What he is proposing is to reduce the number of transistors on a chip, to increase its speed and reduce power usage. So in fact he is trying to reverse more's law.
I had a case of food poisoning while I was riding on the train to work. While in the train I forgot how to read the news paper, I could read words, but I could not understand sentences anymore. When I got at work I was a little bit better, but I should not have been working, but I was to sick to figure that out myself.
I programmed a tight piece (around 100 lines) of code that converted the data from a cobol structure to xml and back, using a template written in xml. After two hours a colleague saw me and said I should not be working and he drove me to the hotel.
That piece of code was used for months in production handling a lot of transactions without problem. But when I got back to that piece of code, I could not understand how it could ever work, following the code was hard because of the many recursions and it looked like it should never be able to do what it did. I ended up rewriting it, with a lot of comments, although still using parts of the design I thought up in that state.
So you might be surprised on what you would program when in under influence.
They have a few of those now in amsterdam, as a pedestrian it is very useful to decide if you want to wait for the green light, or just cross against the red light instead.
- If you ever want to write plugin for most applications (c++ plugins are funny and scary). - work on most operating system kernels (only a few are in c++). - if you need to know exactly how the language constructs its data structures and functions (c++ is to complex and implementation differs wildly).
You could of course program in C++ with only the C subset, but that is the same thing. And I often need to fall back on the C subset when I work with shared memory (putting instances of classes in shared memory is scary, I now use structs with methods), atomic operations and other weird things. But I also don't shy away of writing little pieces of assembly for special instructions that the compiler did not export.
And there are still quite a few people in the world that use an abacus instead of a calculator. In fact the user interface of an abacus is much faster for certain operations like adding a list of prices at a shop. But it is a lost art in most of the world.
actually, if this is reliable it seems to be much quicker than normal parking. it may help congestion (from ppl slowly parking, while othe ppl wait) in the streets of amsterdam.
unless the ball is tracking your eyes. See, if it is a hallucinations in your visual cortex it would move with your eyes.
Same thing as with floaters which you can sometimes see when the light is right, it is like it is moving around the room and you somehow try to follow it with your eyes, while it is actually following your eye movement instead. http://en.wikipedia.org/wiki/Floater
Actually you can buy the mpeg2 codec from apple as well.
http://www.apple.com/quicktime/mpeg2/
It costs $19.99, I guess most of that goes to mpeg2 licensing or something, which is why they are not including it in the main os. Maybe Microsoft has a better deal with them.
IPv6 will use less space in the router tables as the address space is split into a geographical tree. Or at least it used to be when I looked at it the last time, which was a long time ago. Anyway on that scheme the first few bits show the continent, then the country, then the ISP.
1. To use Java well you need to be a good programmer. 2. A good programmer doesn't want to use Java.
The problems I've seen most programmers make is initializing and configuring objects (like full blown XML parsers, which takes over a ms to initialize and only a few us to use and reset) in the performance critical part of the program. This can be easily solved by creating object on demand, but putting them on a shared stack when you are done with it for reuse by an other thread.
You can list AAAA records for a website on the same name.
The browser will use standard library gethostbyname() function call to query the DNS for all IP address listed, including IPv4 and IPv6 addresses all at the same time. Then the browser should prioritize (Safari, IE and Firefox all do) IPv6 addresses over IPv4 addresses and try to connect to the IPv6 address first, and fall back to IPv4 address if the connection failed.
Most software actually already handles IPv6 fine for years. Most software only need around 20 lines of C code extra to handle IPv6 properly.
All websites of google only list AAAA records if your ISP opted-in for AAAA. Normally google's DNS only shows A records, and use magic to show the AAAA record when your source IP address is from one from the opted-in ISPs.
However I believe youtube's content servers always list AAAA records (even if your ISP didn't opt-in), the content servers have a different name from www.youtube.com
As my ISP now did opt in I get with dig -t AAAA www.youtube.com:
www.youtube.com. 6057 IN CNAME youtube-ui.l.google.com.
youtube-ui.l.google.com. 300 IN AAAA 2a00:1450:8001::64
youtube-ui.l.google.com. 300 IN AAAA 2a00:1450:8001::65
youtube-ui.l.google.com. 300 IN AAAA 2a00:1450:8001::66
I am not sure if it works for you, but try (this is a host where the actual contents comes from): dig -t AAAA v10.lscache4.c.youtube.com
v18.lscache1.c.youtube.com. 6904 IN CNAME v18.lscache1.l.google.com.
v18.lscache1.l.google.com. 266 IN AAAA 2a00:1450:4001:2::15
For web it may be quite easy to have one or more companies set up a ipv4/ipv6 http proxy. As web server owner you only need to point your A record to this proxy. As the webbrowser sends its destination hostname to the proxy, the proxy can look up the AAAA record for this to find your actual web server.
But this is not really a solution for other services or even https.
In any case, there is quite a buzz now with quite a few ISP running "large" scale pilots for native IPv6 to the consumer. Currently these tests are for people who want to try it out. When these tests are done and when ISPs have locked down consumer router configuration (which they do now for Ipv4 as well), they will force it on to every customer.
The chicken and egg problem is beginning to be solved. 1. Youtube delivers the content (not the website, just the movies) over Ipv6 (by default, without provider Ipv6 agreement with google). 2. Apple routers by default set up an IPv6 tunnel. (for their back-to-your-mac system). 3. Computers are configured by default to use IPv6 when available.
These 3 caused an enormous spike in IPv6 Internet traffic since february. This traffic, because of the limited number of tunnel machines available, is not routed via an ideal path. These traffic patterns may cost more in peering fees for the provider, which they can solve by routing native IPv6 (shorter route to the mulihomed network like youtube probably has).
1. customer at 196.0.1.1 asks to "wget www.someone.com" 2. query for www.someone.com to the ISPs DNS/NAT 3. The DNS does a DNS lookup and only receives a AAAA address back. 2001::2345 4. The DNS request from the NAT box a free IPv4 address (on the NAT box) for host 196.0.1.1 to 2001::2345. The NAT replies 10.254.1.1 5. The DNS server returns A 10.254.1.1 as a reply for the www.someone.com request.
Now the NAT only needs to translate Ipv4->ipv6 packets. (196.0.1.1->20.254.1.1) -> 2001::2345
This scheme requires the NAT device to have a lot of IPv4 addresses on the inside (which may be from a private range). But those IPv4 addresses may be shared with multiple customers. I.e. a customer on 196.0.1.2 can also use 10.254.1.1 for an other translation, as the combination of source and destination need to be unique.
I am one of the customers of XS4ALL who is participating with the native IPv6 pilot. I myself have the Draytek Vigor 120 + Vigor 2130n combination. The Vigor 120 only does PPPoA to PPPoE translation. And the Vigor 2130n connects to the PPPoE and does DHCPv6 to retrieve the IPv6 addresses.
From what I understand a modern Airport Extreme or a modern TimeCapsule also can do DHCPv6 and connect over PPPoE through the Vigor 120. But sadly I have an old TimeCapsule which does not allow its firmware to be upgraded.
Both of the draytek 2130 and the time capule has a nice user interface and the only thing to do is select DHCPv6. I am expecting modems from ISPs that will be pre-configured so that the customer doesn't need to know anything about networks.
For the google webservers yes. In fact the youtube website and the google website themselves were (I now have native Ipv6 and my provider seems to have now an agreement with google) still published on IPv4 for me. The media for youtube however is published as IPv6 for everyone.
There are currently two companies forcing the hand of the consumer ISPs to adopt IPv6.
Since February this year Youtube has put all the actual media reachable on IPv6 as default when you access the youtube website through their normal DNS name. Apple's time capsule and airport extreme by default sets up IPv6 through tunnels.
This means that a lot of people with Apple computers browsing youtube movies are heavy users of IPv6. As there are only a few tunnel brokers, the load on those will be quite high.
It does actually. As a programmer I can write my application to use only lines, splines, circles and so to design my widgets. there is also an API to convert floating point coordinates to pixel coordinates to align lines to pixel boundaries so that antialiasing does cause a grey double width line. There is even a tool to change the DPI of the display to check if your application draws pretty.
The point probably is, that there are no high DPI displays to enable the feature for the customer. Which also causes the application writers to ignore pixel boundaries, because DPI changes are not enabled for the customer. Which is maybe why the LCD display manufactures are not creating high DPI displays.
I think at some point, when all the young generation is grown up. They expect to see compromising pictures on your facebook. And if you don't have those they wont trust you and you won't get a job.
Actually, these machines are probably doing "high speed machining" which can only be done when they are software controlled, i.e. it can not be done manually.
In high speed machining, the cutting speed is much higher than is intuitive. As one increases the cutting speed (from zero), the temperature of the tool increases, until it breaks. But there is a sweet spot at a much higher cutting speed where the tool actually is cooled by the cuttings it is removing.
However this sweet spot is tight and it depends on the angle of attack, the depth of the tool, the forward speed, the rotational speed, and the amount of material being removed. And these must be maintained, as the material is entering the material, leaving the material, makes corners, etc.
I've noticed this for a while, it is not just the harddisk. From the first clone XT PC to a modern DELL computer, the prices haven't changed much at all.
Although one of the thread children says he could get a 1GB drive with a SCSI controller for $1300, he should not forget to compare that to the highend of today, which would be a SAS controller (400 dollar for real RAID (not fake raid)) with SAS disk (500 dollar for a 10,000 rpm).
Intel was not the only one that was shipping 64 bit processors.
DEC Alpha was a 64 bit processor that was manufactured in the mid 1990s. Linux was running on the Alpha processor (native 64 bit) way before the year 1999. The MIPS R4000 was a 64 bit processor that was manufactured at the beginning of the 1990s. In this case Linux was running on these as well. The sparc64 was made in the mid 1990s, not sure about the status of linux on those at the time, but I am pretty sure it was ported to these systems as well.
In any case Linux was already running on quite a few processors natively in 64 bit (this includes a complete 64 bit kernel and a 64 bit user-land system, as most of these processors didn't have a 32 bit mode). Therefor it was relatively trivial to port to the new Itanium processors and was running in the labs before the processor was released to the public.
We actually build that model at work, it certainly does OCR autonomic. It scans each character it finds by moving the sensor over the character. Our version needed quite a few tries before it was able to read all the numbers correctly though.
Actually for at least a formula 1 car 20% slip during acceleration is optimal for acceleration, although not for tire wear. From the article below it says that traction control can set the amount of slip during acceleration as ideal as possible. Where ideal may be for either tire wear or acceleration depending on the circumstances of the race.
If the result is "yes": 10,000 + 10,000 * 6/1 = 70,000 (original stake + winnings)
If the result is "no": 50,000 + 50,000 * 2/5 = 70,000 (original stake + winnings)
In this case it is always good to bet, because you will always win 10,000.
Re:Motion blur and bloom effects
on
Framerates Matter
·
· Score: 4, Informative
But, if you follow the hand with your eyes, your hand will appear sharp. You'll be supprised how quickly and stable eyes can track moving objects.
The BBC has been experimenting with fast frame rate TV, running at 300 frames-per-second. Moving objects will appear much sharper with such a broadcast compared to the standard 50 frames-per-second (not fields). They showed a side by side example, both were 1080 progressive scan. Great for sports broadcasting.
Also Silicon Graphics (when they were called that) have done test with fighter pilots when designing flight simulators. Motion sickness is a problem with those flight simulators, compared to an actual jet plane. When they got a constant frame rate above 80 frames (160 frames per second when doing stereo imaging) per second the motion sickness was greatly reduced. They solved the processing power problem by being able to reduce the rendering resolution on each frame.
Slashdot Mirror
What he is proposing is to reduce the number of transistors on a chip, to increase its speed and reduce power usage.
So in fact he is trying to reverse more's law.
While not on drugs.
I had a case of food poisoning while I was riding on the train to work. While in the train I forgot how to read the news paper, I could read words, but I could not understand sentences anymore. When I got at work I was a little bit better, but I should not have been working, but I was to sick to figure that out myself.
I programmed a tight piece (around 100 lines) of code that converted the data from a cobol structure to xml and back, using a template written in xml.
After two hours a colleague saw me and said I should not be working and he drove me to the hotel.
That piece of code was used for months in production handling a lot of transactions without problem. But when I got back to that piece of code, I could not understand how it could ever work, following the code was hard because of the many recursions and it looked like it should never be able to do what it did. I ended up rewriting it, with a lot of comments, although still using parts of the design I thought up in that state.
So you might be surprised on what you would program when in under influence.
They have a few of those now in amsterdam, as a pedestrian it is very useful to decide if you want to wait for the green light, or just cross against the red light instead.
- If you ever want to write plugin for most applications (c++ plugins are funny and scary).
- work on most operating system kernels (only a few are in c++).
- if you need to know exactly how the language constructs its data structures and functions (c++ is to complex and implementation differs wildly).
You could of course program in C++ with only the C subset, but that is the same thing. And I often need to fall back on the C subset when I work with shared memory (putting instances of classes in shared memory is scary, I now use structs with methods), atomic operations and other weird things. But I also don't shy away of writing little pieces of assembly for special instructions that the compiler did not export.
And there are still quite a few people in the world that use an abacus instead of a calculator. In fact the user interface of an abacus is much faster for certain operations like adding a list of prices at a shop. But it is a lost art in most of the world.
actually, if this is reliable it seems to be much quicker than normal parking.
it may help congestion (from ppl slowly parking, while othe ppl wait) in the streets of amsterdam.
unless the ball is tracking your eyes.
See, if it is a hallucinations in your visual cortex it would move with your eyes.
Same thing as with floaters which you can sometimes see when the light is right, it is like it is moving around the room and you somehow try to follow it with your eyes, while it is actually following your eye movement instead. http://en.wikipedia.org/wiki/Floater
Actually you can buy the mpeg2 codec from apple as well.
http://www.apple.com/quicktime/mpeg2/
It costs $19.99, I guess most of that goes to mpeg2 licensing or something, which is why they are not including it in the main os. Maybe Microsoft has a better deal with them.
IPv6 will use less space in the router tables as the address space is split into a geographical tree. Or at least it used to be when I looked at it the last time, which was a long time ago. Anyway on that scheme the first few bits show the continent, then the country, then the ISP.
That is the biggest problem with Java.
1. To use Java well you need to be a good programmer.
2. A good programmer doesn't want to use Java.
The problems I've seen most programmers make is initializing and configuring objects (like full blown XML parsers, which takes over a ms to initialize and only a few us to use and reset) in the performance critical part of the program. This can be easily solved by creating object on demand, but putting them on a shared stack when you are done with it for reuse by an other thread.
If it is something Alice and Bob are likely to do it is encryption.
You can list AAAA records for a website on the same name.
The browser will use standard library gethostbyname() function call to query the DNS for all IP address listed, including IPv4 and IPv6 addresses all at the same time. Then the browser should prioritize (Safari, IE and Firefox all do) IPv6 addresses over IPv4 addresses and try to connect to the IPv6 address first, and fall back to IPv4 address if the connection failed.
Most software actually already handles IPv6 fine for years. Most software only need around 20 lines of C code extra to handle IPv6 properly.
All websites of google only list AAAA records if your ISP opted-in for AAAA. Normally google's DNS only shows A records, and use magic to show the AAAA record when your source IP address is from one from the opted-in ISPs.
However I believe youtube's content servers always list AAAA records (even if your ISP didn't opt-in), the content servers have a different name from www.youtube.com
As my ISP now did opt in I get with dig -t AAAA www.youtube.com:
www.youtube.com. 6057 IN CNAME youtube-ui.l.google.com.
youtube-ui.l.google.com. 300 IN AAAA 2a00:1450:8001::64
youtube-ui.l.google.com. 300 IN AAAA 2a00:1450:8001::65
youtube-ui.l.google.com. 300 IN AAAA 2a00:1450:8001::66
I am not sure if it works for you, but try (this is a host where the actual contents comes from):
dig -t AAAA v10.lscache4.c.youtube.com
v18.lscache1.c.youtube.com. 6904 IN CNAME v18.lscache1.l.google.com.
v18.lscache1.l.google.com. 266 IN AAAA 2a00:1450:4001:2::15
For web it may be quite easy to have one or more companies set up a ipv4/ipv6 http proxy. As web server owner you only need to point your A record to this proxy. As the webbrowser sends its destination hostname to the proxy, the proxy can look up the AAAA record for this to find your actual web server.
But this is not really a solution for other services or even https.
In any case, there is quite a buzz now with quite a few ISP running "large" scale pilots for native IPv6 to the consumer. Currently these tests are for people who want to try it out. When these tests are done and when ISPs have locked down consumer router configuration (which they do now for Ipv4 as well), they will force it on to every customer.
The chicken and egg problem is beginning to be solved.
1. Youtube delivers the content (not the website, just the movies) over Ipv6 (by default, without provider Ipv6 agreement with google).
2. Apple routers by default set up an IPv6 tunnel. (for their back-to-your-mac system).
3. Computers are configured by default to use IPv6 when available.
These 3 caused an enormous spike in IPv6 Internet traffic since february. This traffic, because of the limited number of tunnel machines available, is not routed via an ideal path. These traffic patterns may cost more in peering fees for the provider, which they can solve by routing native IPv6 (shorter route to the mulihomed network like youtube probably has).
An integrated DNS/NAT solution:
1. customer at 196.0.1.1 asks to "wget www.someone.com"
2. query for www.someone.com to the ISPs DNS/NAT
3. The DNS does a DNS lookup and only receives a AAAA address back. 2001::2345
4. The DNS request from the NAT box a free IPv4 address (on the NAT box) for host 196.0.1.1 to 2001::2345. The NAT replies 10.254.1.1
5. The DNS server returns A 10.254.1.1 as a reply for the www.someone.com request.
Now the NAT only needs to translate Ipv4->ipv6 packets. (196.0.1.1->20.254.1.1) -> 2001::2345
This scheme requires the NAT device to have a lot of IPv4 addresses on the inside (which may be from a private range). But those IPv4 addresses may be shared with multiple customers. I.e. a customer on 196.0.1.2 can also use 10.254.1.1 for an other translation, as the combination of source and destination need to be unique.
It is not pretty, but it probably works.
http://www.xs4all.nl/klant/ipv6/modems.php
AVM FRITZ!box 7270 internationaal [1]
AVM FRITZ!box 7570 VDSL internationaal [1]
Draytek Vigor 2130n icm Vigor 120 [2]
Cisco 876/877 (release 12.4T) [3]
Cisco 886/887 (release 12.4T) [3]
I am one of the customers of XS4ALL who is participating with the native IPv6 pilot.
I myself have the Draytek Vigor 120 + Vigor 2130n combination. The Vigor 120 only does PPPoA to PPPoE translation.
And the Vigor 2130n connects to the PPPoE and does DHCPv6 to retrieve the IPv6 addresses.
From what I understand a modern Airport Extreme or a modern TimeCapsule also can do DHCPv6 and connect over PPPoE through the Vigor 120. But sadly I have an old TimeCapsule which does not allow its firmware to be upgraded.
Both of the draytek 2130 and the time capule has a nice user interface and the only thing to do is select DHCPv6. I am expecting modems from ISPs that will be pre-configured so that the customer doesn't need to know anything about networks.
For the google webservers yes. In fact the youtube website and the google website themselves were (I now have native Ipv6 and my provider seems to have now an agreement with google) still published on IPv4 for me. The media for youtube however is published as IPv6 for everyone.
http://www.networkworld.com/news/2010/020110-youtube-ipv6.html
There are currently two companies forcing the hand of the consumer ISPs to adopt IPv6.
Since February this year Youtube has put all the actual media reachable on IPv6 as default when you access the youtube website through their normal DNS name.
Apple's time capsule and airport extreme by default sets up IPv6 through tunnels.
This means that a lot of people with Apple computers browsing youtube movies are heavy users of IPv6.
As there are only a few tunnel brokers, the load on those will be quite high.
It does actually. As a programmer I can write my application to use only lines, splines, circles and so to design my widgets. there is also an API to convert floating point coordinates to pixel coordinates to align lines to pixel boundaries so that antialiasing does cause a grey double width line.
There is even a tool to change the DPI of the display to check if your application draws pretty.
The point probably is, that there are no high DPI displays to enable the feature for the customer. Which also causes the application writers to ignore pixel boundaries, because DPI changes are not enabled for the customer. Which is maybe why the LCD display manufactures are not creating high DPI displays.
I think at some point, when all the young generation is grown up. They expect to see compromising pictures on your facebook. And if you don't have those they wont trust you and you won't get a job.
Actually, these machines are probably doing "high speed machining" which can only be done when they are software controlled, i.e. it can not be done manually.
In high speed machining, the cutting speed is much higher than is intuitive. As one increases the cutting speed (from zero), the temperature of the tool increases, until it breaks. But there is a sweet spot at a much higher cutting speed where the tool actually is cooled by the cuttings it is removing.
However this sweet spot is tight and it depends on the angle of attack, the depth of the tool, the forward speed, the rotational speed, and the amount of material being removed. And these must be maintained, as the material is entering the material, leaving the material, makes corners, etc.
I've noticed this for a while, it is not just the harddisk.
From the first clone XT PC to a modern DELL computer, the prices haven't changed much at all.
Although one of the thread children says he could get a 1GB drive with a SCSI controller for $1300, he should not forget to compare that to the highend of today, which would be a SAS controller (400 dollar for real RAID (not fake raid)) with SAS disk (500 dollar for a 10,000 rpm).
Ok, ok, 30% drop of price :-)
Intel was not the only one that was shipping 64 bit processors.
DEC Alpha was a 64 bit processor that was manufactured in the mid 1990s. Linux was running on the Alpha processor (native 64 bit) way before the year 1999.
The MIPS R4000 was a 64 bit processor that was manufactured at the beginning of the 1990s. In this case Linux was running on these as well.
The sparc64 was made in the mid 1990s, not sure about the status of linux on those at the time, but I am pretty sure it was ported to these systems as well.
In any case Linux was already running on quite a few processors natively in 64 bit (this includes a complete 64 bit kernel and a 64 bit user-land system, as most of these processors didn't have a 32 bit mode). Therefor it was relatively trivial to port to the new Itanium processors and was running in the labs before the processor was released to the public.
We actually build that model at work, it certainly does OCR autonomic. It scans each character it finds by moving the sensor over the character. Our version needed quite a few tries before it was able to read all the numbers correctly though.
Actually for at least a formula 1 car 20% slip during acceleration is optimal for acceleration, although not for tire wear. From the article below it says that traction control can set the amount of slip during acceleration as ideal as possible. Where ideal may be for either tire wear or acceleration depending on the circumstances of the race.
http://www.atlasf1.com/2000/dec27/shoebotham.html
If you don't play:
10,000 + 50,000 = 60,000
If the result is "yes":
10,000 + 10,000 * 6/1 = 70,000 (original stake + winnings)
If the result is "no":
50,000 + 50,000 * 2/5 = 70,000 (original stake + winnings)
In this case it is always good to bet, because you will always win 10,000.
But, if you follow the hand with your eyes, your hand will appear sharp. You'll be supprised how quickly and stable eyes can track moving objects.
The BBC has been experimenting with fast frame rate TV, running at 300 frames-per-second. Moving objects will appear much sharper with such a broadcast compared to the standard 50 frames-per-second (not fields). They showed a side by side example, both were 1080 progressive scan. Great for sports broadcasting.
Also Silicon Graphics (when they were called that) have done test with fighter pilots when designing flight simulators. Motion sickness is a problem with those flight simulators, compared to an actual jet plane. When they got a constant frame rate above 80 frames (160 frames per second when doing stereo imaging) per second the motion sickness was greatly reduced. They solved the processing power problem by being able to reduce the rendering resolution on each frame.