My impression is that the OP was thinking about distribution for all of the low power devices that currently use wall warts - you might be looking at 200W for a typical house.
If I was going to pick a distribution voltage - I'd go for 42V. Why? Because that is going to be the new standard voltage for cars and there should be lots of stuff that will run off that voltage. Other issue is that it is comfortably below the cut-off for low-voltage safety issues.
I know that DC power transmission went out with Westinghouse's AC transmission, mostly because DC doesn't really work over significant distances.
That's true of the low voltage DC used in Edison's distribution system (which, by the way, is why American homes have 120/240V single phase service). High voltage DC (+/- 400 KV or so) is actually better than AC for long distance power transmission - due to lower losses and not having to keep the neds in synchronism.
Assuming that it can work like that, why isn't there DC power distribution alongside AC in many situations?
Hmmm, the Edison system was still in active use in some cities until recently, that is you could get 120V DC power from the utility.
Seriously, a DC power distribution does have a lot of merit (power over ethernet??) - it would make a hell of a lot more sense for a UPS to put out DC instead of having to convert the battery voltage from DC to AC. Even if a voltage conversion step is necessary, it would be more effiecient than a wall wart. It would also make sense for those living "off grid".
If you turn the clock on your DeLorean back to the 1930's - a lot of the ranch and farm houses would have a windcharger for the lights and would have a collection of 32 volt appliances. 32 volts was also common lighting voltage used in railroad passenger cars of that era.
Question is, what exactly is the hardware required to shift DC voltage?
If no galvanic isolation is needed, a chopper works fine. Voltage step-down or step-up is adjusted by varying the duty cycle of the switch and free-wheeling diode.
FAT is an old time file system created in the days of DOS.
The original FAT filesystem was developed by Paul Allen for Disk BASIC (originally 1976??) - i.e. a standalone BASIC interperter that did not require an OS. A version of Disk BASIC was ported over to Seattle Computer's 8086 system about November 1979 - 3 months later Tim Patterson of SCP got tired of waiting for Digital research to ship CP/M-86 and proceeded to write QDOS. Tim adopted the FAT system from Disk-BASIC to QDOS, figuring it was a bit more flexible than the bitmap allocation used on CP/M.
The 8.3 filename convention was adopted from CP/M which was presumably adopted from DEC's RSX-11. Needless to say, there is no way that MS could assert patent rights to FAT using 8.3 filenames.
Somewhere in the 1980's, an article in SA used North Korea as an example of how planned economies were better than market economies.
It also has been a couple of decades since they've correctly identified the name of the chunk of granite that the Palomar Observatory sits on. That's being nitpicky - but I figure if they're slopy about a simple thing as a place name, then where else are they falling down on accuracy.
The correct name is Palomar Mountain - SA used to use that name.
I've seen references to increased melting of glaciers since 1850 - the reading somewhere else that the littice ice age ended in 1850.
FWIW, other studies have shown man's effect on the CO2 levels starting at least 8,000 years ago - primarily due to the development of agriculture. Even now, the agricultural production of CO2 probably rivals the production from buring fossil fuels.
IIRC, they produced an IBM PC-like computer (not a full clone in the Compaq sense). The founder was celebrating a successful IPO with a fast drive down Highway 17 when he lost control of the car.
I don't think you could tax foreign workers working in foreign countries even if they worked from American firms.
I wouldn't think so either.
There was a recent change to the US tax code that treats dividend income differently than interest income. The rationale is that beforehand, dividends were subject to double taxation - first at the corporate level (before dividends are paid out) and at the individual level. In actuality, dividends are often subject to triple taxation, first the workers wages are taxed, then the corporate profits and finally the dividends received by the individuals.
My point is that by outsourcing labor, corporations are reducing payroll tax revenues. I can accept tax credits if those credits lead to increased domestic employment (and increased payroll tax revenues), but if the corporations want to outsource, well let them pay for the work on their own ticket.
That's all well and good, but if your job was the one making the product at $1, and they decided to outsource it to [insert country here] for production and you're now unemployed and have no income, does it matter that the item which used to cost $1 is now $0.50? You can't afford it because you're worrying about your [insert payment schedule here] bills.
And what happens when a good portion of the workforce ends up like that? Who is going to buy your products? Outsourcing comes across to me as the current equivalent of the dot-com pseudo-boom. At first people are making money, but after a while reality starts catching up.
I'd like to propose a few changes to the corporate tax structure.
No investment or R&D tax credits for offshore work.
No dividend tax exclusion for profits earned offshore.
A return to the 90% marginal tax rates for offshore profits.
Henry Ford was able to make money by paying his workers much higher than the average salaries and at the same time, he increased the market for his cars because more people could afford them.
They are only required by law to support it for a limited number of years (like 15)
To qualify for California's Partial Zero Emission Vehicle (PZEV) status, the emissions control systems have to have 15 year/ 150,000 mile warranties. If you to seel cars in California, a certain percentage of the cars need to be PZEV compliant.
Perhaps it made sense at one time for KDE to try to replicate every Microsoft desktop software.
KOffice replicates MS Office the same way that Linux with a window manager replicates M$-Windows. The most intriguing part about KOffice is that it doesn't slavishly follow M$-Office. KWrite borrows heavily from DTP, where OOo's Text personna is primarily a word processor.
On the CDC-6600 and 7600, single precision reals were 60 bits and double precision reals were 120 bits (well 1 bit for sign, 11 for exponent and 96 bits for the mantissa).
Sun Fortran has support for 128 bit reals (IIRC copied from DEC) although handling of those beasts is done by software.
Note that while Intels are 80 bit internally, as soon as you move off the CPU, you're back down to 64 bits. Someone correct me if I'm wrong, but I believe there's an option in gcc to force IA32 code to be IEEE compliant, and it works by inserting a load and save before every floating point operation (forcing the value off chip, which kills performance but causes things to get truncated to 64 bits).
Dunno about the Pentia, but the original 8087 implementation had a control word that would force rounding to either 32 or 64 bits. Also bear in mind that the SSE facilities in the latest Pentia do not use the 80 bit internal representation (which probably disgusts Kahan, but...).
There would be one slight problem with running a train from Scotland to Africa vis Spain - Spain uses a wider gauge than the rest of Europe (IIRC 5'6" versus 4'8.5" - I'm using Imperial units since the track gauges were originally defined in Imperial units).
There are some trains that can shift gauge, but most railway engineers think that's asking for trouble.
Another issue is loading gauge (essentially how big the cross section of the train can be and not create problems). The UK has a really tight loading gauge and the Chunnel has a HUGE loading gauge.
From what I've read - the Opteron version of Solaris is supposed to be released summer 2004.
Equally important is Opteron support for the development tools (i.e. compiler collection) to ease the transition from Sparc to Opteron and vice versa. Done right, Opteron support for Solaris could help Sparc sales as a wider installed base for 64 bit Solaris would encourage porting more applications to Solaris.
The reviewer comes across as a Windoze user who then became a linux geek - Way too focused on GUI administration. Real Solaris admin's prefer to use command line tools over the serial port.
My install went a lot more smoothly than Jem's, the only issue was not getting more than 8 bits of color. The network came up with no problem when installing Solaris x86 - was even able to browse the web during installation.
My experience is that Linux beats Solaris in device support, SMB support and eye candy. Solaris comes across as more polished, header files are in standard locations, IPsec is supported semi-natively and has better PostScript support built into the OS.
I haven't found anything as nice to do simple photo editing (cropping, resizing, printing) as sdtimage - though it would be nice to have png and JPEG2000 support.
I was doing the jamming in that situation, I would probably be looking at one hell of a lawsuit - especially if someone's life was lost due to not getting the call.
If the CTIA really cares about the customers they would take a much more active role in going after the manufacturers/distributors of jamming devices - much as the *cough* *cough* xxAA's are going after the purveyors of (ahem) copyright infringement technology.
One of the biggest differences between PC-DOS and SCP's DOS was that it used BillG's filesystem
The 86-DOS documentation that came with my SCP computer stated that 86-DOS version 1 and 1.14 both used a FAT with 12 bits per allocation entry. Cluster size was 512 bytes for 8" SSSD and DSSD (4 sectors) and 1024 bytes for 8" DSDD (1 sector)(rule of thumb was cluster size should be the square root of the disk size). There were some minor changes in the filesystem between 86-DOS v1.14 and MS-DOS v1.25, primarily in the size of the directory. In both 86-DOS and PC-DOS v1.x, the contents of the FAT were stored in RAM, which avoided disk accesses for reading the FAT (and introduced an exciting bug for systems that couldn't tell when a drive door had been opened).
The directory layout on PC-DOS was not changed from 86-DOS to PC-DOS. The directory entries included the date and time stamps - a 4 byte filelength entry replaced CP/M's bitmap.
There were two major areas of difference between 86-DOS and PC-DOS. One was that many of the utilities in PC-DOS relied on the IBM firmware. The other was that PC-DOS left out many of the SCP utilities, e.g. READCPM, the SCP assembler, HEX2BIN.
The first major change from 86-DOS came with DOS v2.0.
SCP's QDOS was written in Feb/March 1980 and released as 86-DOS in late 1980. While the idea of using a FAT was based on BillG's Disk Basic, the specific implementation of FAT-12 was Tim Patterson's work while at SCP. Also recall that 86-DOS was written in such a way that it was very easy to port CP/M programs - i.e. similar data structures for the file control blocks.
And it still wouldn't get you the same service as 911.
One of the benefits of a landline is that the 911 operators have your address on the screen when they take the call - and many jurisdictions will send someone out if there is a hang-up.
In California, 911 calls from cell phones are handled by the CHP, which may not be as helpful as the normal 911 dispatch.
Slashdot Mirror
If I was going to pick a distribution voltage - I'd go for 42V. Why? Because that is going to be the new standard voltage for cars and there should be lots of stuff that will run off that voltage. Other issue is that it is comfortably below the cut-off for low-voltage safety issues.
Then there's Ward Kimball's Emma Nevada, a real 3 foot gauge Mogul (2-6-0) - though the last time it ran under steam was in the late 40's.
That's true of the low voltage DC used in Edison's distribution system (which, by the way, is why American homes have 120/240V single phase service). High voltage DC (+/- 400 KV or so) is actually better than AC for long distance power transmission - due to lower losses and not having to keep the neds in synchronism.
Assuming that it can work like that, why isn't there DC power distribution alongside AC in many situations?
Hmmm, the Edison system was still in active use in some cities until recently, that is you could get 120V DC power from the utility.
Seriously, a DC power distribution does have a lot of merit (power over ethernet??) - it would make a hell of a lot more sense for a UPS to put out DC instead of having to convert the battery voltage from DC to AC. Even if a voltage conversion step is necessary, it would be more effiecient than a wall wart. It would also make sense for those living "off grid".
If you turn the clock on your DeLorean back to the 1930's - a lot of the ranch and farm houses would have a windcharger for the lights and would have a collection of 32 volt appliances. 32 volts was also common lighting voltage used in railroad passenger cars of that era.
Question is, what exactly is the hardware required to shift DC voltage?
If no galvanic isolation is needed, a chopper works fine. Voltage step-down or step-up is adjusted by varying the duty cycle of the switch and free-wheeling diode.
The original FAT filesystem was developed by Paul Allen for Disk BASIC (originally 1976??) - i.e. a standalone BASIC interperter that did not require an OS. A version of Disk BASIC was ported over to Seattle Computer's 8086 system about November 1979 - 3 months later Tim Patterson of SCP got tired of waiting for Digital research to ship CP/M-86 and proceeded to write QDOS. Tim adopted the FAT system from Disk-BASIC to QDOS, figuring it was a bit more flexible than the bitmap allocation used on CP/M.
The 8.3 filename convention was adopted from CP/M which was presumably adopted from DEC's RSX-11. Needless to say, there is no way that MS could assert patent rights to FAT using 8.3 filenames.
It also has been a couple of decades since they've correctly identified the name of the chunk of granite that the Palomar Observatory sits on. That's being nitpicky - but I figure if they're slopy about a simple thing as a place name, then where else are they falling down on accuracy.
The correct name is Palomar Mountain - SA used to use that name.
I've seen references to increased melting of glaciers since 1850 - the reading somewhere else that the littice ice age ended in 1850.
FWIW, other studies have shown man's effect on the CO2 levels starting at least 8,000 years ago - primarily due to the development of agriculture. Even now, the agricultural production of CO2 probably rivals the production from buring fossil fuels.
You're being too hard on yourself - Windblows 1.0 was announced in 1983.
IIRC, they produced an IBM PC-like computer (not a full clone in the Compaq sense). The founder was celebrating a successful IPO with a fast drive down Highway 17 when he lost control of the car.
I wouldn't think so either.
There was a recent change to the US tax code that treats dividend income differently than interest income. The rationale is that beforehand, dividends were subject to double taxation - first at the corporate level (before dividends are paid out) and at the individual level. In actuality, dividends are often subject to triple taxation, first the workers wages are taxed, then the corporate profits and finally the dividends received by the individuals.
My point is that by outsourcing labor, corporations are reducing payroll tax revenues. I can accept tax credits if those credits lead to increased domestic employment (and increased payroll tax revenues), but if the corporations want to outsource, well let them pay for the work on their own ticket.
I have no qualm with oversea's R&D, but I do object to subsidizing it through the tax structure.
And what happens when a good portion of the workforce ends up like that? Who is going to buy your products? Outsourcing comes across to me as the current equivalent of the dot-com pseudo-boom. At first people are making money, but after a while reality starts catching up.
I'd like to propose a few changes to the corporate tax structure.
No investment or R&D tax credits for offshore work.
No dividend tax exclusion for profits earned offshore.
A return to the 90% marginal tax rates for offshore profits.
Henry Ford was able to make money by paying his workers much higher than the average salaries and at the same time, he increased the market for his cars because more people could afford them.
To qualify for California's Partial Zero Emission Vehicle (PZEV) status, the emissions control systems have to have 15 year/ 150,000 mile warranties. If you to seel cars in California, a certain percentage of the cars need to be PZEV compliant.
KOffice replicates MS Office the same way that Linux with a window manager replicates M$-Windows. The most intriguing part about KOffice is that it doesn't slavishly follow M$-Office. KWrite borrows heavily from DTP, where OOo's Text personna is primarily a word processor.
doubles are 64 bits.
Sez who?
On the CDC-6600 and 7600, single precision reals were 60 bits and double precision reals were 120 bits (well 1 bit for sign, 11 for exponent and 96 bits for the mantissa).
Sun Fortran has support for 128 bit reals (IIRC copied from DEC) although handling of those beasts is done by software.
Note that while Intels are 80 bit internally, as soon as you move off the CPU, you're back down to 64 bits. Someone correct me if I'm wrong, but I believe there's an option in gcc to force IA32 code to be IEEE compliant, and it works by inserting a load and save before every floating point operation (forcing the value off chip, which kills performance but causes things to get truncated to 64 bits).
Dunno about the Pentia, but the original 8087 implementation had a control word that would force rounding to either 32 or 64 bits. Also bear in mind that the SSE facilities in the latest Pentia do not use the 80 bit internal representation (which probably disgusts Kahan, but...).
There are some trains that can shift gauge, but most railway engineers think that's asking for trouble.
Another issue is loading gauge (essentially how big the cross section of the train can be and not create problems). The UK has a really tight loading gauge and the Chunnel has a HUGE loading gauge.
k) You think that CDE and Motif are Sun creations.
l) You wonder why CDE (actually xsun) doesn't have better TrueType support. m) You think any deviation from the GNU interface is non-standard.
Equally important is Opteron support for the development tools (i.e. compiler collection) to ease the transition from Sparc to Opteron and vice versa. Done right, Opteron support for Solaris could help Sparc sales as a wider installed base for 64 bit Solaris would encourage porting more applications to Solaris.
My install went a lot more smoothly than Jem's, the only issue was not getting more than 8 bits of color. The network came up with no problem when installing Solaris x86 - was even able to browse the web during installation.
My experience is that Linux beats Solaris in device support, SMB support and eye candy. Solaris comes across as more polished, header files are in standard locations, IPsec is supported semi-natively and has better PostScript support built into the OS.
I haven't found anything as nice to do simple photo editing (cropping, resizing, printing) as sdtimage - though it would be nice to have png and JPEG2000 support.
If the CTIA really cares about the customers they would take a much more active role in going after the manufacturers/distributors of jamming devices - much as the *cough* *cough* xxAA's are going after the purveyors of (ahem) copyright infringement technology.
The parent posting is well worth reading.
From the Open Boot Prompt simply type: boot net
The 86-DOS documentation that came with my SCP computer stated that 86-DOS version 1 and 1.14 both used a FAT with 12 bits per allocation entry. Cluster size was 512 bytes for 8" SSSD and DSSD (4 sectors) and 1024 bytes for 8" DSDD (1 sector)(rule of thumb was cluster size should be the square root of the disk size). There were some minor changes in the filesystem between 86-DOS v1.14 and MS-DOS v1.25, primarily in the size of the directory. In both 86-DOS and PC-DOS v1.x, the contents of the FAT were stored in RAM, which avoided disk accesses for reading the FAT (and introduced an exciting bug for systems that couldn't tell when a drive door had been opened).
The directory layout on PC-DOS was not changed from 86-DOS to PC-DOS. The directory entries included the date and time stamps - a 4 byte filelength entry replaced CP/M's bitmap.
There were two major areas of difference between 86-DOS and PC-DOS. One was that many of the utilities in PC-DOS relied on the IBM firmware. The other was that PC-DOS left out many of the SCP utilities, e.g. READCPM, the SCP assembler, HEX2BIN.
The first major change from 86-DOS came with DOS v2.0.
SCP's QDOS was written in Feb/March 1980 and released as 86-DOS in late 1980. While the idea of using a FAT was based on BillG's Disk Basic, the specific implementation of FAT-12 was Tim Patterson's work while at SCP. Also recall that 86-DOS was written in such a way that it was very easy to port CP/M programs - i.e. similar data structures for the file control blocks.
PC-DOS 1.0 was a re-warmed 86-DOS 1.14
Ummm, CDE is basically HP's VUE.
And it still wouldn't get you the same service as 911.
One of the benefits of a landline is that the 911 operators have your address on the screen when they take the call - and many jurisdictions will send someone out if there is a hang-up.
In California, 911 calls from cell phones are handled by the CHP, which may not be as helpful as the normal 911 dispatch.