Now that these giant computer manufacturers are preloading Linux, what incentive does Joe Consumer have to buy from a company like Penguin? Those companies found a niche and started selling products. They were here first! Are they just going to get shoved aside or remain small? Or do you think that they still have a chance to become major manufacturers while the older companies are still getting their acts together?
They might, but it would be an uphill battle. The older companies have name recognition and past contracts which only need to be renewed, which count for a lot.
The main selling point of, say, Penguin is that Penguin presumably has more experience building and configuring Linux systems, and so (hopefully) does a better job at it.
In order to become a major player, a company like Penguin would have to:
Build an ad campaign around their expertise.
Spend lots and lots of money pushing this ad campaign like crazy across television and computer magazines to get comsumer awareness up.
Spend lots of money pushing this ad campaign in trade magazines where retailers, OEMs, and businesses' purchasing managers will see it.
Spend lots and lots of effort approaching small- to medium-sized retailers offering prebuilt Linux systems with the Penguin logo, to gain sales volume and credibility.
This can be done, and has a chance of working, but takes a lot of effort and a very large amount of money. There would also be the risk of failing spectacularly, as there are a glut of suppliers on the market.
A quick check at CompUSA Onlineshows that full suite price to be $1980.95 and upgrade price is $931.85.
Doing my own search, I find $225 for the stripped down version, and $490 for the full version of Corel Draw alone.
High, but not nearly as high as the figure you quoted.
This is typical of office software, and quite reasonable when compared to the cost of the machine it's going to be used on or of the employee who's going to be using it.
There are almost certianly student versions available for a much lower price, too (around here, student versions are half off or better).
Also, your hyperlink seems to have been munged. Further inspection reveals that they're using some kind of bizzare scratch keys to encode query data, making linking to specific results unreliable.
I'd suspect that the disk subsystem would be the main bottleneck for compiling, followed by the memory subsystem.
When you do a compile, you walk through many, many source files and nested levels of include files (take a look at your dependencies list for an idea of how nasty this is). A fast disk with scatter-gather (coughcoughSCSI) and a large disk cache in memory will work wonders for eliminating that grinding noise that tends to accompany compiles.
Now, the memory subsystem. Compilation with a modern optimizing compiler involves building monstrously huge structures in memory and shuffling them around (I spent much of the last two semesters writing such a compiler). There's enough data here to overflow the cache, so you run into the system memory access speed bottleneck.
100 MHz FSB is Your Friend. 133 is even better, if you have RAM with low enough latency (another poster correctly pointed out that 133 MHz does you no good if wait states eat the extra speed). A larger chip cache will also help, if that's an option.
Lastly, if neither of these seems to be the bottleneck, having a faster CPU or a second CPU may help. Just make sure that the CPU is the bottleneck before forking out the money for a faster one. You can also safely ignore floating-point performance - almost all of the computation performed by a compiler is integer math (building bit vectors to calculate dominators and so forth with).
As far as benchmarks go, the disk- and memory-limited ones should still be useful to you.
Err... did you bother to read the link he gave to magram? No? Well, apparently, magram is faster than DRAM, non-volotile, and not much more expensive that hard drive space... or something...
Yes, I've read the link to magram.
This is a non-technical article describing what they *hope* magram will accomplish when the technology matures.
It is *not* a description of what they actually have working.
There are many, many interesting memory technologies in this stage of development. They only become relevant to debates when engineering prototypes with decent density and speed are produced.
We've become sloppy and trained to the notion that memory should be divided into segments varying by speed, size, volatility, and cost. We all spend months in college or in the field learning about the subtle differences between L1 caching, L2 caching, main memory, hard drive memory, ROM, and the trillion variations of RAM. We don't see the forest for the trees: this model of data management is the single most crucial hindrance to the advancement of computer science in our entire industry.
The problem is that this multilevel approach to memory exists for a very good reason - all serious candidates for implementing RAM have an inverse relationship between speed and cost.
Sure, you could put a gig of SRAM in your machine, but you'd have to spend a pretty penny to do it. This isn't because of some evil manufacturer jacking up prices - it's because SRAM is many times less dense than DRAM, and you're paying for the silicon area that you use.
There's also an inverse relationship between size and speed. This is why a chip's L2 SRAM is slower than its L1 SRAM. The larger RAM array has longer traces and a number of other features that increase settling time and in other ways make the response time longer.
Similarly, even if you did decide to fill your motherboard with SRAM or something comparably fast, you wouldn't be able to access it at full speed. Motherboard traces are long, which increases signal propagation latency, and noise injection from the environment is worse, which decreases the rate at which you can transfer data even if you can tolerate high latency.
These signalling problems aren't going to go away, and won't change even with different RAM types. At best, you can learn to live with them (RAMBus uses a clever scheme to reduce noise, allowing them a faster throughput while keeping the lousy latency of external memory).
As for using disk as memory or vice versa - this just isn't practical. Disk is a thousand times slower than memory and a hundred times cheaper. Take out most of your RAM and watch your machine swap to see what using this as system memory is like. Check out the price tags for solid state drives to see what the cost of using memory as a drive is like (or just ask your local store how much 30 gigs of DIMMs would cost).
In conclusion, I think that your suggestion is not compatible with existing or proposed storage technologies.
Some time ago, when a friend of mine had a cell phone and I didn't, I'd send him text messages via a web page helpfully provided by Fido (the company selling us the service).
This was very useful, but is trivially easy to spam via scripts. My friend even wrote such a script, to forward email from his account to his phone (before purchasing phone email service).
It would only take one or two knowledgeable people saying "hey, that's neat!" to do that here in Toronto, and I'm sure Fido isn't the only company set up this way.
I made it a few pages down the FAQ, and I'm afraid I have to say that it's one of the worst that I've ever seen. The author's argument seems to amount to "Well, you can describe the universe mathematically using four dimensions, but because I only see three of them this is purely bunk.".
At this point, I stopped reading.
In point of fact, if you want to see the effects of space being four-dimensional, you need only look at two reference that are moving quickly with respect to each other. The time and space directions measured by observers in each frame are different - this is the Lorentz transform (if I remember the term correctly). Space in one frame corresponds to a skewed space-and-time axis in another frame.
This has been verified experimentally by very careful measurements of atomic clocks moving at different speeds with respect to each other. A more dramatic illustration is measuring the decay times of unstable particles moving at different speeds. As they approach the speed of light, the lifetime as measured by the observer gets longer. This is called "time dilation", and is one of many effects caused by the "time" and "space" axes not being the same for observers moving with respect to each other.
There's still some argument over whether all particles are influenced by gravity; I seem to remember recently that there is talk of the gravitational lens effect (ie light being affected by gravity) being disproved.
That would be quite a trick. If I remember correctly, one of the measurements used to support relativity was of gravitational lensing of starlight by the sun (mask out the disc and take careful measurements of the apparent positions of stars next to it).
Gravitational lensing studies are also being performed to measure the mass of distant galaxies (by looking at multiple images of even more distant galaxies). If gravitational lensing didn't exist, it would have been noticed in a big way by these studies.
As far as anyone's been able to determine, photons are constrained to follow the geometry of spacetime like everything else. Can you give a citation for the articles that you refer to that propose otherwise?
Today CPUs spend a lot of energy trying to extract parallelism out of code designed to be run linearly. The ability to take advantage of parallelism is strongly limited by your ability to find it, rather than the ability of the chip to carry out instructions in parallel.
Well if the chip is emulating a dual processor machine, then you have pushed a lot of that work down to having the OS identify 2 processes that can run in parallel. I would think that this would be a huge win.
This is nice, if you can overcome two concerns:
Duplicate register sets. Each thread will need to work with its own copy of the virtual chip's registers. For processes, instead of threads, each will also need its own page table (though you could just remap them to different sections of the same table). You might be able to emulate dual x86 processors, but something with a larger register set would pose problems.
Different instruction pointers. Each thread is an independent instruction stream, that can wander and jump any which way. You'd need explicit hardware support to be able to emulate this without prohibitive overhead.
Greater memory load. As the processes would be (hopefully) independent, you'd be hitting two completely different sets of pages when accessing memory. This means you'd need a cache twice as large to avoid thrashing. You could get around this by supporting multiple threads, not processes, but this limits the advantage of your proposal.
More hardware complexity. Transmeta's fundamental approach was to reduce hardware complexity and cost. As some additional hardware support is needed (in fact, a fair bit of additional support), emulation of SMP systems is unlikely to be embraced by Transmeta.
This is a fascinating idea, but there are substantial hurdles that would have to be overcome implementing it in practice.
As all of the MPEG formats offer a sliding scale of compression, you should be able to make a video as compact as you like - at the espense of quality.
What is the minimum size per minute of footage for decent visual quality video, for MP4 and for its predecessor?
Possible solution to the problem of people reading the book and not paying for it:
Publish only the first few chapters online.
This will get people interested enough in the book to buy the real version (especially if you have a "click here to order the book" link).
Possible solution to portability:
Release the electronically published chapters in PDF format (prettiest), HTML format (looks decent), or plain text format (ugly and not fully portable).
The drawback to PDF is that PDF files tend to be huge and you'll be stuck paying for the software to write the PDF file (only the reader is free). The good news is that plain text shouldn't be that big, and the reader probably isn't horribly expensive. Also, this will far and away give the prettiest looking output.
The drawback to HTML is that it's a moderate pain to write portably (don't to "save as HTML" from Word; find someone who knows how to write HTML that looks good on all platforms and hire them to do it). It also doesn't look spectacular. However, it looks good enough for most purposes if written well, and can be viewed absolutely anywhere.
The drawback to text is that it looks ugly and that you'll have a hard time supporting all of [Windows, Mac, Unix]. Windows uses CRLF for line breaks, Mac uses CR, and Unix uses LF. On the plus side, text converters for these forms are abundant, and producing the text in the first place is fairly easy ("save as text" usually works adequately).
As another reader points out, your main problem will probably be advertising. Electronic publication is probably best viewed as a supplement to conventional publishing instead of a replacement, unless you're well enough known that "click here to order" traffic through your web site will be enough to sustain you.
Try this search and tell me that doesn't make an impact.
Follow your own links. I see a couple of hundred names on the most populated sites - and fewer than ten names on most links.
The search returned about 280 pages. The Fermis out to maybe ten thousand names, if you're *EXTREMELY* optimistic.
Contrast that to the vast number of albums sold. Looking considerably less significant, isn't it?
Of course, the industry has a few words about the impact on sales as well.
Looks like a typical fluff piece written by a third-party journalist. What of it?
Now, about those figures... look here, here and since some fans even feel they are directly harming metallica they've setup a site to pay lars back.
Your first link states that CD sales are declining near universities. By 7 percent.
What fraction of the population goes to university? Relatively small. What fraction of the population buys CDs? Relatively large.
You become Enlightened.
Or, take the geographic approach. Assume that 5% of all stores selling CDs are near a university. They've had sales drop by 7%. This gives you a whopping 0.35% sales hit. Not looking terribly significant.
Your second link seems to be another third-party journalist article advertising "GoodNoise". There is a link included to an RIAA statement providing "evidence", but this link is dead. This article, lacking a critical component, supports neither of our cases.
Your third link, to the "pay Lars" site, lists a grand total of $399.00 raised at the time of my viewing (during the writing of this message).
If the number of people on this site is supposed to be representative of the number of people boycotting, then the boycott is in a sorry state indeed.
Let's take the more plausible approach, and say that the ratio of payLars-ers to boycotters is equal to the ratio of boycotters to CD buyers. This boosts the number of assumed boycotters very substantially.
It's still not enough. To make the numbers easier, let's assume that Metallica makes the piddlingly small sum of $4 million on sales. Using the ratios above, this gives an estimate of the number of CD buyers per boycotter as being sqrt($4m / $400), or 100. Again, the boycotters are at at most a 1% level.
Try this again with Metallica's real income. Or better yet, stick to counting boycotters themselves so that you have a real number.
Summary: Your figures serve only to underscore how few boycotters there are, compared to the CD-buying masses.
Bands like Metallica and Dr. Dre who have sided with you and the pro-IP fight have met with icy criticism and a mass exodus of their fans.
I seriously doubt that most Metallica/Dr. Dre fans even know what Napster is.
However, I'm sure you wouldn't have posted without evidence. Please show your figures demonstrating that enough fans of these bands are boycotting to make a significant impact on their bottom lines.
t's not the ping pong balls...it's the NUMBER of balls. There are a few hundred there, each made up of how many polygons? Then realize that all of those balls are falling, bouncing and recoiling from the snaps of the mouse-traps (atleast that's what they look like to me) which are ALSO bouncing around in the room. That means a lot of physics (figuring the motion for all the balls and all of the traps).
The physics, at least, takes a negligeable amount of processing power for only a few hundred balls. Collision detection is the only thing that might be a concern, and there are algorithms that can do that efficiently.
That leaves the rendering. Now, you could probably render that scene in real-time with multi-polygon balls (you have a graphics card that can render hundreds of thousands to millions of polygons per second). However, if you're being evil about it, you don't need to. Pre-render pictures of the balls with the "X" at various positions, and render them as 2D sprites (texture maps on 2-triangle quadrilaterals). It would only give visual artifacts on shadows _on_the_balls_, which would be uncommon and difficult to see when present for the demo being used.
They probably rendered this mostly-legitimately, but don't underestimate the power of creative cheating:).
Worse still, unless it is extremely upgradable (i.e. rip the CPU and GPU out and upgrade) it will be obsolete hardware when compared with the standard PC of the time only 6-12 months after release. So what is going on?
The same argument could be made about any other gaming console, and holds. However, the reason why Playstations and Dreamcasts still sell is that they are far, far cheaper than the PC that blows them away.
For anyone who has a good gaming PC, a console can be argued to be redundant (aside from the little matter of getting all of the console _games_ on the PC).
However, Joe Average doesn't have a good gaming PC. Joe Average may not even have a PC at all. However, Joe Parent can more likely afford to shell out for a cheap console than for a full computer when their kid finally convinces them to buy a game system.
For people who don't already have PCs, a console is a good investment. Heck, it may cost less than your PC's next video card.
But what have they actually got to show for all of their venture capital and demonstrations? Not a lot, just a fair of fancy chips that run at about a quarter of the speed of most chips today. Sure, they use less power, but their performance is a joke, and their oh-so-fancy "code morphing" technology looks a bit overrated since the chips are just running Linux.
There are a few things that you seem to be overlooking, here:
Crusoe chips also run Windows. Anyone who feels like dual-booting to Windows on their Crusoe notebook is perfectly free to. They aren't Linux-only. They're emulating the x86 _architecture_, not just a particular operating system.
Power consumption is very, very important for notebooks and palmtops. If you want to play Quake III, use an Athlon or a Pentium III. Power consumption is not an issue for game machines. However, the Crusoe is not _targetted_ at game machines. There is a vast market for extremely low-power, reasonably good chips for PDAs and notebooks, and this chip looks perfectly targetted to that market.
Code morphing gives Transmeta an "in" on Arm and Dragonball territory. Most PDAs do not use x86 chips at the moment. You could build a PDA around a Crusoe with x86 binary translation and bash Linux into shape for PDA work - or you could write a new code morphing later for ARM and/or Dragonball emulation, and use PalmOS et. al. without modification.
Summary: Crusoe does many things well, and is well-targetted for its niche. You seem to be assessing the chip purely from a desktop standpoint, which leads to questionable conclusions.
It turns out that arithmetic operators cost half as many transistors/operator if the bit order goes from lowest-order bit to highest-order bit, as in an Intel chip, than they do if the bit order is highest-order bit first.
Having just finished my final year in Computer Engineering learning chip design, I can say with confidence that this isn't true. The implementation is the same no matter which order the bits are in (just use a mirror image of the layout).
They just want to make music, want to sell it, want to have it spread. We want to hear it, obviously, and share it. Can some genius, someone with the right insight and the right knowledge, right now work a system up that puts all of this together and create a win-win situation?
Someone in a previous article suggested a solution that might help - the band releasing a lower-quality MP3 preemptively.
The MP3 would be high enough quality to sound decent, but low enough that you'd still notice the difference buying the CD. Laziness being what it is, this would stop most people from ripping the CD themselves.
The band wins because people who like several of their songs are likely to buy their CDs to get better quality (free advertising). The people win because they can listen to free songs.
For readers who haven't waded through the article or were confused:
As far as I can tell, the main points Lars makes are as follows:
Metallica should decide how Metallica's works are distributed. Allowing distribution via Napster to go unchallenged removes this control (Metallica hadn't OKd this distribution of their work).
Distribution over the 'net has a much bigger impact than distribution via tape dubbing. Tape dubs degrade and generally aren't spread very widely from the source. Files shared across the 'net are always perfect copies and are distributed very far afield from the original purchaser.
Metallica is investigating 'net distribution options, but Napster won't be it. Metallica is aware of the 'net (now), but wants to retain control over distribution with whatever distribution method is chosen.
The Napster prosecution was an act of the band itself, not their legal department or their record company.
As far as I can tell, these are the main points stated in the interview. Please post addendums if I've missed any.
Remember, one time pads are only genuinely secure if the pad is at least as long as the message, the pad is genuinely random, and the pad is never reused. If any of these is not true, the encryption is breakable.
All of this is true. However, these conditions are easy to meet:
1) Your pad is several gigabytes, or more. How much space will your text messages take up?
2) Diode noise and other schemes can easily produce truly randum pads. Hardware exists for this in many places already; there just isn't much demand right now.
3) You are unlikely to send more than several gigabytes of text in communications in your lifetime. You never need to reuse the pad.
I don't see a problem.
Anyway, even setting aside that a video stream is not necessarily totally random
Perhaps I was not clear enough in my original post - you aren't using a video stream as a key. I was giving "video clips" as an example of _message_data_ that would quickly overflow your pad. The same applies if you want to email your pr0n collection to a friend. OTPs are only really, _really_ practical for text messages.
if pre-agreeing on a giant key to draw from for small messages were a practical scheme, folks would use it now.
Not true. It's practical, and it works - it's just more conventient to use a public key scheme. People will usually go with the most convenient option unless there is a reason not to (for instance, quantum computing breaking public-key systems).
The speed at which quantum computers could break current key based technology is phenomenal. Entirely new encryption methods based on new (quantum?) Algorithms would have to be invented 'cos if a quantum computer got in the hands of a criminal they could breeze through most anything.
One-time pads would still work. As long as you're just trying to send text instead of video clips, you should be able to give your prospective message recipient a few gigabytes of dedicated pad _once_, and be able to communicate securely for many years after.
That having been said, quantum encryption methods have indeed been postulated. The main problem with the ones that I know of is that they have engineering difficulties (sending a single photon to your recipient half a continent away is difficult, for instance).
You can hook icons into a script or use a vague query and have the parser build long scripts for you.
The point of the article is not that vague searches can be performed - it is that an efficient algorithm for performing searches, vague and non-vague, has been found for quantum computers.
Quantum computers and quantum computing algorithms allow you to solve many problems of exponential difficulty in polynomial time. These problems would be intractible on conventional computers, no matter how fast.
Searching large databases can be very, very time consuming. With this algorithm on quantum hardware, it is far, far less so.
Vague query matching is just something that shows up as an added bonus.
Slashdot Mirror
They might, but it would be an uphill battle. The older companies have name recognition and past contracts which only need to be renewed, which count for a lot.
The main selling point of, say, Penguin is that Penguin presumably has more experience building and configuring Linux systems, and so (hopefully) does a better job at it.
In order to become a major player, a company like Penguin would have to:
This can be done, and has a chance of working, but takes a lot of effort and a very large amount of money. There would also be the risk of failing spectacularly, as there are a glut of suppliers on the market.
Out of curiosity, are any retailers offering to ship systems preloaded with *BSD, or even making noise about doing so down the road?
The ultimate goal, recall, is to have systems shipped with your choice of operating system - not just with any given one.
A quick check at CompUSA Onlineshows that full suite price to be $1980.95 and upgrade price is $931.85.
Doing my own search, I find $225 for the stripped down version, and $490 for the full version of Corel Draw alone.
High, but not nearly as high as the figure you quoted.
This is typical of office software, and quite reasonable when compared to the cost of the machine it's going to be used on or of the employee who's going to be using it.
There are almost certianly student versions available for a much lower price, too (around here, student versions are half off or better).
Also, your hyperlink seems to have been munged. Further inspection reveals that they're using some kind of bizzare scratch keys to encode query data, making linking to specific results unreliable.
I'd suspect that the disk subsystem would be the main bottleneck for compiling, followed by the memory subsystem.
When you do a compile, you walk through many, many source files and nested levels of include files (take a look at your dependencies list for an idea of how nasty this is). A fast disk with scatter-gather (coughcoughSCSI) and a large disk cache in memory will work wonders for eliminating that grinding noise that tends to accompany compiles.
Now, the memory subsystem. Compilation with a modern optimizing compiler involves building monstrously huge structures in memory and shuffling them around (I spent much of the last two semesters writing such a compiler). There's enough data here to overflow the cache, so you run into the system memory access speed bottleneck.
100 MHz FSB is Your Friend. 133 is even better, if you have RAM with low enough latency (another poster correctly pointed out that 133 MHz does you no good if wait states eat the extra speed). A larger chip cache will also help, if that's an option.
Lastly, if neither of these seems to be the bottleneck, having a faster CPU or a second CPU may help. Just make sure that the CPU is the bottleneck before forking out the money for a faster one. You can also safely ignore floating-point performance - almost all of the computation performed by a compiler is integer math (building bit vectors to calculate dominators and so forth with).
As far as benchmarks go, the disk- and memory-limited ones should still be useful to you.
Err... did you bother to read the link he gave to magram? No? Well, apparently, magram is faster than DRAM, non-volotile, and not much more expensive that hard drive space... or something...
Yes, I've read the link to magram.
This is a non-technical article describing what they *hope* magram will accomplish when the technology matures.
It is *not* a description of what they actually have working.
There are many, many interesting memory technologies in this stage of development. They only become relevant to debates when engineering prototypes with decent density and speed are produced.
We've become sloppy and trained to the notion that memory should be divided into segments varying by speed, size, volatility, and cost. We all spend months in college or in the field learning about the subtle differences between L1 caching, L2 caching, main memory, hard drive memory, ROM, and the trillion variations of RAM. We don't see the forest for the trees: this model of data management is the single most crucial hindrance to the advancement of computer science in our entire industry.
The problem is that this multilevel approach to memory exists for a very good reason - all serious candidates for implementing RAM have an inverse relationship between speed and cost.
Sure, you could put a gig of SRAM in your machine, but you'd have to spend a pretty penny to do it. This isn't because of some evil manufacturer jacking up prices - it's because SRAM is many times less dense than DRAM, and you're paying for the silicon area that you use.
There's also an inverse relationship between size and speed. This is why a chip's L2 SRAM is slower than its L1 SRAM. The larger RAM array has longer traces and a number of other features that increase settling time and in other ways make the response time longer.
Similarly, even if you did decide to fill your motherboard with SRAM or something comparably fast, you wouldn't be able to access it at full speed. Motherboard traces are long, which increases signal propagation latency, and noise injection from the environment is worse, which decreases the rate at which you can transfer data even if you can tolerate high latency.
These signalling problems aren't going to go away, and won't change even with different RAM types. At best, you can learn to live with them (RAMBus uses a clever scheme to reduce noise, allowing them a faster throughput while keeping the lousy latency of external memory).
As for using disk as memory or vice versa - this just isn't practical. Disk is a thousand times slower than memory and a hundred times cheaper. Take out most of your RAM and watch your machine swap to see what using this as system memory is like. Check out the price tags for solid state drives to see what the cost of using memory as a drive is like (or just ask your local store how much 30 gigs of DIMMs would cost).
In conclusion, I think that your suggestion is not compatible with existing or proposed storage technologies.
Some time ago, when a friend of mine had a cell phone and I didn't, I'd send him text messages via a web page helpfully provided by Fido (the company selling us the service).
This was very useful, but is trivially easy to spam via scripts. My friend even wrote such a script, to forward email from his account to his phone (before purchasing phone email service).
It would only take one or two knowledgeable people saying "hey, that's neat!" to do that here in Toronto, and I'm sure Fido isn't the only company set up this way.
I sorted out my confusion on this FAQ web site.
/physics/relativity.html
I made it a few pages down the FAQ, and I'm afraid I have to say that it's one of the worst that I've ever seen. The author's argument seems to amount to "Well, you can describe the universe mathematically using four dimensions, but because I only see three of them this is purely bunk.".
At this point, I stopped reading.
In point of fact, if you want to see the effects of space being four-dimensional, you need only look at two reference that are moving quickly with respect to each other. The time and space directions measured by observers in each frame are different - this is the Lorentz transform (if I remember the term correctly). Space in one frame corresponds to a skewed space-and-time axis in another frame.
This has been verified experimentally by very careful measurements of atomic clocks moving at different speeds with respect to each other. A more dramatic illustration is measuring the decay times of unstable particles moving at different speeds. As they approach the speed of light, the lifetime as measured by the observer gets longer. This is called "time dilation", and is one of many effects caused by the "time" and "space" axes not being the same for observers moving with respect to each other.
You can find a very good FAQ on relativity here:
http://math.ucr.edu/home/baez
This links to a FAQ on general physics, and many other FAQs that may be of interest.
There's still some argument over whether all particles are influenced by gravity; I seem to remember recently that there is talk of the gravitational lens effect (ie light being affected by gravity) being disproved.
That would be quite a trick. If I remember correctly, one of the measurements used to support relativity was of gravitational lensing of starlight by the sun (mask out the disc and take careful measurements of the apparent positions of stars next to it).
Gravitational lensing studies are also being performed to measure the mass of distant galaxies (by looking at multiple images of even more distant galaxies). If gravitational lensing didn't exist, it would have been noticed in a big way by these studies.
As far as anyone's been able to determine, photons are constrained to follow the geometry of spacetime like everything else. Can you give a citation for the articles that you refer to that propose otherwise?
Well if the chip is emulating a dual processor machine, then you have pushed a lot of that work down to having the OS identify 2 processes that can run in parallel. I would think that this would be a huge win.
This is nice, if you can overcome two concerns:
Each thread will need to work with its own copy of the virtual chip's registers. For processes, instead of threads, each will also need its own page table (though you could just remap them to different sections of the same table). You might be able to emulate dual x86 processors, but something with a larger register set would pose problems.
Each thread is an independent instruction stream, that can wander and jump any which way. You'd need explicit hardware support to be able to emulate this without prohibitive overhead.
As the processes would be (hopefully) independent, you'd be hitting two completely different sets of pages when accessing memory. This means you'd need a cache twice as large to avoid thrashing. You could get around this by supporting multiple threads, not processes, but this limits the advantage of your proposal.
Transmeta's fundamental approach was to reduce hardware complexity and cost. As some additional hardware support is needed (in fact, a fair bit of additional support), emulation of SMP systems is unlikely to be embraced by Transmeta.
This is a fascinating idea, but there are substantial hurdles that would have to be overcome implementing it in practice.
As all of the MPEG formats offer a sliding scale of compression, you should be able to make a video as compact as you like - at the espense of quality.
What is the minimum size per minute of footage for decent visual quality video, for MP4 and for its predecessor?
Hard? Just start with one file, and use sed to convert whatever linebreaks the file has into the correct format.
:).
I'm assuming that the author in question is using Windows. For my own use, I hacked together a Perl script ("tconv - "). But I'm using *nix.
Or if making the files available to download, make seperate links "For Mac", "For Windows", "For Linux".
Yes, this isn't exactly optimal, but what else can you do when platforms aren't compatible?
Use HTML
Using separate links would be my solution too for text.
This will get people interested enough in the book to buy the real version (especially if you have a "click here to order the book" link).
Possible solution to portability:
The drawback to PDF is that PDF files tend to be huge and you'll be stuck paying for the software to write the PDF file (only the reader is free). The good news is that plain text shouldn't be that big, and the reader probably isn't horribly expensive. Also, this will far and away give the prettiest looking output.
The drawback to HTML is that it's a moderate pain to write portably (don't to "save as HTML" from Word; find someone who knows how to write HTML that looks good on all platforms and hire them to do it). It also doesn't look spectacular. However, it looks good enough for most purposes if written well, and can be viewed absolutely anywhere.
The drawback to text is that it looks ugly and that you'll have a hard time supporting all of [Windows, Mac, Unix]. Windows uses CRLF for line breaks, Mac uses CR, and Unix uses LF. On the plus side, text converters for these forms are abundant, and producing the text in the first place is fairly easy ("save as text" usually works adequately).
As another reader points out, your main problem will probably be advertising. Electronic publication is probably best viewed as a supplement to conventional publishing instead of a replacement, unless you're well enough known that "click here to order" traffic through your web site will be enough to sustain you.
Try this search and tell me that doesn't make an impact.
Follow your own links. I see a couple of hundred names on the most populated sites - and fewer than ten names on most links.
The search returned about 280 pages. The Fermis out to maybe ten thousand names, if you're *EXTREMELY* optimistic.
Contrast that to the vast number of albums sold. Looking considerably less significant, isn't it?
Of course, the industry has a few words about the impact on sales as well.
Looks like a typical fluff piece written by a third-party journalist. What of it?
Now, about those figures... look here, here and since some fans even feel they are directly harming metallica they've setup a site to pay lars back.
Your first link states that CD sales are declining near universities. By 7 percent.
What fraction of the population goes to university? Relatively small.
What fraction of the population buys CDs? Relatively large.
You become Enlightened.
Or, take the geographic approach. Assume that 5% of all stores selling CDs are near a university. They've had sales drop by 7%. This gives you a whopping 0.35% sales hit. Not looking terribly significant.
Your second link seems to be another third-party journalist article advertising "GoodNoise". There is a link included to an RIAA statement providing "evidence", but this link is dead. This article, lacking a critical component, supports neither of our cases.
Your third link, to the "pay Lars" site, lists a grand total of $399.00 raised at the time of my viewing (during the writing of this message).
If the number of people on this site is supposed to be representative of the number of people boycotting, then the boycott is in a sorry state indeed.
Let's take the more plausible approach, and say that the ratio of payLars-ers to boycotters is equal to the ratio of boycotters to CD buyers. This boosts the number of assumed boycotters very substantially.
It's still not enough. To make the numbers easier, let's assume that Metallica makes the piddlingly small sum of $4 million on sales. Using the ratios above, this gives an estimate of the number of CD buyers per boycotter as being sqrt($4m / $400), or 100. Again, the boycotters are at at most a 1% level.
Try this again with Metallica's real income. Or better yet, stick to counting boycotters themselves so that you have a real number.
Summary: Your figures serve only to underscore how few boycotters there are, compared to the CD-buying masses.
I eagerly await your rebuttal.
Bands like Metallica and Dr. Dre who have sided with you and the pro-IP fight have met with icy criticism and a mass exodus of their fans.
I seriously doubt that most Metallica/Dr. Dre fans even know what Napster is.
However, I'm sure you wouldn't have posted without evidence. Please show your figures demonstrating that enough fans of these bands are boycotting to make a significant impact on their bottom lines.
t's not the ping pong balls...it's the NUMBER of balls. There are a few hundred there, each made up of how many polygons? Then realize that all of those balls are falling, bouncing and recoiling from the snaps of the mouse-traps (atleast that's what they look like to me) which are ALSO bouncing around in the room. That means a lot of physics (figuring the motion for all the balls and all of the traps).
:).
The physics, at least, takes a negligeable amount of processing power for only a few hundred balls. Collision detection is the only thing that might be a concern, and there are algorithms that can do that efficiently.
That leaves the rendering. Now, you could probably render that scene in real-time with multi-polygon balls (you have a graphics card that can render hundreds of thousands to millions of polygons per second). However, if you're being evil about it, you don't need to. Pre-render pictures of the balls with the "X" at various positions, and render them as 2D sprites (texture maps on 2-triangle quadrilaterals). It would only give visual artifacts on shadows _on_the_balls_, which would be uncommon and difficult to see when present for the demo being used.
They probably rendered this mostly-legitimately, but don't underestimate the power of creative cheating
Worse still, unless it is extremely upgradable (i.e. rip the CPU and GPU out and upgrade) it will be obsolete hardware when compared with the standard PC of the time only 6-12 months after release. So what is going on?
The same argument could be made about any other gaming console, and holds. However, the reason why Playstations and Dreamcasts still sell is that they are far, far cheaper than the PC that blows them away.
For anyone who has a good gaming PC, a console can be argued to be redundant (aside from the little matter of getting all of the console _games_ on the PC).
However, Joe Average doesn't have a good gaming PC. Joe Average may not even have a PC at all. However, Joe Parent can more likely afford to shell out for a cheap console than for a full computer when their kid finally convinces them to buy a game system.
For people who don't already have PCs, a console is a good investment. Heck, it may cost less than your PC's next video card.
There are a few things that you seem to be overlooking, here:
Anyone who feels like dual-booting to Windows on their Crusoe notebook is perfectly free to. They aren't Linux-only. They're emulating the x86 _architecture_, not just a particular operating system.
If you want to play Quake III, use an Athlon or a Pentium III. Power consumption is not an issue for game machines. However, the Crusoe is not _targetted_ at game machines. There is a vast market for extremely low-power, reasonably good chips for PDAs and notebooks, and this chip looks perfectly targetted to that market.
Most PDAs do not use x86 chips at the moment. You could build a PDA around a Crusoe with x86 binary translation and bash Linux into shape for PDA work - or you could write a new code morphing later for ARM and/or Dragonball emulation, and use PalmOS et. al. without modification.
Summary: Crusoe does many things well, and is well-targetted for its niche. You seem to be assessing the chip purely from a desktop standpoint, which leads to questionable conclusions.
Here is an aritcle with more details on the work being done:
http://newsbytes.com/pubNews/00/149716. html
It turns out that arithmetic operators cost half as many transistors/operator if the bit order goes from lowest-order bit to highest-order bit, as in an Intel chip, than they do if the bit order is highest-order bit first.
Having just finished my final year in Computer Engineering learning chip design, I can say with confidence that this isn't true. The implementation is the same no matter which order the bits are in (just use a mirror image of the layout).
They just want to make music, want to sell it, want to have it spread. We want to hear it, obviously, and share it. Can some genius, someone with the right insight and the right knowledge, right now work a system up that puts all of this together and create a win-win situation?
Someone in a previous article suggested a solution that might help - the band releasing a lower-quality MP3 preemptively.
The MP3 would be high enough quality to sound decent, but low enough that you'd still notice the difference buying the CD. Laziness being what it is, this would stop most people from ripping the CD themselves.
The band wins because people who like several of their songs are likely to buy their CDs to get better quality (free advertising). The people win because they can listen to free songs.
As far as I can tell, the main points Lars makes are as follows:
Allowing distribution via Napster to go unchallenged removes this control (Metallica hadn't OKd this distribution of their work).
Tape dubs degrade and generally aren't spread very widely from the source. Files shared across the 'net are always perfect copies and are distributed very far afield from the original purchaser.
Metallica is aware of the 'net (now), but wants to retain control over distribution with whatever distribution method is chosen.
As far as I can tell, these are the main points stated in the interview. Please post addendums if I've missed any.
Remember, one time pads are only genuinely secure if the pad is at least as long as the message, the pad is genuinely random, and the pad is never reused. If any of these is not true, the encryption is breakable.
All of this is true. However, these conditions are easy to meet:
1) Your pad is several gigabytes, or more. How much space will your text messages take up?
2) Diode noise and other schemes can easily produce truly randum pads. Hardware exists for this in many places already; there just isn't much demand right now.
3) You are unlikely to send more than several gigabytes of text in communications in your lifetime. You never need to reuse the pad.
I don't see a problem.
Anyway, even setting aside that a video stream is not necessarily totally random
Perhaps I was not clear enough in my original post - you aren't using a video stream as a key. I was giving "video clips" as an example of _message_data_ that would quickly overflow your pad. The same applies if you want to email your pr0n collection to a friend. OTPs are only really, _really_ practical for text messages.
if pre-agreeing on a giant key to draw from for small messages were a practical scheme, folks would use it now.
Not true. It's practical, and it works - it's just more conventient to use a public key scheme. People will usually go with the most convenient option unless there is a reason not to (for instance, quantum computing breaking public-key systems).
The speed at which quantum computers could break current key based technology is phenomenal. Entirely new encryption methods based on new (quantum?) Algorithms would have to be invented 'cos if a quantum computer got in the hands of a criminal they could breeze through most anything.
One-time pads would still work. As long as you're just trying to send text instead of video clips, you should be able to give your prospective message recipient a few gigabytes of dedicated pad _once_, and be able to communicate securely for many years after.
That having been said, quantum encryption methods have indeed been postulated. The main problem with the ones that I know of is that they have engineering difficulties (sending a single photon to your recipient half a continent away is difficult, for instance).
You can hook icons into a script or use a vague query and have the parser build long scripts for you.
The point of the article is not that vague searches can be performed - it is that an efficient algorithm for performing searches, vague and non-vague, has been found for quantum computers.
Quantum computers and quantum computing algorithms allow you to solve many problems of exponential difficulty in polynomial time. These problems would be intractible on conventional computers, no matter how fast.
Searching large databases can be very, very time consuming. With this algorithm on quantum hardware, it is far, far less so.
Vague query matching is just something that shows up as an added bonus.