While the device in question must of course prove itself to not be bogus, there is some precedent for using a "special arrangement of magnets" to increase efficiency. LLNL had a demonstration maglev train that would glide quite nicely for a long distance after you pushed it to a certain speed.
The following posts provide some additional information:
http://www.llnl.gov/str/Post.html
http://www.alienbaby.com/levitron.html
I could see this principle applied to a package like an electric motor. Indeed, the article describes that the applied energy is used to "kick" the rotor (sorry, my knowledge of such motors is limited, so please pardon my misuse of terminology) into a position in which the array can then finish guiding it around.
In essence, there's nothing really fishy about it. Its just being really, really clever in harnessing the underlying principles. Again, though, the design remains to be proven, both as workable and truly economical.
A single hard disk magnet can usually lift at least 10 pounds, maybe more.
In my experience, that's a bit much. I could see them lifting a couple pounds of metalic stuff, but not really much more. They are hella fun to play with though;) Plus, thy make awesome fridge magnets.
Not to mention every modern hard drive has a nice, big pair of rare-earth magnets driving the armature for the heads. They're placement is barely a centimeter away from the disks in the older (1.2 GB) Western Digital and Seagate drives I pulled apart a couple years ago.
Makes for a nice source of rare earth magnets, which are super strong at short distances.
Many times, characters weren't either "bad guys" or "good guys," but just had their own way of getting through life.
This was really attractive to me in watching the show. I knew I *really, really* liked Farscape when several main characters mutilated another character in order to further their own, selfish goals.
Its not that the characters are good or evil, its that they're human and have their own reasons and ambitions which they follow as the plot moves along. The characters and their relationships to one another were up front and central to the plot. Trust between characters was a recurring theme.
Setting aside the posibilty that this is an April Fool's joke, (Although it does say March 31st on the story..) perhaps advertising is exactly what they're after.
Not to mention that if you're logged in they can use your specific information and analysis of your email to target regular Google ads to you. Think about it, the worlds largest, most respected search engine, with one of the nicest ad systems around is able to accurately target ads to its customers. Not to mention they can store information about your search habits and interests at all times.
This has the potential for massive abuse, but I really, really want to continue believing that Goolge is a truly ethical company. So far they've done a fine job of catering to pay customers (advertisers, et al.) and regular Joes/Janes.
They should closely scrutinize the downloading habits, then create an album based on the popularity of certain tracks.
They don't see this as a tool, only as a threat. They're idiots.
Actually, the real fun with the RIAA and major labels is that they already do such things. Please view the Wired article:
BigChampagne is Watching You.
I say this is fun because the RIAA talks out both sides of its mouth: it wants to limit major expansion of free P2P downloads (control the download market) and simultaneously use the data from such spontaneous sources to make smart investments on marketing.
Of course, when they say "CD sales" have gone down, I'm not so sure they mean all CD-based formats (singles, albums, collections, etc.) or just some sub-categories, like CD sigles. I can believe CD singles have been decimated by P2P filesharing, but I'm more reluctant to agree to a rapid, major decline in album sales without proper evidence. In other words, I don't believe what the RIAA claims is exactly what is happening, merely what they want you to think is happening.
When building beowulf-style clusters, RedHat kickstart has been one of the quintessential installation tools. Its not the only way do things, but its the one I find most useful in my particular setups.
The lack of a kickstart-like installation automation tool for Debian-based systems has kept my clusters RedHat-based exclusively. Does the new installation tool help with this? If not, why not? I know its been requested many, many times. This functionality is entirely too useful to really ignore.
A use for it that even run-of-the-mill boxes might like is that if your box needs a reinstall, simply reinstall using the kickstart script provided after the original install is complete. The machine will then reinstall in exactly the same manner as before, though you may or may not have to apply updates.
This reminds me of good-ol' Win95. You could install it from CD-Rom, but after its first reboot into Windows to finish the install, the CD-Rom driver had not been installed and you couldn't complete the immediate installation.
I ended up pulling the generic cdrom driver off of a Win98 boot disk and using that disk to install Win95. It was years later that I realized the Win95 install would finish and then install the cdrom driver, which could then be used to finish the install from cdrom.
Of course, this was years ago, but it was a major pain in the arse for me at the time.
Something that often crops up in the C/C++ versus Fortran discussions is pointer-aliasing. The inability of C/C++ compilers to really optimize around the problem of pointer-aliasing is a non-trivial problem and impedes the performance of those codes.
Now, I don't program in Java, but I understand that Java pointers are "smarter" than C/C++ pointers and that the JVM may be smarter at determining aliasing and be more aggressive about optimizing. If (again, I don't know, but I'd be interested in knowing) this is true, then it could be relatively easy to imagine situations in which a JVM (with appropriately aggressive optimizations) can execute code much faster than C/C++ compiled code.
Which is also a way to turn waste (biomass and otherwise) into fuels. This method can turn plastics, biomasses, tires, and many other common waste products into gasses, oil, water, carbon, and minerals. All of these things are commercially useful.
This is not a hoax, as they do have a plant operating for ConAgra's Butterball turkey plant in Carthage, MO. It produces oil (convertable to other petro-based fules and plastics) from turkey wastes!
Changing World Technologies
Now, sit back and imagine that this type of facility could be used to recycle are landfills into commercially useful products.
Its also useful to purify and reduce other waste products from current petroleum and coal supplies (think sulfur and mercury).
>other is sharing of the bus bandwidth between two >CPUs,
While your points are quite valid, I do wish to correct this one error. Intel x86 chips use a shared-bus architecture for their multi-processor offerings. However, AMD Athlon MP SMP offerings use a point-to-point architecture in which each processor has a dedicated bus to the memory. So, they do not share bus bandwidth.
Hans Reiser has a different take: commit atomically on flush(). This does the commit at about the most performance-friendly time and in a manner that keeps the filesystem state nice and clean. This is how Reiser v4 works.
However, one of my former professors does research on removing nasty birthmark splotches (think Mikail Gorbechov) using lasers. Before any burning starts they freeze the skin above the blood vessels to be cauterized. I understand this can be done to a pretty good precision. It could be possible to combine such techniques with a dye and create laser-imprinted tatoos or just "burned-in" images. I've seen people who purposefully brand themselves, so such a market may exist, though it may be somewhat expensive.
What's wrong with Dover texts for many things? Has calculus and algebra, classical physics, "modern" physics, or classical literature changed in the last 70-90 years?
Could MIT's OpenCourseware (please google!) be used as a starting point? The courses at which I've looked use known standard books as their references for assignments, but often have lecture notes and such posted.
My experiences are somewhat different, as may be my usage pattern. First, I've noticed on sites with Flash animations or lots of graphics, Firbird tends to be *slower* than Mozilla. This is slower in gui responsiveness (hitting the close tab button or switching tabs or opening a menu) and page rendering. Next, startup time: I have none. My usage is to keep the browser open for as long as possible as I'm always using it (webmail, slashdot, google, etc. all day long). I never close the browser unless I absolutely have to. I figure any amount of load time on my applications is acceptable as long as they stay open, noncrashed, for days or weeks at a time. There simply is no load time, and Mozilla and Firebird are both good for this. Oh, BTW, OpenOffice is also like this for me. I used Calc to grade papers and commonly kept it open for weeks at a time. Finally, the thing that's more important for these usage patterns is memory, physical and virtual as the penalty for swapping in and out an unused process is sometimes noticeable enough to be equivalent to "load time".
A couple of points about a language I've learned to really love and you do not seem to understand.
Languages like C (or C++, Java, Perl, Python, Fortran, etc.) are inherently serial in nature.
Yes, Fortran started as an inherently serial language but many, many of its modern revisions is to increase the areas in which a compiler can safely apply parallel optimizations.
For example: your "incrementing" code snippet:
(incrementmyArray + 1, myArray, count);
is written thus in modern Fortran using two arrays of dimensions 1 to n (arbitrarily defined indices):
REAL:: array1(1:n), array2(1:n)
array2=array1 + 1
And if you only wish slices of the array:
array2(x:y) = array1(a:b) + 1
Anywho, just a few sentences an examples to show you that not all of those languages have to be or continue to remain "primarily serial".
I'm not sure about over charging, but they used to have problems with filing to get a full charge if used before they were fully charged. This has been fixed in the last couple years, though.
I can second this. I bought a used iPod (original series) from a friend for real cheap because he thought the battery was dead. Turns out thare is a defect in these units in which a capacitor in the charging circuit will not release its charge, preventing the battery from either a) charging to capacity, or b) charging but being detected as empty.
Solution: unplug the battery and wait 5 minutes before plugging it back in.
My friend only saw ~2 hours runtime with the unit. I've used it for ~8 hours at work and traveling home with charge still left. So, the problem and its solution are real enough in my unit.
Apple has deliberately turned off processor features on the other platforms that would have led to their 'fastest in the world' claim being untrue. That's the point of the article. Cross-platform benchmarking IS hard, but deliberately crippling what you benchmark against in order to look better makes it seem that your software/hardware/whatever just isn't as good as what you're comparing it to...
A surprisingly reverse example of this is AMD and the Opteron benchmarks. When benchmarking against Intel processors, AMD often uses the 32-bit Intel compiler. This cuts off the reportedly large benefit the Opteron gains from the extra registers and such in 64-bit mode. I'm really waiting for someone to use a compiler that supports Intel ia64, ia32, and amd64 on the same platforms. The Portland Group have their pgi compiler, with Opteron support in 5.0 Beta. I think those tests may be a bit more of a fair playing ground as they also will support all of the vector unit extensions as well as 64-bit mode on the Opteron.
IIRC, it was reported for several years in a row (I remember 1999, 2000, and I sort of remember 2001) that CD album sales skyrocketed, hitting all-time highs each year. However, in his response the RIAA spokesman says sales declined in some of these years. How can both be true? Can someone provide references to actual sales data?
Now, something that raises my suspicion is the term "unit sales" are declining. Units of what? My suspicion is that CD albums are a steady and healthy source of revenue, but things such a cassette tapes and CD singles are a dying market. The P2P services really trump these markets (getting small amounts of music cheap). So, again, where is the sales data, preferably in some form we can infer which markets are actually declining.
I have no doubt that P2P services and the availabilty of cheap music online is putting a dent in their revenue, but I'm very skeptical of where that dent is being placed. It might also be interesting to correlate their sales data with national figures on how much we spend on "entertainment" and "luxury" items over the same years. I'm not sure how to correlate these things, but it would be very interesting to confirm and infer some more useful results than "unit sales are declining".
I'm sorry if I missed someting, but what is wrong with a small, simple digital audio recorder and your notepad. You can pick one up at Wal-Mart, Radio Shack (i.e. anywhere) for under $100 and many accept external memory. They all can record several hours worth of audio and you can than use your notepad to take notes. I'd say this is an order of magnitude cheaper than a tablet computer for the same functionality. In other words, unless the digital ink application is just too compelling for you, nothing stops you from doing what you want by this afternoon for under $100.
Last night I just discovered fansubbed versions of _Last Exile_. Awesome visual style harkening to the WWI era of aircraft mixed with Star Wars pod racing. Just plain *Frikin' A*. Get it from Anime-Kraze as a Bit Torrent download.
Slashdot Mirror
While the device in question must of course prove itself to not be bogus, there is some precedent for using a "special arrangement of magnets" to increase efficiency. LLNL had a demonstration maglev train that would glide quite nicely for a long distance after you pushed it to a certain speed.
The following posts provide some additional information:
http://www.llnl.gov/str/Post.html
http://www.alienbaby.com/levitron.html
I could see this principle applied to a package like an electric motor. Indeed, the article describes that the applied energy is used to "kick" the rotor (sorry, my knowledge of such motors is limited, so please pardon my misuse of terminology) into a position in which the array can then finish guiding it around.
In essence, there's nothing really fishy about it. Its just being really, really clever in harnessing the underlying principles. Again, though, the design remains to be proven, both as workable and truly economical.
A single hard disk magnet can usually lift at least 10 pounds, maybe more.
;) Plus, thy make awesome fridge magnets.
In my experience, that's a bit much. I could see them lifting a couple pounds of metalic stuff, but not really much more. They are hella fun to play with though
Makes for a nice source of rare earth magnets, which are super strong at short distances.
Hello? Oh, hi mom. Yeah, I can help you install a program on your computer. What do you want to install? Oh, cool. Have you downloaded it?
Okay, hang on for a moment.
$ ssh moms.computer.net
It'll be done in just a sec, Mom!
This was really attractive to me in watching the show. I knew I *really, really* liked Farscape when several main characters mutilated another character in order to further their own, selfish goals.
Its not that the characters are good or evil, its that they're human and have their own reasons and ambitions which they follow as the plot moves along. The characters and their relationships to one another were up front and central to the plot. Trust between characters was a recurring theme.
Not to mention that if you're logged in they can use your specific information and analysis of your email to target regular Google ads to you. Think about it, the worlds largest, most respected search engine, with one of the nicest ad systems around is able to accurately target ads to its customers. Not to mention they can store information about your search habits and interests at all times.
This has the potential for massive abuse, but I really, really want to continue believing that Goolge is a truly ethical company. So far they've done a fine job of catering to pay customers (advertisers, et al.) and regular Joes/Janes.
Very tenuous. It doesn't take into account that I *may* purchase the album and others like it after that download.
They don't see this as a tool, only as a threat. They're idiots.
Actually, the real fun with the RIAA and major labels is that they already do such things. Please view the Wired article:
BigChampagne is Watching You.
I say this is fun because the RIAA talks out both sides of its mouth: it wants to limit major expansion of free P2P downloads (control the download market) and simultaneously use the data from such spontaneous sources to make smart investments on marketing.
Of course, when they say "CD sales" have gone down, I'm not so sure they mean all CD-based formats (singles, albums, collections, etc.) or just some sub-categories, like CD sigles. I can believe CD singles have been decimated by P2P filesharing, but I'm more reluctant to agree to a rapid, major decline in album sales without proper evidence. In other words, I don't believe what the RIAA claims is exactly what is happening, merely what they want you to think is happening.
When building beowulf-style clusters, RedHat kickstart has been one of the quintessential installation tools. Its not the only way do things, but its the one I find most useful in my particular setups.
The lack of a kickstart-like installation automation tool for Debian-based systems has kept my clusters RedHat-based exclusively. Does the new installation tool help with this? If not, why not? I know its been requested many, many times. This functionality is entirely too useful to really ignore.
A use for it that even run-of-the-mill boxes might like is that if your box needs a reinstall, simply reinstall using the kickstart script provided after the original install is complete. The machine will then reinstall in exactly the same manner as before, though you may or may not have to apply updates.
This reminds me of good-ol' Win95. You could install it from CD-Rom, but after its first reboot into Windows to finish the install, the CD-Rom driver had not been installed and you couldn't complete the immediate installation.
I ended up pulling the generic cdrom driver off of a Win98 boot disk and using that disk to install Win95. It was years later that I realized the Win95 install would finish and then install the cdrom driver, which could then be used to finish the install from cdrom.
Of course, this was years ago, but it was a major pain in the arse for me at the time.
Something that often crops up in the C/C++ versus Fortran discussions is pointer-aliasing. The inability of C/C++ compilers to really optimize around the problem of pointer-aliasing is a non-trivial problem and impedes the performance of those codes.
Now, I don't program in Java, but I understand that Java pointers are "smarter" than C/C++ pointers and that the JVM may be smarter at determining aliasing and be more aggressive about optimizing. If (again, I don't know, but I'd be interested in knowing) this is true, then it could be relatively easy to imagine situations in which a JVM (with appropriately aggressive optimizations) can execute code much faster than C/C++ compiled code.
This is not a hoax, as they do have a plant operating for ConAgra's Butterball turkey plant in Carthage, MO. It produces oil (convertable to other petro-based fules and plastics) from turkey wastes!
Changing World Technologies
Now, sit back and imagine that this type of facility could be used to recycle are landfills into commercially useful products.
Its also useful to purify and reduce other waste products from current petroleum and coal supplies (think sulfur and mercury).
>other is sharing of the bus bandwidth between two >CPUs,
While your points are quite valid, I do wish to correct this one error. Intel x86 chips use a shared-bus architecture for their multi-processor offerings. However, AMD Athlon MP SMP offerings use a point-to-point architecture in which each processor has a dedicated bus to the memory. So, they do not share bus bandwidth.
Hans Reiser has a different take: commit atomically on flush(). This does the commit at about the most performance-friendly time and in a manner that keeps the filesystem state nice and clean. This is how Reiser v4 works.
However, one of my former professors does research on removing nasty birthmark splotches (think Mikail Gorbechov) using lasers. Before any burning starts they freeze the skin above the blood vessels to be cauterized. I understand this can be done to a pretty good precision. It could be possible to combine such techniques with a dye and create laser-imprinted tatoos or just "burned-in" images. I've seen people who purposefully brand themselves, so such a market may exist, though it may be somewhat expensive.
What's wrong with Dover texts for many things? Has calculus and algebra, classical physics, "modern" physics, or classical literature changed in the last 70-90 years?
Could MIT's OpenCourseware (please google!) be used as a starting point? The courses at which I've looked use known standard books as their references for assignments, but often have lecture notes and such posted.
My experiences are somewhat different, as may be my usage pattern. First, I've noticed on sites with Flash animations or lots of graphics, Firbird tends to be *slower* than Mozilla. This is slower in gui responsiveness (hitting the close tab button or switching tabs or opening a menu) and page rendering. Next, startup time: I have none. My usage is to keep the browser open for as long as possible as I'm always using it (webmail, slashdot, google, etc. all day long). I never close the browser unless I absolutely have to. I figure any amount of load time on my applications is acceptable as long as they stay open, noncrashed, for days or weeks at a time. There simply is no load time, and Mozilla and Firebird are both good for this. Oh, BTW, OpenOffice is also like this for me. I used Calc to grade papers and commonly kept it open for weeks at a time. Finally, the thing that's more important for these usage patterns is memory, physical and virtual as the penalty for swapping in and out an unused process is sometimes noticeable enough to be equivalent to "load time".
Languages like C (or C++, Java, Perl, Python, Fortran, etc.) are inherently serial in nature.
Yes, Fortran started as an inherently serial language but many, many of its modern revisions is to increase the areas in which a compiler can safely apply parallel optimizations.
For example: your "incrementing" code snippet:
(incrementmyArray + 1, myArray, count);
is written thus in modern Fortran using two arrays of dimensions 1 to n (arbitrarily defined indices):
REAL
array2=array1 + 1
And if you only wish slices of the array:
array2(x:y) = array1(a:b) + 1
Anywho, just a few sentences an examples to show you that not all of those languages have to be or continue to remain "primarily serial".
From my Intro. Literature class I remember Lady Murasaki's _The Tale of Genji_ to be the first novel, written in Japan around 1000 A.D.
I can second this. I bought a used iPod (original series) from a friend for real cheap because he thought the battery was dead. Turns out thare is a defect in these units in which a capacitor in the charging circuit will not release its charge, preventing the battery from either a) charging to capacity, or b) charging but being detected as empty.
Solution: unplug the battery and wait 5 minutes before plugging it back in.
My friend only saw ~2 hours runtime with the unit. I've used it for ~8 hours at work and traveling home with charge still left. So, the problem and its solution are real enough in my unit.
Apple has deliberately turned off processor features on the other platforms that would have led to their 'fastest in the world' claim being untrue. That's the point of the article. Cross-platform benchmarking IS hard, but deliberately crippling what you benchmark against in order to look better makes it seem that your software/hardware/whatever just isn't as good as what you're comparing it to...
A surprisingly reverse example of this is AMD and the Opteron benchmarks. When benchmarking against Intel processors, AMD often uses the 32-bit Intel compiler. This cuts off the reportedly large benefit the Opteron gains from the extra registers and such in 64-bit mode. I'm really waiting for someone to use a compiler that supports Intel ia64, ia32, and amd64 on the same platforms. The Portland Group have their pgi compiler, with Opteron support in 5.0 Beta. I think those tests may be a bit more of a fair playing ground as they also will support all of the vector unit extensions as well as 64-bit mode on the Opteron.
IIRC, it was reported for several years in a row (I remember 1999, 2000, and I sort of remember 2001) that CD album sales skyrocketed, hitting all-time highs each year. However, in his response the RIAA spokesman says sales declined in some of these years. How can both be true? Can someone provide references to actual sales data?
Now, something that raises my suspicion is the term "unit sales" are declining. Units of what? My suspicion is that CD albums are a steady and healthy source of revenue, but things such a cassette tapes and CD singles are a dying market. The P2P services really trump these markets (getting small amounts of music cheap). So, again, where is the sales data, preferably in some form we can infer which markets are actually declining.
I have no doubt that P2P services and the availabilty of cheap music online is putting a dent in their revenue, but I'm very skeptical of where that dent is being placed. It might also be interesting to correlate their sales data with national figures on how much we spend on "entertainment" and "luxury" items over the same years. I'm not sure how to correlate these things, but it would be very interesting to confirm and infer some more useful results than "unit sales are declining".
I'm sorry if I missed someting, but what is wrong with a small, simple digital audio recorder and your notepad. You can pick one up at Wal-Mart, Radio Shack (i.e. anywhere) for under $100 and many accept external memory. They all can record several hours worth of audio and you can than use your notepad to take notes. I'd say this is an order of magnitude cheaper than a tablet computer for the same functionality. In other words, unless the digital ink application is just too compelling for you, nothing stops you from doing what you want by this afternoon for under $100.
Hey, thanks for that tip. I'll put it to good use from now on.
Last night I just discovered fansubbed versions of _Last Exile_. Awesome visual style harkening to the WWI era of aircraft mixed with Star Wars pod racing. Just plain *Frikin' A*. Get it from Anime-Kraze as a Bit Torrent download.