I could possibly, but not likely, see it happening if two failures happened at the same time. Lack of ground to the IDE drive through it's power connector[1], and hooking up the IDE cable wrong. It would still likely not happen. My guess is you dropped something on the motherboard that shorted the +12DC supply to the CPU's Supply or knocked off the voltage control jumpers[2] for the CPU voltage(s).
[1] The pin/socket style used on most HD power connectors isn't that good. It tends to get weaker with every use and fails after only a few hundred cycles. It will also fail faster if it isn't aligned right and you force it. This could cause the ground and +5VDC lines to not connect up properly. Thus the drive would be biased at +12V.
[2] Removing, knocking off the voltage control jumpers or putting them into an undefined state may make the switching power supply that feeds the caps to go over voltage. This could also destroy the cpas.
Yep, you have to rethink alot. It's possible. I expect we'll see async processors first show up in embedded situations where all parts of the system are integrated on one chip.
Busses can be made asynchronous. Handshaking is the key. New statigies will be needed, but people are bright so I feel they will be developed. With a little thinking I've sketched out a packet type asyncronous bus in my head. It would work nicely for up to a meter or so. Longer lengths would be slower than shorter ones. One thing I feel may work best is for any signal/data that needs to travel significant distances is to then go into synchronous transmition. Otherwise you end up adding in delays from the back handshake signals.
I remember some of the first articles in SIGARCH and how they sparked my interest. I've always felt that async was the way to go when you don't know how long an operation will take. I'm happy to see it's still getting research dollars.
These comments are spot on. I'm also using small MCU devices for anything that needs hard real time CPU responce. That is motor control, touch sensors, etc. They are cheep and you can easily dedicate them to the task. You can use RS-232 links and have a serial port for each, or you can use RS-485 and use common links. I have two RS-485 links. One for the motors and the other is for touch sensors and slow update devices like compases. I also have a RS-232 set aside for GPS in the future. My central processor is currently only a 486 with Linux loaded. Rather slow, but it's fast enough. I have plans to replace it with something really fast so I can start to do vision work. But that won't happen for awhile. Standard PC hardware isn't quite fast enough yet.
With the PIC CPUs, you can get cheep C compilers. The C2C compiler is horrible, but if you combine it with GCC's C preprocessor it can be made passable, but just barely. The guy who wrote it needs to get himself a copy of K&R and really read and follow it. I do all my development under Linux and download code to the PIC using the PICSTART programmer. With another program (picp) I can use the PICSTART programmer under Linux. This allows me to do all my development under Linux.
For you host processor and motherboard you can go cheep and use and old PC, but then you have to have a huge robot. Otherwise there are many PC compatible single board computers out there. You can even find then surplus. Buying new you should be able to stay under $500 easily and still get a pentium class CPU. The main thing when selecting a board ig getting enough memory. You want atleast 16MBytes due to some of the current install scripts with the available distributions. I only have 8MBytes on my robot, but I loaded linux by putting the HD in another system and only copying the parts I wanted.
Generally yes, though some are using prizmatic cell shapes but they are more costly than the good ol' cylindrical cell. If you go with a non standard cell size or shape it costs more per unit due to lack of volume. The standard cell sizes are also available from many different manufacturers. That means you have second and third sources available in case your manufacturer of choice fails to meet your demand for some reason.
First backup the hole system as is and keep a copy for yourself. If that isn't possible, backup the configuration files and any data files you can fit.
Tripwire is your friend. Run tripwire on all files (even ones known to change). Save the tripwire file on both the system and keep a backup copy. When you get a support call you can use this to check what the guys have changed. For the most part you can run tripwire without checking checksums, just length, data, perms, etc. This will give you a list of new, changed, and deleted files. Not doing the chacksums lessens some of the utility of tripwire, but it gets you a list fast as tripwire dosen't have to read the file in.
Security tool can be used for your bennifit. They aren't just for security. When administering systems for developers who have root I always use tripwire on their systems. Often it tells me what they are changing so I can keep on top of their needs.
What if we start designing things such that each program gets its own IP address?
I doubt you will have that many programs running on your systems. Even at a company level.
With each/64 subnet having a full 64bits for specific machine identification you could easily assign machine addresses randomly and not really worry about collisions. You're talking about a huge address space. So what you assign machines a few billion address so they can assign one to each program. Current process tables are only measured in the thousands. It's a non issue at this point. I can see a senario where it could be an issue, but then having more than 2^64 objects is rather unlikely.
Yes you can get it setup for cheep, but only if your willing to do some hardware hacking as well as possibly some software hacking. My sugestion is go with the wireless lan. For a two PC setup I'm told you don't need the access point. That brings the price down to a reasonable level.
Digikey sells the Linx RF modules which cost $200-$250 for the evaluation kits with all the parts you should need. That's about as cheep as you will get. Note you will need some more hardware linke cases and power supplies. The EV kits just provide 9 volt batteries. Works for testing but not for long term use. After you get the physical link working, now you need to get the software working. That means more effort. Oh, as an after thought. If you are doing this for yourself (not a company) you need to get yourself a HAM license to muck about with RF devices. Otherwise if it's for work, go get only FCC approved devices. Dealing with the FCC regs will quickly cost more than the fancyest gear on the market.
This is typical of the media, a lengthy article about something that is reallly quite trivial.
Actually not so trivial. The article isn't really about the kid. He's the example. It's about self and roles, how we generate them, and how others perceive them. It's about the changes that technology and the internet brought about to the process of how roles are built and validated. Please go back and reread the article and ignore the kid. The author focused way to much time on him and it draws the readership off of the main point. He should have sought out other examples. He also should have had a better, more direct ending. An early paragraph puts it well:
"What was happening on the Internet buttressed a school of thought in sociology known as role theory. The role theorists argue that we have no "self" as such. Our selves are merely the masks we wear in response to the social situations in which we find ourselves. The Internet had offered up a new set of social situations, to which people had responded by grabbing for a new set of masks. People take on the new tools they are ready for and make use of only what they need, how they need it. If they were using the Internet to experiment with their identities, it was probably
because they found their old identities inadequate. If the Internet was giving the world a shove in a certain direction, it was probably because the world already felt inclined to move in that direction. The Internet was telling us what we wanted to become."
For the record, underground utilities actually have more problems with lightning. When lightning strikes a spot, it goes through the ground for some distance. As such it usually finds the underground wiring.
Unfortunately you can't always unplug the computers as you are away from home or running servers on them.
I run a paranoid power system. Everything goes through the sacrificial surge protector. After that is the TripLite line conditioner. Then there is a UPS for each computer. Computers are either directly pluged into the UPS or have a surge supressor between them and the UPS. The phone line for the DSL has it's own surge supressor too. After my appartment building was hit I didn't have to replace any of my electronics, but I did replace the sacrificial surge supressor and line conditioner. Other people in the building were replacing everthing.
My guess is they haven't figured out how to place different colored dye capsules precise enough on the paper. Getting the alignment between the wires, and capsules could be tricky, but doable. They may also be trying to come up with a truely multi color capsel. That is one with all three colors in it. That would be a control problem. I can't see a way to control it. The filter method is hokey in my mind, but it works.
I only have a Xilinx CPLD programmer on my parallel port. On the other hand my serial ports regulairly get to talk to my Parallax SX Blitz and SX Key, a Microchip PICSTART programmer, and a computerised timing gate for recording the velocity of the balls my coilgun shoots. Another one of my computers talks to a couple of motor controllers and a few touch sensors.
Get an oversized Alumimun case and mount many fans inside it blowing every which way but out. Seal the case up tight. On the outside, have a bunch of fans blowing air over the sides and top of the case. Some friends are using this on a computer that is outside. The case internal air temp runs about 10 degrees above ambient. The CPU they have is only a K6-2-350. If you need more cooling, you can mount large area AL heatsinks on both sides of a case wall and direct fans over them. One thing to remember is cooling air wants to go down and air warming up wants to go up. Take this into account when placing fans otherwise it will take alot more fans to get the job done. www.Aavid.com has many large area heatsinks. Some large enough to serve as a side wall for the case.
Alumunum is a very good heat conductor. All the better heat conductors cost much much more to use. Look into using a copper CPU heatsink to better transfer heat away from it, but check the overclocker sites to really find out which heatsink is the best bang for the $$$ you are willing to spend.
I'm looking at having to make a PC for outside use in the near term future. I'm going to make it's case out of large area AL heatsinks. I'll be using a heat pipe to transfer the CPU's heat to one side, and use a dual sided heatsink on the other side to cool the rest of the components. Items like hard disks will be directly mounted to the side walls to provide better heat conduction. The idea is I want to make it so it doesn't need fans at all. It needs to run off of solar and batteries only.
Have they solved the problems with blue OLEDs having a much reduced life expectancy. Last numbers I heard were on the orders of 1000-2000 hours use before fading to much.
Money is always in issue. On interviews I mention that shirt and tie = $5,000 more for me, Add a jacket or suit, that's another $5k a year. Dressing up all the time costs money. Money I could better spend on electronics parts for my robots and coil guns.
OpenBSD allows you to encrypt the swap disk if you so desire. Not perfect, but better. After all your data is still unencrypted in memory. Any root process could peak at it.
In my opinion the HW crypto card is there to both speed things up and provide secure key space for inuse keys. Despite popular opinion keys are easy to extract from a running program given the memory map of the program. just look for areas of high randomness. They will likely be word bounded, and have pointers pointing to their starting position. With a hardware card you can put the key on the card and erase it from memory. A good hardware crypto card has it's key space write only. Better ones the input buffer space is also write only.
With tape drives you get what you pay for. The cheep ones are just that, cheep. Go for the good quality drives, but they will set you back a few grand. On the otherhand they will operate flawlessly for years. I use a midrange DAT here at home, but it's old and dated. I bought it over a decade ago. I know I can read back all my tapes as I do a verify read after writing the tape out.
Now for doing cheep backups. Go with IDE disks in removable carriers. As another said, make a backup system. You don't even need to give it a monitor. Log into it over your local net. Get a bunch of 70+Gig hard disks and put them in carriers. Plug two in and backup to them, then remove then and store one off site.
You mentioned surge dammage. Where are your surge suppressors? UPS? Line conditioner? A transformer style line conditioner does alot for filter out spikes just by it's design. The power for my computers first goes into a line conditioner, then through the UPSes and finally power strips. Each stage has surge absorbers. Kinda a defence in depth. Kinda seams like overkill, but then I haven't lost computers when other in the building had all their electronics toasted. Yes I replaced the power equipment after that surge. The MOVs in the line conditioner let the magic smoke out.
Slashdot Mirror
[1] The pin/socket style used on most HD power connectors isn't that good. It tends to get weaker with every use and fails after only a few hundred cycles. It will also fail faster if it isn't aligned right and you force it. This could cause the ground and +5VDC lines to not connect up properly. Thus the drive would be biased at +12V.
[2] Removing, knocking off the voltage control jumpers or putting them into an undefined state may make the switching power supply that feeds the caps to go over voltage. This could also destroy the cpas.
Maby it because the last one ported to Linux kindof fizzled. I didn't even get one copy sent to me. I feel so left out. :(
I'd say keep a swap space just incase you run out of memory. Linux dosen't handle running out of memory gracefully.
Busses can be made asynchronous. Handshaking is the key. New statigies will be needed, but people are bright so I feel they will be developed. With a little thinking I've sketched out a packet type asyncronous bus in my head. It would work nicely for up to a meter or so. Longer lengths would be slower than shorter ones. One thing I feel may work best is for any signal/data that needs to travel significant distances is to then go into synchronous transmition. Otherwise you end up adding in delays from the back handshake signals.
I remember some of the first articles in SIGARCH and how they sparked my interest. I've always felt that async was the way to go when you don't know how long an operation will take. I'm happy to see it's still getting research dollars.
With the PIC CPUs, you can get cheep C compilers. The C2C compiler is horrible, but if you combine it with GCC's C preprocessor it can be made passable, but just barely. The guy who wrote it needs to get himself a copy of K&R and really read and follow it. I do all my development under Linux and download code to the PIC using the PICSTART programmer. With another program (picp) I can use the PICSTART programmer under Linux. This allows me to do all my development under Linux.
For you host processor and motherboard you can go cheep and use and old PC, but then you have to have a huge robot. Otherwise there are many PC compatible single board computers out there. You can even find then surplus. Buying new you should be able to stay under $500 easily and still get a pentium class CPU. The main thing when selecting a board ig getting enough memory. You want atleast 16MBytes due to some of the current install scripts with the available distributions. I only have 8MBytes on my robot, but I loaded linux by putting the HD in another system and only copying the parts I wanted.
Generally yes, though some are using prizmatic cell shapes but they are more costly than the good ol' cylindrical cell. If you go with a non standard cell size or shape it costs more per unit due to lack of volume. The standard cell sizes are also available from many different manufacturers. That means you have second and third sources available in case your manufacturer of choice fails to meet your demand for some reason.
Enabling process accounting only gets you what commands were run. You don't get the arguments or actions taken within those commands.
First backup the hole system as is and keep a copy for yourself. If that isn't possible, backup the configuration files and any data files you can fit.
Tripwire is your friend. Run tripwire on all files (even ones known to change). Save the tripwire file on both the system and keep a backup copy. When you get a support call you can use this to check what the guys have changed. For the most part you can run tripwire without checking checksums, just length, data, perms, etc. This will give you a list of new, changed, and deleted files. Not doing the chacksums lessens some of the utility of tripwire, but it gets you a list fast as tripwire dosen't have to read the file in.
Security tool can be used for your bennifit. They aren't just for security. When administering systems for developers who have root I always use tripwire on their systems. Often it tells me what they are changing so I can keep on top of their needs.
I rarely use C's or C++'s overflowable library routines. If I do it's only in a quick hack. One dosen't need to use the standard library routines.
"The BSA is targeting NYC".
If I know my typical, jaded NYC dwellers, I sincerely doubt that they'd be cowed into software subservience by the Boy Scouts of America.
Silly boy, don't you know its really the "Boy Sprouts of America".
I doubt you will have that many programs running on your systems. Even at a company level.
With each /64 subnet having a full 64bits for specific machine identification you could easily assign machine addresses randomly and not really worry about collisions. You're talking about a huge address space. So what you assign machines a few billion address so they can assign one to each program. Current process tables are only measured in the thousands. It's a non issue at this point. I can see a senario where it could be an issue, but then having more than 2^64 objects is rather unlikely.
Third generation beer, yuck!
Digikey sells the Linx RF modules which cost $200-$250 for the evaluation kits with all the parts you should need. That's about as cheep as you will get. Note you will need some more hardware linke cases and power supplies. The EV kits just provide 9 volt batteries. Works for testing but not for long term use. After you get the physical link working, now you need to get the software working. That means more effort. Oh, as an after thought. If you are doing this for yourself (not a company) you need to get yourself a HAM license to muck about with RF devices. Otherwise if it's for work, go get only FCC approved devices. Dealing with the FCC regs will quickly cost more than the fancyest gear on the market.
Set your routers to block all traffic to port 80 for your printers.
Actually not so trivial. The article isn't really about the kid. He's the example. It's about self and roles, how we generate them, and how others perceive them. It's about the changes that technology and the internet brought about to the process of how roles are built and validated. Please go back and reread the article and ignore the kid. The author focused way to much time on him and it draws the readership off of the main point. He should have sought out other examples. He also should have had a better, more direct ending. An early paragraph puts it well:
"What was happening on the Internet buttressed a school of thought in sociology known as role theory. The role theorists argue that we have no "self" as such. Our selves are merely the masks we wear in response to the social situations in which we find ourselves. The Internet had offered up a new set of social situations, to which people had responded by grabbing for a new set of masks. People take on the new tools they are ready for and make use of only what they need, how they need it. If they were using the Internet to experiment with their identities, it was probably because they found their old identities inadequate. If the Internet was giving the world a shove in a certain direction, it was probably because the world already felt inclined to move in that direction. The Internet was telling us what we wanted to become."
Personally I like C because I can shoot myself in the foot faster and with less effort with it.
For the record, underground utilities actually have more problems with lightning. When lightning strikes a spot, it goes through the ground for some distance. As such it usually finds the underground wiring.
I run a paranoid power system. Everything goes through the sacrificial surge protector. After that is the TripLite line conditioner. Then there is a UPS for each computer. Computers are either directly pluged into the UPS or have a surge supressor between them and the UPS. The phone line for the DSL has it's own surge supressor too. After my appartment building was hit I didn't have to replace any of my electronics, but I did replace the sacrificial surge supressor and line conditioner. Other people in the building were replacing everthing.
My guess is they haven't figured out how to place different colored dye capsules precise enough on the paper. Getting the alignment between the wires, and capsules could be tricky, but doable. They may also be trying to come up with a truely multi color capsel. That is one with all three colors in it. That would be a control problem. I can't see a way to control it. The filter method is hokey in my mind, but it works.
I only have a Xilinx CPLD programmer on my parallel port. On the other hand my serial ports regulairly get to talk to my Parallax SX Blitz and SX Key, a Microchip PICSTART programmer, and a computerised timing gate for recording the velocity of the balls my coilgun shoots. Another one of my computers talks to a couple of motor controllers and a few touch sensors.
Alumunum is a very good heat conductor. All the better heat conductors cost much much more to use. Look into using a copper CPU heatsink to better transfer heat away from it, but check the overclocker sites to really find out which heatsink is the best bang for the $$$ you are willing to spend.
I'm looking at having to make a PC for outside use in the near term future. I'm going to make it's case out of large area AL heatsinks. I'll be using a heat pipe to transfer the CPU's heat to one side, and use a dual sided heatsink on the other side to cool the rest of the components. Items like hard disks will be directly mounted to the side walls to provide better heat conduction. The idea is I want to make it so it doesn't need fans at all. It needs to run off of solar and batteries only.
Have they solved the problems with blue OLEDs having a much reduced life expectancy. Last numbers I heard were on the orders of 1000-2000 hours use before fading to much.
Money is always in issue. On interviews I mention that shirt and tie = $5,000 more for me, Add a jacket or suit, that's another $5k a year. Dressing up all the time costs money. Money I could better spend on electronics parts for my robots and coil guns.
In my opinion the HW crypto card is there to both speed things up and provide secure key space for inuse keys. Despite popular opinion keys are easy to extract from a running program given the memory map of the program. just look for areas of high randomness. They will likely be word bounded, and have pointers pointing to their starting position. With a hardware card you can put the key on the card and erase it from memory. A good hardware crypto card has it's key space write only. Better ones the input buffer space is also write only.
Now for doing cheep backups. Go with IDE disks in removable carriers. As another said, make a backup system. You don't even need to give it a monitor. Log into it over your local net. Get a bunch of 70+Gig hard disks and put them in carriers. Plug two in and backup to them, then remove then and store one off site.
You mentioned surge dammage. Where are your surge suppressors? UPS? Line conditioner? A transformer style line conditioner does alot for filter out spikes just by it's design. The power for my computers first goes into a line conditioner, then through the UPSes and finally power strips. Each stage has surge absorbers. Kinda a defence in depth. Kinda seams like overkill, but then I haven't lost computers when other in the building had all their electronics toasted. Yes I replaced the power equipment after that surge. The MOVs in the line conditioner let the magic smoke out.