The article discusses ultrawideband and narrowband technologies. For the narrowband weapons to be effective, the amount of power required to generate a continous pulse (more than a second) would be impractical (from the print article):
To drive the Sinus-6's beam continuously for an entire second, you'd need to supply about 25 gigajoules--'the entire output of a typical coal-fired electrical plant for 10 full seconds'.
Also, the article mentions that, at those power levels, the air around the emitter "would heat to a plasma that in turn would interfere with a continuous beam". The practical pulse durations mentioned were around 10-nanoseconds, at 200 pulses a second, which turns out to be a dutycyle of somewhere around 2e-7. Not much harm to humans.
Now, it's too bad we didn't have this weapon last year for use in the invasions of Afghanistan and Iraq. Imagine the effectiveness of such a weapon!
I believe the Spectrum article suggests that the US might have already tried some of these types of weapons to disable television and radio stations in Iraq. To be effective, the weapons have to be detonated at relatively close range (unless they are powered by a nuclear device!), so I don't think that an entire battlefield could be easily "disabled", though.
but it never went anywhere. Digital/Compaq had a prototype that used a tilt sensor for navigation (I think it even had a touch screen). If there ever was a cool credit card PDA, this would have been it.
Hence: To draw an illegitimate profit from; to speculate on; to put upon.
That being said, I think the term Communism is missued in the grandparent's post. The software is not being placed in the public domain, and the state does not have any special rights to the software beyond what the license grants to individuals.
Isn't using the TCP stack a bad idea for hard realtime applications? I suppose you could try to UDP data out, or send out via some kind of streaming protocol, but you wouldn't need a hard realtime kernel to do *that* part. RTLinux provides a software FIFO to send data to the non-(hard) realtime side for interfacing to networks, disk drives, etc. It's true that you'd have to have the scsi drivers on the realtime side if you were trying to stream data directly to disk, but then you could probably get away with sending the data to the non-realtime side and just add enough buffering to handle the timing jitter. Usually the hard realtime requirements are for interrupt-driven data devices like A/D and D/A boards, or specialty interfaces.
This is really the only answer. Currently it's not against the law to send email with forged/spoofed return address. If I started selling stuff over the regular mail using someone else's name/address, I'd be arrested for mail fraud. That's all spam is, after all, and should have the existing laws covering fraud modified to apply.
You've probably already checked out the various approaches to realtime applications for Linux:
RTLinux with a few case studies I believe the latency for this approach is at the low microsecond level (from the website):
FSMLabs RTLinuxPro building block is a tested and validated, hard real-time, POSIX operating system that runs Embedded Linux as an application platform. The RTCore real-time kernel at the heart of RTLinuxPro provides rock-solid, low microsecond worst case interrupt latency and scheduling jitter plus seamless access to Linux.
Its main purpose is to let engineers evaluate a new way of propelling spacecraft, on far-ranging space missions. Power from SMART-1's solar panels will drive an electric propulsion system called an 'ion engine'. The demonstration task is to overcome the Earth's gravity and put the spacecraft into orbit around the Moon.
This appears to be a demonstration project, with the final application in extended range projects (where traditional chemical fuel rockets would have to carry too much fuel mass). Because it's constantly accellerating, this thing (in theory) could hit velocities not practical for chemical rockets.
Re:which taxes? Income taxes? Social Security tax?
on
Tech Rich Get Richer
·
· Score: 1
Tax cuts aren't "financed" by *anything*. Government programs are financed by taxes and the sale of government bonds. The people benefiting from the government programs (or at least the ones approving the budgets) are the ones "transferring money from future taxpayers", not the current taxpayers. It's a little more than troubling to hear some insisting that the federal budget has a tax cut listed on the expenditures side of the federal ledger.
Re:Why do you think Bush gave them tax cuts?
on
Tech Rich Get Richer
·
· Score: 1
If the goal of the government (or at least one of them) is to redistribute income (or to help the less fortunate), then they should institute a program designed to meet this goal, instead of using the tax code (i.e. the revenue-generating part of the government). I would think that if the government really cared about helping individuals with need, it would increase the standard deduction for individuals to somewhere in the neighborhood of $10K-20K (so that the government doesn't tax those in need at *all*).
Pretty much on the mark. The daily cleaning thing, however, is my idea -- mostly as a maintenance thing for our hardwood floors. The one thing that will destroy your hardwoods is dust an dirt that is ground in when you walk on it. Also, by the time dinner is over, there is a debris field scattered throughout the entire first floor (Cheerios, crumbs, etc). If the Roomba doesn't get it, my 10-month-old daughter will certainly the next day.
From my experience with MkLinux [mklinux.org], Mach is a microkernel that lives
between the OS kernel and the hardware. OS X is supposed to have most of the BSD kernel (aka Darwin [apple.com]), with modifications to run on top of the Mach microkernel.
OS kernel ----------- microkernel -----------
hardware
We got a Roomba last December, and it took a little while to get our 1st floor completely "roomba-ready". All the kid's toys get thrown in baskets, I pick up a tassled rug and throw it over a radiator, and I use a virtual wall unit to block a threshold that the Roomba *always* gets stuck on (even though it's below the advertised limit). Now, after dinner, we get our entire 1st floor roomba-ready (pick up toys, move chairs, etc.), and all go upstairs for the kid's baths + stories + bedtime. By the time the kid's are in bed, the Roomba's done (45-60 minutes or so). We still have to vacume every other week or so, but the Roomba keeps the floors cleaner between manual cleaning. Sure, I could sweep every night, but the Roomba does it while I'm doing something else. I figure it saves me 10-15 minutes each day, which is great for those of us with small children. So in the past year, I figure the Roomba has saved me 60-90 hours of free time, which translates to about $2-3/hour. If the thing lasts another year without any maintenance costs, then we're talking about $1-1.50/hour. Would I spend a dollar for an extra hour of free time? (I guess I already do!)
It's gotten to the point where my 3-year-old daughter can pick it up, put it in the middle of the floor, and turn it on by herself.
I think the Electrolux version costs upwards of $1k (US), while the Roomba is $200. The Electrolux version also doesn't have "cliff" sensors to keep it from falling off stair edges.
There's a fairly well-documented theory that the Hindenburg accident was really caused by the flamable skin of the airship. A quick Google search renders a few sites: Rice U. Clean-Air.org AmericanHistory.about.com
Just to name a few. At least let's not have a bunch of people using the Hindenburg as a reason not to think about hydrogen.
Re:The last of the Apple-based OSX machines?
on
G5s Start Shipping
·
· Score: 1
OS X is based on the Mach kernel with a BSD userland.
From my experience with MkLinux, Mach is a microkernel that lives between the OS kernel and the hardware. OS X is supposed to have most of the BSD kernel (aka Darwin), with modifications to run on top of the Mach microkernel.
I've used TeXmacs too, and it is under active development. You can import/export LaTeX code with LyX (reLyX works fairly well on LaTeX import) if you need.tex output. It's relatively simple to add new document classes to LyX, although I don't have any experience with this in TeXmacs. I stayed with LyX because I like the "Document processor" model where formatting is secondary to content. LyX does not allow things like two spaces in a row, or more than one linefeed (that's what the document class is for).
I do like TeXmacs' interface to things like Octave, which renders things quite nicely.
The 1.3.2 version is quite a bit better than some of the older versions. I like the in-line equation editor, which give instant feedback on equation formatting. Other environments give instant feedback, as well as figure/table float previews. Most of the "flexibility" people like about LaTeX reduces to the same "fiddling" that a Document processor is supposed to reduce. If you need a different document format, you really should work on a new document class (.cls). LyX does (as of 1.3.2) allow one to adjust things like margins and paper size within a class, though. If I need special packages, I just include them in the preamble, and then insert the raw LaTeX code with a '\' (like./clearpage when I start getting too many unprocessed floats!).
LyX makes LaTeX almost useable. The latest development version (1.3.2) supports MacOS X, Windoze (via cygwin), and can use either an xforms gui or Qt. For technical writing, it really can't be beat.
I had GEOS for my C64, and it really was very impressive for the level of hardware it was running on. It was supposed to be an entire graphical OS written entirely in assembly language (yikes!). It didn't have anything like a TCP/IP stack (neither did anything else in those days!).
If you look at the dpi (dots per inch) of each display, they *are* keeping the display resolution the same while increasing the display *size* (as well as aspect ratio). There are alot of constraints in selecting the pixel dimensions of the screen, and, past a certain point, increased dpi doesn't buy you much (except for additional eyestrain).
the difference between the 12 and 17 inch models is (106.67-99.9)/106.67 ~= 0.65, or a 6.5% reduction in dpi. I think the idea was to try to maintain dpi while increasing display area (two pictures at the same zoom level should be the same physical size on each screen). I would imagine that the 17" model was designed to have a 16:10 aspect ratio to fit the 16:9 widescreen video with some space left on the screen (for controls or whatever).
I wrote a small, heavily CPU intensive program which performed many typical DSP calculations, compiled it with gcc and ran it on a desktop athlon-xp 1700. Without optimizations, this test routine took eight seconds to complete, with optimizations it took four. Then, I compiled and ran it on an 867 Mhz G4 (same compiler): the typical routine took 30 seconds to complete without optimizations and 8 seconds with optimizations. Of course, you can't really make any scientific conclusions here, but it is suspicious at least.
Only the more recent versions of gcc (or patched versions) have AltiVec support out of the box, and only through use of the special vector routines (just turning on optimization won't help). Otherwise, you might as well compare performance to a PPC 750 (aka G3)
Short of paying to have an "unlisted" phone number, or changing my phone number every 6 months, there's not much we can do until everyone refuses to buy junk they don't need from people they don't know. Telemarketing is just another one of those "ruined it for the rest of us" cases where now I refuse to buy anything solicited to me over the phone (why would I anyway?). The real problem is human nature/behavior, because the only reason Telemarketers try to sell us something we don't need over the phone or via email is because they think there are enough (as a percentage of population) simple-minded people that would fall for the trick.
There has to be a change of behavior on the consumer's part before telemarketing becomes cost ineffective.
Slashdot Mirror
but it never went anywhere. Digital/Compaq had a prototype that used a tilt sensor for navigation (I think it even had a touch screen). If there ever was a cool credit card PDA, this would have been it.
Roomba
That being said, I think the term Communism is missued in the grandparent's post. The software is not being placed in the public domain, and the state does not have any special rights to the software beyond what the license grants to individuals.
Isn't using the TCP stack a bad idea for hard realtime applications? I suppose you could try to UDP data out, or send out via some kind of streaming protocol, but you wouldn't need a hard realtime kernel to do *that* part. RTLinux provides a software FIFO to send data to the non-(hard) realtime side for interfacing to networks, disk drives, etc. It's true that you'd have to have the scsi drivers on the realtime side if you were trying to stream data directly to disk, but then you could probably get away with sending the data to the non-realtime side and just add enough buffering to handle the timing jitter. Usually the hard realtime requirements are for interrupt-driven data devices like A/D and D/A boards, or specialty interfaces.
This is really the only answer. Currently it's not against the law to send email with forged/spoofed return address. If I started selling stuff over the regular mail using someone else's name/address, I'd be arrested for mail fraud. That's all spam is, after all, and should have the existing laws covering fraud modified to apply.
RTLinux with a few case studies
I believe the latency for this approach is at the low microsecond level (from the website):
See this article for the rest.
This appears to be a demonstration project, with the final application in extended range projects (where traditional chemical fuel rockets would have to carry too much fuel mass). Because it's constantly accellerating, this thing (in theory) could hit velocities not practical for chemical rockets.
Tax cuts aren't "financed" by *anything*. Government programs are financed by taxes and the sale of government bonds. The people benefiting from the government programs (or at least the ones approving the budgets) are the ones "transferring money from future taxpayers", not the current taxpayers. It's a little more than troubling to hear some insisting that the federal budget has a tax cut listed on the expenditures side of the federal ledger.
If the goal of the government (or at least one of them) is to redistribute income (or to help the less fortunate), then they should institute a program designed to meet this goal, instead of using the tax code (i.e. the revenue-generating part of the government). I would think that if the government really cared about helping individuals with need, it would increase the standard deduction for individuals to somewhere in the neighborhood of $10K-20K (so that the government doesn't tax those in need at *all*).
Pretty much on the mark. The daily cleaning thing, however, is my idea -- mostly as a maintenance thing for our hardwood floors. The one thing that will destroy your hardwoods is dust an dirt that is ground in when you walk on it. Also, by the time dinner is over, there is a debris field scattered throughout the entire first floor (Cheerios, crumbs, etc). If the Roomba doesn't get it, my 10-month-old daughter will certainly the next day.
To clarify:
OS kernel
-----------
microkernel
-----------
hardware
We got a Roomba last December, and it took a little while to get our 1st floor completely "roomba-ready". All the kid's toys get thrown in baskets, I pick up a tassled rug and throw it over a radiator, and I use a virtual wall unit to block a threshold that the Roomba *always* gets stuck on (even though it's below the advertised limit). Now, after dinner, we get our entire 1st floor roomba-ready (pick up toys, move chairs, etc.), and all go upstairs for the kid's baths + stories + bedtime. By the time the kid's are in bed, the Roomba's done (45-60 minutes or so). We still have to vacume every other week or so, but the Roomba keeps the floors cleaner between manual cleaning. Sure, I could sweep every night, but the Roomba does it while I'm doing something else. I figure it saves me 10-15 minutes each day, which is great for those of us with small children. So in the past year, I figure the Roomba has saved me 60-90 hours of free time, which translates to about $2-3/hour. If the thing lasts another year without any maintenance costs, then we're talking about $1-1.50/hour. Would I spend a dollar for an extra hour of free time? (I guess I already do!)
It's gotten to the point where my 3-year-old daughter can pick it up, put it in the middle of the floor, and turn it on by herself.
I think the Electrolux version costs upwards of $1k (US), while the Roomba is $200. The Electrolux version also doesn't have "cliff" sensors to keep it from falling off stair edges.
There's a fairly well-documented theory that the Hindenburg accident was really caused by the flamable skin of the airship. A quick Google search renders a few sites:
Rice U.
Clean-Air.org
AmericanHistory.about.com
Just to name a few. At least let's not have a bunch of people using the Hindenburg as a reason not to think about hydrogen.
From my experience with MkLinux, Mach is a microkernel that lives between the OS kernel and the hardware. OS X is supposed to have most of the BSD kernel (aka Darwin), with modifications to run on top of the Mach microkernel.
I've used TeXmacs too, and it is under active development. You can import/export LaTeX code with LyX (reLyX works fairly well on LaTeX import) if you need .tex output. It's relatively simple to add new document classes to LyX, although I don't have any experience with this in TeXmacs. I stayed with LyX because I like the "Document processor" model where formatting is secondary to content. LyX does not allow things like two spaces in a row, or more than one linefeed (that's what the document class is for).
I do like TeXmacs' interface to things like Octave, which renders things quite nicely.
The 1.3.2 version is quite a bit better than some of the older versions. I like the in-line equation editor, which give instant feedback on equation formatting. Other environments give instant feedback, as well as figure/table float previews. Most of the "flexibility" people like about LaTeX reduces to the same "fiddling" that a Document processor is supposed to reduce. If you need a different document format, you really should work on a new document class (.cls). LyX does (as of 1.3.2) allow one to adjust things like margins and paper size within a class, though. If I need special packages, I just include them in the preamble, and then insert the raw LaTeX code with a '\' (like ./clearpage when I start getting too many unprocessed floats!).
LyX makes LaTeX almost useable. The latest development version (1.3.2) supports MacOS X, Windoze (via cygwin), and can use either an xforms gui or Qt. For technical writing, it really can't be beat.
I had GEOS for my C64, and it really was very impressive for the level of hardware it was running on. It was supposed to be an entire graphical OS written entirely in assembly language (yikes!). It didn't have anything like a TCP/IP stack (neither did anything else in those days!).
If you look at the dpi (dots per inch) of each display, they *are* keeping the display resolution the same while increasing the display *size* (as well as aspect ratio). There are alot of constraints in selecting the pixel dimensions of the screen, and, past a certain point, increased dpi doesn't buy you much (except for additional eyestrain).
12"(diag) Powerbook: 1024x768 -> 7.2"x9.6" -> 106.7 dpi, 1.33 aspect ratio
15"(diag) Powerbook: 1280x854 -> 8.325"x12.5" -> 102.58 dpi, 1.5 aspect ratio
17"(diag) Powerbook: 1440x900 -> 9"x14.4" -> 99.9 dpt, 1.6 aspect ratio
the difference between the 12 and 17 inch models is (106.67-99.9)/106.67 ~= 0.65, or a 6.5% reduction in dpi. I think the idea was to try to maintain dpi while increasing display area (two pictures at the same zoom level should be the same physical size on each screen). I would imagine that the 17" model was designed to have a 16:10 aspect ratio to fit the 16:9 widescreen video with some space left on the screen (for controls or whatever).
Hot swapping?
Only the more recent versions of gcc (or patched versions) have AltiVec support out of the box, and only through use of the special vector routines (just turning on optimization won't help). Otherwise, you might as well compare performance to a PPC 750 (aka G3)
I agree.
Short of paying to have an "unlisted" phone number, or changing my phone number every 6 months, there's not much we can do until everyone refuses to buy junk they don't need from people they don't know. Telemarketing is just another one of those "ruined it for the rest of us" cases where now I refuse to buy anything solicited to me over the phone (why would I anyway?). The real problem is human nature/behavior, because the only reason Telemarketers try to sell us something we don't need over the phone or via email is because they think there are enough (as a percentage of population) simple-minded people that would fall for the trick.
There has to be a change of behavior on the consumer's part before telemarketing becomes cost ineffective.