition. Some of those bugs are going to be really bad - erasing disks, corrupting documents, leaving your Mac unbootable, etc. The loss of productivity isn't going to be worth it for "regular users" until the OS is pretty far along.
-Mark
Let's make sure we're on the same page. In order to perform the "meet in the middle" attack, you first need to have a known Plaintext and Ciphertext pair, P and C. In our example, we'll assume these are 64 bit blocks, or 8 bytes. We're also assuming that the keys are 64 bits.
For each possible 64-bit key K, you need to generate and store Ek(P) in a table. It's the size of that table I was trying to calculate. So that's 8 bytes * 2^64 keys, which is 147,573,952,589,676,412,928. That sure looks like 147 Billion Billion to me.
In any case, even using your figure of 2.2 Billion GB, you've still got the number of drives very wrong. When you divide 2.2 Billion by 750, you get 2.9 million drives required. First of all, that's probably larger than the total production run of those drives. Second, it's much more like $1.1 Billion, rather than $1.2 Million, in cost.
I don't think hard drives are going to be the way to go. A decent tape library can hold 500 Terabytes or so, so you'd only need to buy (and power, and house) a mere 300,000 of those to hold the whole table.
From the Wikipedia article:
If the keysize is n, this attack uses only 2^(n + 1) encryptions (and O(2^n) space) in contrast to the naive attack, which needs 2^2n encryptions (but only O(1) space).
So for a 64-bit key and 64-bit block size, you'd need about 2^64 * 8 bytes of storage, which is 148 Billion Gigabytes. I doubt you could find any off-the-shelf storage solution you could use for that.
OR am I missing something in my calculation?
Do you happen to have a reference you can cite for this? I've done controller board swaps in the past, and been successful in recovering data, but perhaps more modern controller boards are more finicky. I can certainly believe that some amount of the bad sector mapping is done by the controller board, but my impression was that most of that information was encoded in the servo tracks on the platter itself.
-Mark
What makes you think Mac OS X 10.5 is going to have a 64-bit kernel? Again, the currently shipping version of Mac OS X supports 64-bit applications just fine with a 32-bit* kernel and drivers.
-Mark
* Okay, so there's a *very small* part of the kernel that's aware of the expanded address space. But the kernel itself runs in a 32-bit memory space, in 32-bit mode.
The PowerPC version of Mac OS X 10.4 has 32 bit drivers and kernel, and supports running both 32 and 64-bit applications seamlessly. I expect that theat level of support was a minimum requirement for Intel 64-bit support on Leopard. I don't know enough about the addressing mode differences in the Intel world to know if the same strategy would work there.
Oh, okay. It wasn't clear to me from the original message whether the information was Mac-market specific. That's interesting. Maybe the Transgaming folks are on the right track, then. I'll still be bummed when Intel-only Mac games start appearing, though.
Of course, by the time that actually happens, I'm quite likely going to have an Intel Mac anyway - they're awfully tempting.
If Apple is as successful with the Intel machines as they hope, the day when Intel-based Macs outnumber PowerPC Macs will no doubt arrive within a few years. However, there are still a lot of PowerPC Macs out there, and will be for some time. There are still Macs out there that aren't even running Mac OS X yet...
I think the folks porting games to Mac OS X will still be around for a while.
I did think about the "early adopter==gamer" thing, but I wasn't sure whether that translated to the Mac market. It's certainly true that in the PC world, "serious" gamers are constantly updating their machines, so there's obviously a strong correlation between hardware purchases and game purchases. Among the Macintosh users I know, the rate of upgrades is considerably slower, and there really aren't any of the crazy gamer types that drive the PC hardware business (at least there aren't yet).
Do you happen to know if anybody's done research about the validity of that model in the Mac market, specifically? All I've got is anecdotal evidence.
Okay, I understand that this guy's trying to sell the solution that his company produces. But it's pretty weird to say that these sorts of API translation technologies will be "the way" to bring games to the Mac when Intel-based Macs are a tiny minority of the total Macintosh user base.
I understand it's a lot less effort for the game developer to utilize something like this technology rather than porting the game to native MacOS X. But to the extent that game publishers claim that the Mac market is "too small" to justify porting games, I can't see how a small fraction of that too-small market is going to look any better.
I'm sure they'll claim that this is a zero-effort solution to supporting th Mac, and it's therefore 100% upside to add this in and get a few hundred sales to Intel-based Mac users. I'm sceptical that's really going to work out.
-Mark
Sure, you can play a game of Titan against computer opponents in half an hour, but that's not a realistic comparison with playing the actual boradgame, because of the following differences:
1. The computer players in Colossus are pretty dumb.
2. Dice rolling and counting is immensely faster with the computer doing it.
3. Same thing for setting up the battle maps.
4. Perhaps most importantly, you can play very agressively when you're the only human player. If you lose, you don't have to wait 5 hours for the game to finish before you can play again.
It's not just me - read any review of Titan you can find on the web. Lots of people complain about it taking 6-12 hours to play with more than two players.
Yeah, that's right - worse than Monopoly, even. I just can't understand why some people seem to love it so much. Some of my friends play every now and then, and I've joined them a few times. Fundamentally, I suppose I'm unwilling to devote 8 to 12 HOURS of my time to a single game of Titan, when I could play multiple rounds of any other game in the same amount of time.
I suppose if there were some kind of payoff to be had after all that investment, it'd be one thing, but the climactic final battle between armies is just like every smaller skirmish before it, except with more dice.
Amusingly, the back of the rulebook has a short history of the game, explaining all the changes that were made to the rules to make the game SHORTER. I imagine that the first game played by the original rules is actually STILL GOING ON somewhere, even today.
Only folks attending WWDC have the XCode 3.0 pre-release. They're all covered by a Non-Disclosure Agreement, so they can't (legally) say anything about it. Some information will doubtless leak out, but you won't see anything like a full review of it until it ships (presumably with Leopard).
There's a few additional details and a screenshot up on Apple's site:
http://www.apple.com/macosx/leopard/xcode.html
-Mark
Consider that Intel's current product line has single-core and dual-core processors, and that they're likely to ramp that up to 8-16 cores in a few years (they've said as much, in public statements).
When Core Quad and Core Octo (I made those names up) chips are available, you might see single-chip systems with 8 cores, and dual-chip systems with 4 cores if they're still using Core Duo for some systems.
It's much less confusing to count up the total number of processor cores, and just give people that number. That way, customers don't need a lot of technical know-how to know if the "dual-chip" system is faster than the "single-chip" system, or vice-versa.
The dual-core Xeons (and other dual-core chips) have two (more-or-less) independent processors packaged onto a single piece of Silicon. Apple calls these "quad processor" systems, because as far as the applications can tell, there are four independent processors to work with. The OS may do some things differently in the scheduling department based on which processors are physically located close together, and then again it may not.
Two dual-core processors are essentially equivalent to 4 single-core processors in terms of throughput, so it's a reasonable way to refer to things.
I gotta say that Windows' System Restore and Time Machine are VERY different in intent. System Restore only tracks a subset of changes, and is intended to rescue your system from an unbootable/unusable state after installing some piece of software that completely screws up the system.
System Restore doesn't do anything to preserve your documents, which is more-or-less the whole point of Time Machine. You CAN use Time Machine to roll back the system software, but the intent is mostly to save you from yourself, when you save the wrong version of a file or accidentally delete something.
Not much technical detail at all in the article, but from previous articles on splicing human genes into animals and plants, I gather that the real advantage of this technique is that, if it's done right, the plant or animal concentrates the product in one place for you (the seeds, in this case).
Compared to a big churning vat of E. Coli, their food, and their waste products, it's easier to use something like an impeller or centrifuge to separate out bulk quantities of insulin-laced oil from a plant, and then purify it from there.
Also, those same fermenters that are growing insulin currently are also a great breeding ground for other bacteria, molds, and yeasts in the environment. I'm given to understand that keeping unwanted organisms out of the vats is part of what makes the process difficult and expensive. A single bacterium or wild yeast spore that gets into a fermenter can ruin the whole batch.
The macroscopic plants are a much more robust system - they can still get sick of course, but it's fairly easy to keep plants healthy. And even given that you can't spray these plants willy-nilly with pesticides, organic farmers get pretty good yields on most products with just natural controls.
I like having the Windows apps in a penalty box. A couple more months in there to think about what choices they might have made differently, and they'll then be ready to come back out and rejoin civilized society.
I think when this monster comes screaming in at several times the speed of sound, scattering thousands of pounds of explosives behind it in an orgy of death and destruction, "Silent" is probably not the adjective that the survivors (if any) will use to describe it.
Two quick points.
1. That's CNN's article title, so that's hardly the Slashdot editors' fault.
2. It's unmanned. Unmanned combat aircraft are used almost exclusively for assassinations, at least currently.