I don't know what's the methodology behind those xlr8yourmac figures, but an Xserve G5 I administer gave me the following figures for PPC970FX 2.0GHz power draw: 12W at idle, 34W with a busyloop. (A busyloop obviously doesn't excercise all of the processor's subsystems, so it's conceivable that this is not a totally-maxed-out wattage figure.)
All in all, this compares pretty favourably with some of the P4 chips used in monster PC laptops today. So the real issue seems to be Apple's unwillingness to compromise on the PowerBooks' physical dimensions in order to accommodate the G5 chip.
Well, sure, but my original point was that a soultion based on market/economics considerations seems to me to be far more workable than one based on an ever-escalating spiral of technological measures and countermeasures.
Market economics have nothing to do with this. You make this sound like there's fair competition between the record labels and P2P music sharers. That's just plain false. Two distributors can't compete fairly if one of them is paying for the production and promotion of music and the other one doesn't have to.
The only things that can protect the labels are laws and DRM. You seem to argue that the laws should be enough, but really, will they ever be? I think it's pretty obvious that law and law enforcement (as they exist today) are doing next to nothing in preventing illegal music sharing.
Quality of the user experience is the only trump card Apple has. Apple isn't interested in licensing out technology if that means they're unable to control the quality. But if they are able to retain that control, they can deal -- look at the HP iPods.
"Freedom to innovate" in the consumer space largely means the freedom to see how badly you can fuck a customer over and still get away with it. If Apple was to license their music tech with no strings attached, the licensees would quickly exploit all obvious avenues of cost reduction: negotiate worse (for the customer) DRM terms with the labels, drop features from the portable player (two-row display is enough, right? a 2.5" hd is cheaper, and not THAT much bigger, right?), spend less on QA and software development.
From a business perspective, it's easy to argue that Jobs' unwillingness to compromise on user experience has led to the "failure" of both Apple and NeXT. But it's equally obvious that his customers are thanking him for staying the course.
The funny thing here is to see just how eager Mac fans are to bash Microsoft for this, yet every one of the knows that if a temp at Cupertino were to blog images of shipments arriving at Infinite Loop, he'd be gone faster than he could say "Steve's Wrath".
Which company do you guys think is more secretive?
I think many people here are missing the point of computers in schoolwork these days. Back when our generation was in elementary school, the few computers the schools had were there primarily to teach kids about computers.
But times have changed, and that's no longer the focus.
Those kids in Maine didn't get an iBook each just to learn how to use an iBook. They got the laptops to use them in class, to google up facts, to write essays, to edit short video presentations.
It's not about learning to use a computer. Believe me, the 10-year-olds of today are so computer savvy that they don't need a mouse usage primer. And if they later in life encounter a system with the widgets in slightly different places, the difference is trivial enough to be completely inconsequential to them.
You know, the 1.25GHz macs were never overclocked. Take off the heat sink and see "7455A 1250" on the chips. This rumor started to spread after Motorola didn't update their G4 specs to include 1250MHz availability. The likely explanation is that these parts are only available to Apple.
I admit that I haven't been following the latest developments in the field, but saying that recognizing speech and understanding the content are totally different things seems like an awfully bold statement.
I'd imagine that understanding the content is exactly what makes humans such cunning speech recognizers. Understanding can mean different things to different people, of course, but I'm sure that the first speech recognition machine to pass your test will contain at least some sort of semantic analysis to beef up the score.
All the technology isn't there. The Aqua widgets would still be prohibitively expensive to render without ready-made bitmaps.
For icons, OS X uses a scheme where each power-of-two size from 16x16 to 128x128 has a dedicated bitmap. The one closest to the display size is selected and final tweaking is done with interpolation.
This approach works for OS X icons because per the Aqua Human Interface Guidelines they are photorealistic. The blurring caused by the interpolation doesn't affect them; for widgets it would be unacceptable. This is why the icons can be scaled from bitmaps but widgets would have to be drawn as vector art.
My prediction is that instead of the utopia of totally boundless scaling, we will get the Palm resolution hack adapted to desktop systems. Just like in the 160x160 -> 320x320 migration of Palms, you can probably soon drive a 3840x2400 display with a "virtual" 1920x1200 desktop, with twice the letterform resolution, twice the widget resolution, and twice the OpenGL resolution. Bitmaps (on the web, etc) would just be pixel-doubled to appear the right size.
And then, with a special API, a program could tap into the physical resolution of the display and supply it with native pixels.
I don't know what you'll be paying for the SGI display, but since it hasn't exactly been the budget brand earlier I think the similar-spec Apple displays might be a cheaper alternative.
In short, Apple has a 1600x1024@22" display for $2499 and a 1920x1200@23" display for $3499.
Apple has a custom display connector for which you'll need a $149 adapter, but the proprietary format is quite useful: a single cable will carry power, DVI and USB to the display. This means that you can hide a noisy six fan Athlon beast in your closet (up to 15ft away with an extension) and only have one cable coming to desk, with your keyboard, mouse and speakers attached to the display. Neat!
Apple thinks they can change MPEG-LAs mind
on
More on MPEG4
·
· Score: 2
Its significant to note that Apple seems to be fairly confident that theyll be able to turn MPEG-LAs head and ditch the per-use royalties. Apple is happy with the per-copy charges, though, so I'm not sure what the implications are for open source MPEG-4 implementations.
The press release does a pretty good job at describing the situation. Apple is also encouraging everyone to send their (constructive) views on the issue to licensing@mpegla.com.
It seems to me that this is just an incremental advance from older satellite-to-satellite communications systems.
Military satellite networks, for example MILSTAR have already implemented very narrow beam communications between satellites. This has been necessary to prevent interception or jamming of the signal.
The advances here probably relate mostly to greater-precision mechanics and more powerful CPUs. I don't know if the data rate mentioned is a big leap or not, but considering the fact that the MILSTAR network carries all the photographic and video intelligence gathered by NRO's Improved CRYSTAL satellites the MILSTAR bandwidth must be pretty impressive too...
1. Spock's logic. The Star Trek Encyclopedia states that "Spock was raised with an older half-brother, Sybok, until Sybok was ostracized from Vulcan society because he rejected the Vulcan dogma of pure logic." This means that the "pure logic" isn't a feature of the Vulcan brain per se, but a norm of behavior in the Vulcan society. In other words, they just *try* to be as logical as possible.
2. Dilithium crystals are not used a power sources in the Star Trek universe. It is used to *regulate* the matter/antimatter reactions that provide the energy necessary to warp time-space.
3. Warp speed. Here's a quick roundup of warp factors' correlation to the speed of light:
Warp factor 1, 1 c
Warp factor 2, 10 c
Warp factor 3, 39 c
Warp factor 4, 102 c
Warp factor 5, 214 c
Warp factor 6, 392 c
Warp factor 7, 656 c
Warp factor 8, 1024 c
Warp factor 9, 1516 c
Warp factor 9.2, 1649 c
Warp factor 9.6, 1909 c
Warp factor 9.9, 3053 c
Warp factor 9.99, 7912 c
Warp factor 9.9999, 199516 c
Warp factor 10, infinite c
Of these speeds, factor 5 is the cruising speed of Enterprise-D. Its maximum rated speed is factor 9.6, although 9.9 can be maintained for 10 minutes. Warp factor 9.9999 is the propagation speed of subspace radio, and factor 10, obviously, can never be reached.
4. Plants on Enterprise. Someone already answered this one correctly. The ships life-support systems are fitted to provide breathable air without any biological components.
5. Artificial gravity. Starfleet vessels are fitted with gravitational units that generate the synthetic gravity field aboard the vessel. A key component of the unit is the graviton field generator, which, obviously, generates a field of gravitons; the elementary particles that transmit gravitational force in the Star Trek universe.
6. Beaming down. It is true that the physics of the transporter are pretty much out of this world. Without going into details, all the different physical problems of the concept seem to be taken care of by a separate component to the transporter system: some of the funniest components are the Heisenberg compensator (go figure) or the transporter's "pattern enhancer". The true story is that The Original Series' effects budget couldn't possibly cope with landing the Enterprise or even a shuttle in every episode. The transporter's instantaneus speed also helps to keep out mundane tasks like shuttle travel out of the show.
WTC Towers not designed to be evacuated
on
More WTC News
·
· Score: 3, Interesting
The towers were not designed to be *completely* evacuated in any situation. In case of fire, the plan was to only evacuate the fire floor, the floor above it, and three floors below.
Infrastructure for a timely evacuation of 20-30 thousand people would look a *lot* different than a staircase two persons wide.
I remember the 533MHz model drawing well north of 80W of power -- which when compared to the PPC750 is almost an order of magnitude more, clock for clock.
Apple has shipped at least five million units since those times; imagine if all of them would draw,say, 50W more than they do now? At three hours use per day, that'd be, what, more than 600GWh wasted every year...
"Wired Equivalent Privacy" isn't the only thing WEP was aiming for. They also aimed for Wired Equivalent Simplicity. A single encryption standard for WLANs wouldn't have been a bad thing, had they actually managed to specify a half-decent one.
If the link level security isn't simple and interoperable, then it doesn't really pay to have the security on that layer at all. As suggested earlier, upper layer security works just as well.
Or, alternatively, consider this translation: "It is a known fact that upstream bandwidth in a cable network is an extremely scarce resource. At the market's current price point, we are forced to have a modems-to-headend ratio that only permits a typical web surfing workload on the upstream. The decision to actually enforce the no-server policy was made only after empirical data was gathered, proving that even a single file-sharing server could severely disrupt the service level for hundreds of other customers."
(Disclaimer: I have no association with @home)
You might have a leg to stand on if @home was bringing in huge profits and denying you features just to bring in a cent more. But guess what, they aren't, and those downsides of cable modem service are precisely what's enabling them to offer it at the price you are paying now.
Don't like it? Tough. Go out and buy some real Internet bandwidth. It will cost you at least $200 per Mbps per month, in addition to the circuit costs.
Actually, I consider the color shift to be extremely noticeable, and because of this, I tend to turn any subpixel rendering options off. And yes, I've used the correct subpixel alignment value..
True -- if you aim for cost savings, you need to set up a new process and a manufacturing line for producing LCD panels with 1/3 of the sub-pixel density of a color screen.
But what if you went for resolution instead of savings? You could use more or less the same process, only with monochrome elements instead of R, G and B -- and have a display with three times the horizontal resolution of a color display.
4800x1200 resolution in a greyscale laptop screen? I could exchange color for that..
..at least if the past performance of Palm OS PDAs' processors is any indication, the Clie's CPU doesn't have what it takes to decode high-bitrate MP3 or Ogg audio.
And if we extrapolate from there, it seems obvious that Sony has included a dedicated MPEG audio decoder chip on the Clie. Your odds of hacking that to play Ogg audio are, as they say, slim to none.
Since this is "not the newest technology out there", one would think that it's possible to go out and buy one?
I'm actually in the process of building a local point-to-point radio link *right now*, so this interests me. We have a 3-mile distance we need to get at least a STM-1 over and NEC is quoting us a four-month lead time for their 26GHz STM-1 radio links. I can't wait that long, so I'm looking for alternatives. Can someone recommend a solution for STM-1 -level point-to-point connectivity that is available *today*?
I don't think that a particular license can not work when making a game. A great game doesn't cease to be create if it is released under the GPL.
Perhaps you meant to say that the classic Open Source development model doesn't really shine in game development. This, I think, is mostly true, because iteration and peer review doesn't really improve a work of art after, say, one iteration. Not everything can be succesfully designed in a committee.
On the other hand, I strongly disagree with the idea that the greatest games will always be developed commercially. If the merits of a game are reviewed based on the production value alone, then sure, we probably won't be seeing Lucasfilm-level cut scenes in a free game for a while.
But many great non-commercial games rival the commercial offerings in other areas. Just take a look at nethack and Planetarion, for instance. I'm sure others can come up with more examples of great, non-commercial games.
You can't just blindly compare the datarates of the DV codec and HDTV. The difference is that because the DV codec was designed for acquisition purposes, it only features intraframe compression. This allows the recorded footage to be easily editable while preserving (most of the) quality.
HDTV and MPEG were obviously designed for content delivery, where recompression and/or editing doesn't need to be lossless. This is why interframe compression can be used, resulting in significantly lower bitrates.
The difference is that conceivably the Carbon libraries of the OS contain ppc legacy code that isn't as easily ported to other processor architectures. The apps are as native either way, but the effort needed for a port to a new platform might very well be different.
Slashdot Mirror
Swatch sold its shares of Smart to Daimler-Benz AG in 1998. So Smart has been a fully-owned Daimler subsidiary for six years now.
I don't know what's the methodology behind those xlr8yourmac figures, but an Xserve G5 I administer gave me the following figures for PPC970FX 2.0GHz power draw: 12W at idle, 34W with a busyloop. (A busyloop obviously doesn't excercise all of the processor's subsystems, so it's conceivable that this is not a totally-maxed-out wattage figure.)
All in all, this compares pretty favourably with some of the P4 chips used in monster PC laptops today. So the real issue seems to be Apple's unwillingness to compromise on the PowerBooks' physical dimensions in order to accommodate the G5 chip.
The only things that can protect the labels are laws and DRM. You seem to argue that the laws should be enough, but really, will they ever be? I think it's pretty obvious that law and law enforcement (as they exist today) are doing next to nothing in preventing illegal music sharing.
Quality of the user experience is the only trump card Apple has. Apple isn't interested in licensing out technology if that means they're unable to control the quality. But if they are able to retain that control, they can deal -- look at the HP iPods.
"Freedom to innovate" in the consumer space largely means the freedom to see how badly you can fuck a customer over and still get away with it. If Apple was to license their music tech with no strings attached, the licensees would quickly exploit all obvious avenues of cost reduction: negotiate worse (for the customer) DRM terms with the labels, drop features from the portable player (two-row display is enough, right? a 2.5" hd is cheaper, and not THAT much bigger, right?), spend less on QA and software development.
From a business perspective, it's easy to argue that Jobs' unwillingness to compromise on user experience has led to the "failure" of both Apple and NeXT. But it's equally obvious that his customers are thanking him for staying the course.
The funny thing here is to see just how eager Mac fans are to bash Microsoft for this, yet every one of the knows that if a temp at Cupertino were to blog images of shipments arriving at Infinite Loop, he'd be gone faster than he could say "Steve's Wrath".
Which company do you guys think is more secretive?
I think many people here are missing the point of computers in schoolwork these days. Back when our generation was in elementary school, the few computers the schools had were there primarily to teach kids about computers.
But times have changed, and that's no longer the focus.
Those kids in Maine didn't get an iBook each just to learn how to use an iBook. They got the laptops to use them in class, to google up facts, to write essays, to edit short video presentations.
It's not about learning to use a computer. Believe me, the 10-year-olds of today are so computer savvy that they don't need a mouse usage primer. And if they later in life encounter a system with the widgets in slightly different places, the difference is trivial enough to be completely inconsequential to them.
You know, the 1.25GHz macs were never overclocked. Take off the heat sink and see "7455A 1250" on the chips. This rumor started to spread after Motorola didn't update their G4 specs to include 1250MHz availability. The likely explanation is that these parts are only available to Apple.
I admit that I haven't been following the latest developments in the field, but saying that recognizing speech and understanding the content are totally different things seems like an awfully bold statement.
I'd imagine that understanding the content is exactly what makes humans such cunning speech recognizers. Understanding can mean different things to different people, of course, but I'm sure that the first speech recognition machine to pass your test will contain at least some sort of semantic analysis to beef up the score.
All the technology isn't there. The Aqua widgets would still be prohibitively expensive to render without ready-made bitmaps.
For icons, OS X uses a scheme where each power-of-two size from 16x16 to 128x128 has a dedicated bitmap. The one closest to the display size is selected and final tweaking is done with interpolation.
This approach works for OS X icons because per the Aqua Human Interface Guidelines they are photorealistic. The blurring caused by the interpolation doesn't affect them; for widgets it would be unacceptable. This is why the icons can be scaled from bitmaps but widgets would have to be drawn as vector art.
My prediction is that instead of the utopia of totally boundless scaling, we will get the Palm resolution hack adapted to desktop systems. Just like in the 160x160 -> 320x320 migration of Palms, you can probably soon drive a 3840x2400 display with a "virtual" 1920x1200 desktop, with twice the letterform resolution, twice the widget resolution, and twice the OpenGL resolution. Bitmaps (on the web, etc) would just be pixel-doubled to appear the right size.
And then, with a special API, a program could tap into the physical resolution of the display and supply it with native pixels.
I don't know what you'll be paying for the SGI display, but since it hasn't exactly been the budget brand earlier I think the similar-spec Apple displays might be a cheaper alternative.
In short, Apple has a 1600x1024@22" display for $2499 and a 1920x1200@23" display for $3499.
Apple has a custom display connector for which you'll need a $149 adapter, but the proprietary format is quite useful: a single cable will carry power, DVI and USB to the display. This means that you can hide a noisy six fan Athlon beast in your closet (up to 15ft away with an extension) and only have one cable coming to desk, with your keyboard, mouse and speakers attached to the display. Neat!
Its significant to note that Apple seems to be fairly confident that theyll be able to turn MPEG-LAs head and ditch the per-use royalties. Apple is happy with the per-copy charges, though, so I'm not sure what the implications are for open source MPEG-4 implementations.
Apples QuickTime 6 press release
The press release does a pretty good job at describing the situation. Apple is also encouraging everyone to send their (constructive) views on the issue to licensing@mpegla.com.
Well, I can open a shell, ssh to my Mom's iBook and get some real work done! :)
It seems to me that this is just an incremental advance from older satellite-to-satellite communications systems.
Military satellite networks, for example MILSTAR have already implemented very narrow beam communications between satellites. This has been necessary to prevent interception or jamming of the signal.
The advances here probably relate mostly to greater-precision mechanics and more powerful CPUs. I don't know if the data rate mentioned is a big leap or not, but considering the fact that the MILSTAR network carries all the photographic and video intelligence gathered by NRO's Improved CRYSTAL satellites the MILSTAR bandwidth must be pretty impressive too...
1. Spock's logic. The Star Trek Encyclopedia states that "Spock was raised with an older half-brother, Sybok, until Sybok was ostracized from Vulcan society because he rejected the Vulcan dogma of pure logic." This means that the "pure logic" isn't a feature of the Vulcan brain per se, but a norm of behavior in the Vulcan society. In other words, they just *try* to be as logical as possible.
2. Dilithium crystals are not used a power sources in the Star Trek universe. It is used to *regulate* the matter/antimatter reactions that provide the energy necessary to warp time-space.
3. Warp speed. Here's a quick roundup of warp factors' correlation to the speed of light:
Of these speeds, factor 5 is the cruising speed of Enterprise-D. Its maximum rated speed is factor 9.6, although 9.9 can be maintained for 10 minutes. Warp factor 9.9999 is the propagation speed of subspace radio, and factor 10, obviously, can never be reached.
4. Plants on Enterprise. Someone already answered this one correctly. The ships life-support systems are fitted to provide breathable air without any biological components.
5. Artificial gravity. Starfleet vessels are fitted with gravitational units that generate the synthetic gravity field aboard the vessel. A key component of the unit is the graviton field generator, which, obviously, generates a field of gravitons; the elementary particles that transmit gravitational force in the Star Trek universe.
6. Beaming down. It is true that the physics of the transporter are pretty much out of this world. Without going into details, all the different physical problems of the concept seem to be taken care of by a separate component to the transporter system: some of the funniest components are the Heisenberg compensator (go figure) or the transporter's "pattern enhancer". The true story is that The Original Series' effects budget couldn't possibly cope with landing the Enterprise or even a shuttle in every episode. The transporter's instantaneus speed also helps to keep out mundane tasks like shuttle travel out of the show.
The towers were not designed to be *completely* evacuated in any situation. In case of fire, the plan was to only evacuate the fire floor, the floor above it, and three floors below.
Infrastructure for a timely evacuation of 20-30 thousand people would look a *lot* different than a staircase two persons wide.
Environment 1 - Exponential 0!
I remember the 533MHz model drawing well north of 80W of power -- which when compared to the PPC750 is almost an order of magnitude more, clock for clock.
Apple has shipped at least five million units since those times; imagine if all of them would draw,say, 50W more than they do now? At three hours use per day, that'd be, what, more than 600GWh wasted every year...
"Wired Equivalent Privacy" isn't the only thing WEP was aiming for. They also aimed for Wired Equivalent Simplicity. A single encryption standard for WLANs wouldn't have been a bad thing, had they actually managed to specify a half-decent one.
If the link level security isn't simple and interoperable, then it doesn't really pay to have the security on that layer at all. As suggested earlier, upper layer security works just as well.
Or, alternatively, consider this translation: "It is a known fact that upstream bandwidth in a cable network is an extremely scarce resource. At the market's current price point, we are forced to have a modems-to-headend ratio that only permits a typical web surfing workload on the upstream. The decision to actually enforce the no-server policy was made only after empirical data was gathered, proving that even a single file-sharing server could severely disrupt the service level for hundreds of other customers."
(Disclaimer: I have no association with @home)
You might have a leg to stand on if @home was bringing in huge profits and denying you features just to bring in a cent more. But guess what, they aren't, and those downsides of cable modem service are precisely what's enabling them to offer it at the price you are paying now.
Don't like it? Tough. Go out and buy some real Internet bandwidth. It will cost you at least $200 per Mbps per month, in addition to the circuit costs.
Actually, I consider the color shift to be extremely noticeable, and because of this, I tend to turn any subpixel rendering options off. And yes, I've used the correct subpixel alignment value..
True -- if you aim for cost savings, you need to set up a new process and a manufacturing line for producing LCD panels with 1/3 of the sub-pixel density of a color screen. But what if you went for resolution instead of savings? You could use more or less the same process, only with monochrome elements instead of R, G and B -- and have a display with three times the horizontal resolution of a color display. 4800x1200 resolution in a greyscale laptop screen? I could exchange color for that..
..at least if the past performance of Palm OS PDAs' processors is any indication, the Clie's CPU doesn't have what it takes to decode high-bitrate MP3 or Ogg audio.
And if we extrapolate from there, it seems obvious that Sony has included a dedicated MPEG audio decoder chip on the Clie. Your odds of hacking that to play Ogg audio are, as they say, slim to none.
Since this is "not the newest technology out there", one would think that it's possible to go out and buy one?
I'm actually in the process of building a local point-to-point radio link *right now*, so this interests me. We have a 3-mile distance we need to get at least a STM-1 over and NEC is quoting us a four-month lead time for their 26GHz STM-1 radio links. I can't wait that long, so I'm looking for alternatives. Can someone recommend a solution for STM-1 -level point-to-point connectivity that is available *today*?
I don't think that a particular license can not work when making a game. A great game doesn't cease to be create if it is released under the GPL.
Perhaps you meant to say that the classic Open Source development model doesn't really shine in game development. This, I think, is mostly true, because iteration and peer review doesn't really improve a work of art after, say, one iteration. Not everything can be succesfully designed in a committee.
On the other hand, I strongly disagree with the idea that the greatest games will always be developed commercially. If the merits of a game are reviewed based on the production value alone, then sure, we probably won't be seeing Lucasfilm-level cut scenes in a free game for a while.
But many great non-commercial games rival the commercial offerings in other areas. Just take a look at nethack and Planetarion, for instance. I'm sure others can come up with more examples of great, non-commercial games.
You can't just blindly compare the datarates of the DV codec and HDTV. The difference is that because the DV codec was designed for acquisition purposes, it only features intraframe compression. This allows the recorded footage to be easily editable while preserving (most of the) quality.
HDTV and MPEG were obviously designed for content delivery, where recompression and/or editing doesn't need to be lossless. This is why interframe compression can be used, resulting in significantly lower bitrates.
The difference is that conceivably the Carbon libraries of the OS contain ppc legacy code that isn't as easily ported to other processor architectures. The apps are as native either way, but the effort needed for a port to a new platform might very well be different.