One thing though, the VCS only had half a scan line of raster data. All the playfield graphics in the register was reversed for a mirror image by default. Many games had to update the register just as the scanline was in the middle of the screen to not have a pong-like play field. In fact some games would change colors or sprite, missile, ball coords during the scanline for effects like having more colors and more objects on screen. The game could more easily do effects like multicolored sprites by changing the color on a vertical retrace, but then the sprite would need to be the same color left to right. Also by moving the sprite during the vertical retrace it could double the sprites, but they would seem to flicker. So it was much more impressive of an effect to change sprites and the register during the raster.
I bought two sets of madcatz adapters that allowed me to charge and plug in head phones at the same time. The most disappointing thing about the SP related to the GBA was that when you played four swords you could not charge. The great thing about it (even the first rev SP) was the backlight and small size.
I had cable circa 6 years ago during a promotional period. I had a motorola cable box with firewire and an iBook with a firewire port. I used the code from the Apple firewire SDK to record shows. Initially I could record everything but HBO (that was the only premium channel I had) to the iBook. I could record all shows to the internal drive in the cable box itself. There was a bug I stumbled upon in the cable box though that when I played a recorded show from the drive in the cable box it would also stream over the firewire port and in this way I could record the show to the iBook.
The reason for this was that the software on the cable box was so poor that the only way I could get it working again after about a month's use was restore (which included a format of all the shows you had recorded) and this was the only manner in which to keep those shows across such a restore.
Then one day I could no longer record anything to the iBook. I called and about a week later they restored the old behavior. Then a few months later they did it again. The only channels I could record to the iBook were the broad cast ones, ie ABC, NBC, CBS, FOX, the local PBS, etc. All of the other basic, extended basic, and digital channels no longer recorded to the iBook. I then found out that this in fact was fine.
Cable needs to have every local broadcast channel. In today's terms that means all full power digital stations in your broadcast area (which in two cases has been smaller than what I actually got with a roof top and in attic antenna) and for those stations only the main sub-channel, 7.1 is required but 7.2 and 7.3 are not. They do not need to be in the same quality as broadcast. If you have a cable box with firewire, only those channels are required to be recordable via firewire.
So then I used the record to cable box drive and play while recording on the iBook work around I had discovered earlier. Then there was a new problem later. Comcast slightly changed the format of the file that was used to record analog cable channels to the cable box drive. So when I played back I could record with the iBook but then VLC was unable to play back the file. I did everything I could think of with transcode and ffmpeg but nothing worked. They did look very much like mpeg2 files but something was not right.
Then when I lost a bunch of Ghibli films not available in the US that I had recorded from TCM when the cable box crashed again I decided that was the end of Comcast for me when the promotion ran out. I installed an antenna in the attic and after that month I was done with them.
They also failed to deliver 12 dB of signal into my house for many months which caused all sorts of problems for my TVs (QAM and analog) and cable internet.
You most likely get your IP via DHCP from your ISP. To disable the DNS hijacking of OpenDNS you need to send them message with your new IP address. There are some tools for this, but it gets annoying when you have a lot of machines that are often asleep unless you have one always up or a router you can put it on.
Also I have seen a smaller version of that Samsung LED set at a BestBuy and I was underwhelmed. I brought "A New Hope" on DVD (the bonus disc version that was largely unmodified) and I played it on a BR player connected to the set. The effect of the LEDs during the initial text scroll and star destroyer scene was unwatchable to me in the Magnolia room even though it was much brighter than my room at home. What happened is that rectangular regions would go black then gray once there were enough stars or noise inside them. It was very distracting and I figured-out in the menus how to largely disable the LED dimming feature on the set, but then it was just like any other set, though everything looked too purple and the white point was too high, which was again not too hard to fix in the menus.
Supposedly the LEDs do give a larger color gamut than CCFLs though I could not notice it in the lighting of that room.
One aspect is that Flash 8 adopted VP6. VP6 runs adequately on punier cpus (single threaded 500-800 Mhz or so little cache) like those used by Android compared to H.264 at similar resolutions and bitrates. Another reason is that going forward the licensing for VP6 is going to be free for google compared to using H.264 (there are currently some rates that are essentially 0 for H.264 for those streaming rather than devices but that is set to expire).
One good thing that may come as a side effect of this is that there was some lawyer writing letters to libavcodec from On2 about patent infringement and what not regarding VP6/7. Maybe google will make some assurances about putting an end to that sort of behavior.
CERN was already sketching out a future upgrade with a long term shutdown for a second run. It may be reasonable to pull-off 4-7 TeV upgrade like you propose then.
You really think that non-abelian Q and you mixing and matching the two inverses and then when Q is not enough you do a Cayley extension to get something not quite O which is not even associative and just barely well defined and mix and match those where ever you please waving your hands saying that's Nature simply because reinterpreting Maxwell's equations in Q was so much fun, that that that well garbage makes more sense than SU3?
You are right, if you are on Bell and you use dig with a different DNS server all is well. But if they really wanted to be jerks they could do what that fellow you responded to was afraid of with out deep packet inspection for 99.9% of those that just hard code some sane DNS server IPs. They already have a firewall, now they just redirect everything to port 53 to another spigot connected to one of their many evil DNS servers. There is no need to rewrite any frames or anything of that sort. That server can even be Windows since everything supports SOCK_RAW which is just one way to not have to worry about correct IPs. It just replies to all ARPs, hey that's me, and returns bogus IPs when NXDOMAIN should have instead.
That is the US propaganda on what happened. The propaganda from the other side is that the Chinese brought a radar system they had in development under the cover of being modern weather forecasting equipment prior to the conflict. Then they tweaked it until the F-117 had no chance with lots of opportunities.
You can believe that one propaganda is more valid if you like, I just know that shortly after that F-117 was downed, NATO bombed the rest of the weather stations. Also one side points to the "accidental" bombing of the Chinese embassy after that as more than a simple accident.
Though you should build what you need the most first. In recent trends the largest proportion of troops maimed and killed are primarily by roadside bombs. Comparatively little has been spent on that, mostly for improved armor. How much has been spent on technology to improve roadside bomb detection? It seems that would be a high tech area, but unfortunately due to the the way the industrial military complex operates very little has gone to that. In fact we are still in the stages of specing what are the reqs.
Another point is that we already have built F-22s, so that addresses your point.
Then there is the point that they really are not all that scary since China has tech to detect the F-22, is working on SAM and A2A tech to down them at long range, there are already tactics to deal with the F-22 that rely on concentrating forces on the tankers and AEWCs that serve the F-22. Finally it is likely that the F-22 is not all that hot anyway and plagued with issues. In addition to the dateline bug, there have been two that crashed for not public reasons in routine low altitude flight (a typical mode of operation to evade RADAR) and the F-22 has yet to see combat.
If Turkey had not already been used, that would have been a more apt nickname for the F-22.
Yeah any revision control system should be fine. Give them accounts on one of your machines. There should be a computer lab in the CS dept. When the students need to work together they should go there and grab two stations next to each other and work from there once they ssh into your box. Other times they can just work from where ever and just email, call, or text each other.
I lived for 5 years in the south side of Chicago without a car. A baby stroller worked great for when there was too much to carry for groceries but usually two or three of us would walk by the grocery store and then we could carry back the bags we brought with us now filled with groceries easily enough. When I needed to move larger items I had a friend who had a car. I was never homeless during that time, but I was a student for part of it:)
If you live and work in certain areas BART is very good in the bay area and the light rail system equally so in LA under the same circumstances. You have to consider not having any of the costs for the car you no longer own into your affordability calculation.
I thought I had heard about a technique where a single camera could use rapidly changing the focal length to produce an image with depth information. The rationale I think was for removing the looking into the screen not the camera aspect of web cams. Does anyone remember anything about that?
I think errata 013 and 014 fixed http://www.openbsd.org/errata31.html were the best examples of critical bugs making it in for at least 2.5 years of releases. Now of course they were not remote exploits so do not count for those two, but they led to the OpenBSD guys to carefully look for other possible signed/unsigned mistakes in a whole lot of code, and of course they were very serious bugs.
I'm saying it is an added complication, I can't just put my font files somewhere on a server that gets accessed when needed, no first I need to run some utility on it that I have to learn about first. This is needed basically only for IE, for reasons it does not actually solve. Before I did any of that all those people that work on web browsers and font libraries needed to implement EOT since places started using it instead of OTF because of IE. I bet you there will be some new wrinkle in five years that makes all of us do something different again. See what I mean now about the frustration with the extra work that does not really accomplish much?
Actually it would be pretty hard to go x86 PEX to x86 ELF as relocations are done completely differently in those formats. They are often names or indexes in DLLs with PEX and handles (think dlopen) while there are all sorts of relocation entries in ELF where the instructions themselves get modified by the run time linker to the correct address or offset (the premunged instruction either has an index in it to a name or another relocation table with more info).
That was a very good high level explanation of how rsync works, thanks. It shows though that this approach is uninteresting to rsync. The hashes and heuristics are so that you do not have to compare the two ends over a potentially slow link, that is the beauty of rsync. In google's case they have all the versions locally and can do the comparisons orders of magnitude faster. So even if there was some way to get binary 1-1 (which I bet they do in fact) it would be useless for rsync as you would have to read the whole file to do the comparison, so why not just send the whole file instead of changed to begin with?
The other point is that when google says assembler they do not mean something like MASM or gas, but rather something that can take two binary blobs (one that you have and one that they send) and run them to generate a third binary blob that matches what they intended the result to be after encapsulating it again. I think you get that but I wanted to make that clear to other readers.
One thing is very different from back then and now, disk space was at a premium, so there were a ton of optimizations to code fragments in the object file format to that end. For example most data was zero, so only portions that were not were in the object file bunched-up. The utility you refer to was able to take advantage of this. It was otherwise a simple ed-like that could insert, remove, and replace. I wish I could remember the name.
Yes and it is a generic approach that does not need to know anything about pex, x86, and archives. In fact it is used by this google implementation. And there was something much like what google has here implemented back in 1999 cited in the paper you posted the link to:
B.S. Baker, U. Manber, and R. Muth, Compressing Differences of Executable Code, ACM SIGPLAN Workshop on Compiler Support for System Software, 1999.
Slashdot Mirror
Ha! yes texture mapping like that would be fun.
One thing though, the VCS only had half a scan line of raster data. All the playfield graphics in the register was reversed for a mirror image by default. Many games had to update the register just as the scanline was in the middle of the screen to not have a pong-like play field. In fact some games would change colors or sprite, missile, ball coords during the scanline for effects like having more colors and more objects on screen. The game could more easily do effects like multicolored sprites by changing the color on a vertical retrace, but then the sprite would need to be the same color left to right. Also by moving the sprite during the vertical retrace it could double the sprites, but they would seem to flicker. So it was much more impressive of an effect to change sprites and the register during the raster.
Yes but I only ever saw it on their online store. They still have them in fact:
http://www.nintendo.com/consumer/systems/gameboyadvance/accessories.jsp
I bought two sets of madcatz adapters that allowed me to charge and plug in head phones at the same time. The most disappointing thing about the SP related to the GBA was that when you played four swords you could not charge. The great thing about it (even the first rev SP) was the backlight and small size.
How have you not had to replace some bulging caps?
It is much more restrictive in fact, read on:
I had cable circa 6 years ago during a promotional period. I had a motorola cable box with firewire and an iBook with a firewire port. I used the code from the Apple firewire SDK to record shows. Initially I could record everything but HBO (that was the only premium channel I had) to the iBook. I could record all shows to the internal drive in the cable box itself. There was a bug I stumbled upon in the cable box though that when I played a recorded show from the drive in the cable box it would also stream over the firewire port and in this way I could record the show to the iBook.
The reason for this was that the software on the cable box was so poor that the only way I could get it working again after about a month's use was restore (which included a format of all the shows you had recorded) and this was the only manner in which to keep those shows across such a restore.
Then one day I could no longer record anything to the iBook. I called and about a week later they restored the old behavior. Then a few months later they did it again. The only channels I could record to the iBook were the broad cast ones, ie ABC, NBC, CBS, FOX, the local PBS, etc. All of the other basic, extended basic, and digital channels no longer recorded to the iBook. I then found out that this in fact was fine.
Cable needs to have every local broadcast channel. In today's terms that means all full power digital stations in your broadcast area (which in two cases has been smaller than what I actually got with a roof top and in attic antenna) and for those stations only the main sub-channel, 7.1 is required but 7.2 and 7.3 are not. They do not need to be in the same quality as broadcast. If you have a cable box with firewire, only those channels are required to be recordable via firewire.
So then I used the record to cable box drive and play while recording on the iBook work around I had discovered earlier. Then there was a new problem later. Comcast slightly changed the format of the file that was used to record analog cable channels to the cable box drive. So when I played back I could record with the iBook but then VLC was unable to play back the file. I did everything I could think of with transcode and ffmpeg but nothing worked. They did look very much like mpeg2 files but something was not right.
Then when I lost a bunch of Ghibli films not available in the US that I had recorded from TCM when the cable box crashed again I decided that was the end of Comcast for me when the promotion ran out. I installed an antenna in the attic and after that month I was done with them.
They also failed to deliver 12 dB of signal into my house for many months which caused all sorts of problems for my TVs (QAM and analog) and cable internet.
You most likely get your IP via DHCP from your ISP. To disable the DNS hijacking of OpenDNS you need to send them message with your new IP address. There are some tools for this, but it gets annoying when you have a lot of machines that are often asleep unless you have one always up or a router you can put it on.
Bingo
Also I have seen a smaller version of that Samsung LED set at a BestBuy and I was underwhelmed. I brought "A New Hope" on DVD (the bonus disc version that was largely unmodified) and I played it on a BR player connected to the set. The effect of the LEDs during the initial text scroll and star destroyer scene was unwatchable to me in the Magnolia room even though it was much brighter than my room at home. What happened is that rectangular regions would go black then gray once there were enough stars or noise inside them. It was very distracting and I figured-out in the menus how to largely disable the LED dimming feature on the set, but then it was just like any other set, though everything looked too purple and the white point was too high, which was again not too hard to fix in the menus.
Supposedly the LEDs do give a larger color gamut than CCFLs though I could not notice it in the lighting of that room.
One aspect is that Flash 8 adopted VP6. VP6 runs adequately on punier cpus (single threaded 500-800 Mhz or so little cache) like those used by Android compared to H.264 at similar resolutions and bitrates. Another reason is that going forward the licensing for VP6 is going to be free for google compared to using H.264 (there are currently some rates that are essentially 0 for H.264 for those streaming rather than devices but that is set to expire).
One good thing that may come as a side effect of this is that there was some lawyer writing letters to libavcodec from On2 about patent infringement and what not regarding VP6/7. Maybe google will make some assurances about putting an end to that sort of behavior.
sed has been around for more than 20 years though and I bet a couple of one liners can fix most anything right up.
CERN was already sketching out a future upgrade with a long term shutdown for a second run. It may be reasonable to pull-off 4-7 TeV upgrade like you propose then.
WTF?
You really think that non-abelian Q and you mixing and matching the two inverses and then when Q is not enough you do a Cayley extension to get something not quite O which is not even associative and just barely well defined and mix and match those where ever you please waving your hands saying that's Nature simply because reinterpreting Maxwell's equations in Q was so much fun, that that that well garbage makes more sense than SU3?
Take your meds.
You are right, if you are on Bell and you use dig with a different DNS server all is well. But if they really wanted to be jerks they could do what that fellow you responded to was afraid of with out deep packet inspection for 99.9% of those that just hard code some sane DNS server IPs. They already have a firewall, now they just redirect everything to port 53 to another spigot connected to one of their many evil DNS servers. There is no need to rewrite any frames or anything of that sort. That server can even be Windows since everything supports SOCK_RAW which is just one way to not have to worry about correct IPs. It just replies to all ARPs, hey that's me, and returns bogus IPs when NXDOMAIN should have instead.
That is the US propaganda on what happened. The propaganda from the other side is that the Chinese brought a radar system they had in development under the cover of being modern weather forecasting equipment prior to the conflict. Then they tweaked it until the F-117 had no chance with lots of opportunities.
You can believe that one propaganda is more valid if you like, I just know that shortly after that F-117 was downed, NATO bombed the rest of the weather stations. Also one side points to the "accidental" bombing of the Chinese embassy after that as more than a simple accident.
Though you should build what you need the most first. In recent trends the largest proportion of troops maimed and killed are primarily by roadside bombs. Comparatively little has been spent on that, mostly for improved armor. How much has been spent on technology to improve roadside bomb detection? It seems that would be a high tech area, but unfortunately due to the the way the industrial military complex operates very little has gone to that. In fact we are still in the stages of specing what are the reqs.
Another point is that we already have built F-22s, so that addresses your point.
Then there is the point that they really are not all that scary since China has tech to detect the F-22, is working on SAM and A2A tech to down them at long range, there are already tactics to deal with the F-22 that rely on concentrating forces on the tankers and AEWCs that serve the F-22. Finally it is likely that the F-22 is not all that hot anyway and plagued with issues. In addition to the dateline bug, there have been two that crashed for not public reasons in routine low altitude flight (a typical mode of operation to evade RADAR) and the F-22 has yet to see combat.
If Turkey had not already been used, that would have been a more apt nickname for the F-22.
Yeah any revision control system should be fine. Give them accounts on one of your machines. There should be a computer lab in the CS dept. When the students need to work together they should go there and grab two stations next to each other and work from there once they ssh into your box. Other times they can just work from where ever and just email, call, or text each other.
I lived for 5 years in the south side of Chicago without a car. A baby stroller worked great for when there was too much to carry for groceries but usually two or three of us would walk by the grocery store and then we could carry back the bags we brought with us now filled with groceries easily enough. When I needed to move larger items I had a friend who had a car. I was never homeless during that time, but I was a student for part of it :)
Somehow grandma ends-up buy incredible amounts of fuel then.
If you live and work in certain areas BART is very good in the bay area and the light rail system equally so in LA under the same circumstances. You have to consider not having any of the costs for the car you no longer own into your affordability calculation.
I thought I had heard about a technique where a single camera could use rapidly changing the focal length to produce an image with depth information. The rationale I think was for removing the looking into the screen not the camera aspect of web cams. Does anyone remember anything about that?
I think errata 013 and 014 fixed http://www.openbsd.org/errata31.html were the best examples of critical bugs making it in for at least 2.5 years of releases. Now of course they were not remote exploits so do not count for those two, but they led to the OpenBSD guys to carefully look for other possible signed/unsigned mistakes in a whole lot of code, and of course they were very serious bugs.
I'm saying it is an added complication, I can't just put my font files somewhere on a server that gets accessed when needed, no first I need to run some utility on it that I have to learn about first. This is needed basically only for IE, for reasons it does not actually solve. Before I did any of that all those people that work on web browsers and font libraries needed to implement EOT since places started using it instead of OTF because of IE. I bet you there will be some new wrinkle in five years that makes all of us do something different again. See what I mean now about the frustration with the extra work that does not really accomplish much?
Courgette uses bsdiff as well, after it breaks down into some parts that bsdiff better.
Actually it would be pretty hard to go x86 PEX to x86 ELF as relocations are done completely differently in those formats. They are often names or indexes in DLLs with PEX and handles (think dlopen) while there are all sorts of relocation entries in ELF where the instructions themselves get modified by the run time linker to the correct address or offset (the premunged instruction either has an index in it to a name or another relocation table with more info).
That was a very good high level explanation of how rsync works, thanks. It shows though that this approach is uninteresting to rsync. The hashes and heuristics are so that you do not have to compare the two ends over a potentially slow link, that is the beauty of rsync. In google's case they have all the versions locally and can do the comparisons orders of magnitude faster. So even if there was some way to get binary 1-1 (which I bet they do in fact) it would be useless for rsync as you would have to read the whole file to do the comparison, so why not just send the whole file instead of changed to begin with?
The other point is that when google says assembler they do not mean something like MASM or gas, but rather something that can take two binary blobs (one that you have and one that they send) and run them to generate a third binary blob that matches what they intended the result to be after encapsulating it again. I think you get that but I wanted to make that clear to other readers.
One thing is very different from back then and now, disk space was at a premium, so there were a ton of optimizations to code fragments in the object file format to that end. For example most data was zero, so only portions that were not were in the object file bunched-up. The utility you refer to was able to take advantage of this. It was otherwise a simple ed-like that could insert, remove, and replace. I wish I could remember the name.
Yes and it is a generic approach that does not need to know anything about pex, x86, and archives. In fact it is used by this google implementation. And there was something much like what google has here implemented back in 1999 cited in the paper you posted the link to:
B.S. Baker, U. Manber, and R. Muth, Compressing Differences of Executable Code, ACM SIGPLAN Workshop on Compiler Support for System Software, 1999.