The list is fine as far as it goes, but all of it refers to commercial hardware. Some of us are working on easily customized reference designs for linux based controllers and micro systems.
At http://www.openh.org/ there are several ongoing linux controller projects released under free licenses.
At http://freeio.org/ we are developing and posting frequent updates to linux based controllers based on the ColdFire processor, released under GPL. The conceprt is free hardware designs for the free software community.
For those interested in rolling their own systems, or better yet contributing to free hardware designs and porting / building drivers, these are pretty good opportunities.
Anyone with sensitive data of any kind runs into the question of how to make it gone on demand. We all know ways of starting over on a drive to recycle it - load your favorite GNU/Linux or BSD distribution on it from scratch and enjoy.
However, the real question is this: if someone really wants to do so, how much data can be retrieved from a drive which has been reformatted and written over. Sure, most of us are going to give up fairly quickly, but if one puts enough effort into it, what data is still there waiting between the tracks? Unfortunately, at least some of the stuff from Tom Clancey novels is based on fact, and if you have the budget of a major government, and enough time, at least some data can sometimes be retrieved. If there are secrets which are worth enough to spend major time and money recovering, no data is ever quite gone.
Or at least, that is what those fabled government agencies base their budget requests upon.
When your company develops something under a contract, as an employee of the employer you are bound by that contract, as well as any government regulations regarding contracts. Depending upon the circumstances and the way the contract was written, the company may or may not retain copyright to the work. One common variable is whether the work was done under set-asides for small business, in which case the government frequently allows the copyright to remain with the company, and the government gets an unrestricted license to use the sofware which was developed. If the company has the rights to the copyright, then it can GPL the product at will.
You will have to have the company counsel read both the contract and the federal acquisition regulations (FARs) to determine how to proceed. We were in such a battle two years ago, and in our case it turned out that the copyright was ours to do with whatever we chose, as long as the government had the right to use what we developed.
Note also that FSF has a major section in their explanation of GPL dealing with the dangers of trying to GPL something to which you do not have the rights. There is danger here if you don't get it right.
Trademarking a commonly used text string has been done before, and has been solved before.
More than a decade ago, we saw this with SEA trying to enforce the.arc extension for archived files being trademarked. The result was Phil Katz releasing the pkzip utilities, and the whole community switched over to zip files within a few months. SEA had made itself odious to the user community, and was cut off.
I currently have two hardware projects on openh.org, and am currently working on a third one. All three are meant to make life easier for the person who needs to develop hardware that runs software, free or not.
The first two devices are ISA bus I/O cards for control of hardware devices. In this case, the cards themselves are programmable (in the VHDL language) to allow the hardware functions to be changed, without ever having to remove the I/O card from the computer. You want parallel I/O?... no problem, can do. You want serial I/O? -...reprogram it and it will do it. You need some arbitrary function? You can do that too. I provide the full specification of how it works, plus the CAD files for those who are interested, along with the manufacturing files and sample VHDL code to show how to use it as well - all under GPL.
The current project is a GPLed control board using a Motorola ColdFire processor. Once again, all of the data is available right there on the openh.org site. The goal is to make it run the uClinux code with no customization actually required. This will be fully free software running on fully free hardware.
Now my projects are still available on my own site http://FreeIO.org/ but when Alan at the University of Cape Town approached me about hosting my projects on the openH.org site, I jumped at the opportunity. Just as with Linux or any software project which is developed in the open, there is the opportunity for others to benefit, and their feedback will help improve the project.
Contrary to the statements of some, developing hardware is not really harder than developing software, although the tools are different. Open source tools are not there yet, but free (beer) tools are available which work on smaller projects rather well. An example is the freeware version of the Eagle electronic CAD system http://www.cadsoft.de/ which handles schematic and printed wiring board designs very well, and runs natively under Linux. You do have to learn it to use it, but the same can be said or C/perl/python or whatever.
Several years ago I worked for an "engineering service contractor" for NASA, and had the opportunity to design the sensor electronics for a space-based sun-staring telescope, called Solar X-ray Imager. The task would not have been excessively difficult except that the radiation levels expected were quite high, and we had to design it to be very radiation resistant. This proved to be very expensive to do.
For certain electronic functions, there are radiation hardened integrated circuits. We used a set of CMOS devices manufactured by Harris (now Intersil) which are manufactured using a silicon-on-sapphire technology which makes them tough indeed. (The manufacturer claimed that they were radiation resistant up to "strategic levels" - to run through massive radiation doses and never glitch...) While they were very good parts, their cost was incredible. They were getting US$ 225 for a single 75hc00 equivalent device which usually costs 25 cents. Processors and memories had prices that were proportionally worse. Nobody but government bodies with the power of taxation can afford these things.
The other method of achieving radiation hardness is shielding. In our case, we had to use this in addition to the radiation hardened integrated circuits, because the telescope sensor itself was not radiation hardened (otherwise it would not have been able to see light either!) The shielding we used was made from tantalum, machined in shapes to cover the sensitive parts. Now tantalum is not inexpensive, it does not machine well, and is heavy. The weight is the limiting factor, since whatever you use you have to launch into space, and the cost per kg is very high. Shielding which is sufficient is also too heavy to launch on a secondary payload.
My thoughts? Amsat did the best they could within the budget, and the budget (raised from donations) did not provide enough money to use the premium radiation hardened parts everywhere.
When measurements of polar ice thickness are plotted since 1950, we usually assume that there are no local events which could cause such a thing. However, during that period of time there have been increasing numbers of intentional cases of breaking-up of the polar ice, in order to navigate the surface of the polar region.
Recently there were photos of a commercial expedition, using a Russian nuclear-power ice breaker, for those scientists, journalists, and tourists who travelled to the pole by ship. While this may be a comfortable way to get there, it has the direct effect of causing major damage to the ice pack, and increasing the surface area in contact with the ocean. Add to this the vast quantities of excess heat generated by the ship's nuclear reactors, and there is some reason to beileve that there will be localized dammge to the ice pack. Just as with permafrost regions, damage from human presence can be very long-lasting.
It is also worth noting that undersea measurements of polar ice thickness are taken from submarines. Now the overwhelming majority of those subs which cross those cold waters are nuclear-powered. This is a matter of necessity, since the diesel/electric boats are not suited to long runs under the ice. Those nuclear reactors generate vast quantities of heat, which is dissipated in-place, under the ice. Now no one transit by a single ship may cause any measurable change, but the thousands of transits, each emitting many megajoules of energy as heat, can in concert cause measurable melting.
Whether or not there are external effects caused by long-term climactic changes, there are most certainly local causes which must not be ignored.
The laptop computer is the existence proof that the silent computer exists.
Most systems are overcooled. This is due to the system design requirement that says that if a computer power supply can deliver, say, 200 watts, that it also must be able to remove that much as heat. The fans are sized accordingly.
The best solution is probably to design a cooling system that works on demand. Just like the fan on many autos, it runs when the temperature rises above a certain level. Make the fan with a concinuously variable fan, and it would be quiet most of the time.
I have noted that the K6-III has disappeared from the stores, and this explains it. It is a matter of product positioning. The K6-III had more cache, and fit the SS7 socket, but had a larger die than the K6-2. Why spend acres of silicon in on line 25 in Austin making K6-III parts, with a larger margin on the cheaper K6-2, when a new middle line processor is in the works.
Ah, economics... The new part needs shelf space by itself, so kill off the neato part nearest it.
The lesson here is to be silent about your naming plans for your new site until you act, and then snag the big three, com, org, and net, all at once. That is exactly what we had to do... and our organizational name was not confirmed until we knew we had the three doman names locked up. Perhaps our final name of freeio.org was not our first choice, but all three were available, and it was descriptive of what we do - GPL hardware designs.
Live and Learn...
-- The easiest way to lose your freedom is to fail to exercise it! --
One of the joys of working near Nashville, Tennessee is that our alternative music is actually somewhat local. So consider such older groups as Flatt & Scruggs, Bill Monroe, Roy Acuff, or more modern ones such as Ricky Scaggs & Kentucky Thunder, Allison Krauss, and such.
Ah, the joys of rule by decree. They can always change the decree whenever they wish. Note that this is not a change in law but rather a change in regulation. Oh, joy...
Slashdot Mirror
The list is fine as far as it goes, but all of it refers to commercial hardware. Some of us are working on easily customized reference designs for linux based controllers and micro systems.
At http://www.openh.org/ there are several ongoing linux controller projects released under free licenses.
At http://freeio.org/ we are developing and posting frequent updates to linux based controllers based on the ColdFire processor, released under GPL. The conceprt is free hardware designs for the free software community.
For those interested in rolling their own systems, or better yet contributing to free hardware designs and porting / building drivers, these are pretty good opportunities.
Anyone with sensitive data of any kind runs into the question of how to make it gone on demand. We all know ways of starting over on a drive to recycle it - load your favorite GNU/Linux or BSD distribution on it from scratch and enjoy.
However, the real question is this: if someone really wants to do so, how much data can be retrieved from a drive which has been reformatted and written over. Sure, most of us are going to give up fairly quickly, but if one puts enough effort into it, what data is still there waiting between the tracks? Unfortunately, at least some of the stuff from Tom Clancey novels is based on fact, and if you have the budget of a major government, and enough time, at least some data can sometimes be retrieved. If there are secrets which are worth enough to spend major time and money recovering, no data is ever quite gone.
Or at least, that is what those fabled government agencies base their budget requests upon.
When your company develops something under a contract, as an employee of the employer you are bound by that contract, as well as any government regulations regarding contracts. Depending upon the circumstances and the way the contract was written, the company may or may not retain copyright to the work. One common variable is whether the work was done under set-asides for small business, in which case the government frequently allows the copyright to remain with the company, and the government gets an unrestricted license to use the sofware which was developed. If the company has the rights to the copyright, then it can GPL the product at will.
You will have to have the company counsel read both the contract and the federal acquisition regulations (FARs) to determine how to proceed. We were in such a battle two years ago, and in our case it turned out that the copyright was ours to do with whatever we chose, as long as the government had the right to use what we developed.
Note also that FSF has a major section in their explanation of GPL dealing with the dangers of trying to GPL something to which you do not have the rights. There is danger here if you don't get it right.
Trademarking a commonly used text string has been done before, and has been solved before.
.arc extension for archived files being trademarked. The result was Phil Katz releasing the pkzip utilities, and the whole community switched over to zip files within a few months. SEA had made itself odious to the user community, and was cut off.
More than a decade ago, we saw this with SEA trying to enforce the
It could happen again.
Well, being /.ed is a heavy hit for what is nearly always a low volume technical server. Note three things:
1. openH.org is on a single server, and not a fast one at that;
2. The single server is operating out of the engineering department in a university;
3. The University is in Cape Town, South Africa, which limits the bandwidth and channnel speed a bit compared to what we in the USA are used to.
Marty
I currently have two hardware projects on openh.org, and am currently working on a third one. All three are meant to make life easier for the person who needs to develop hardware that runs software, free or not.
... no problem, can do. You want serial I/O? - ...reprogram it and it will do it. You need some arbitrary function? You can do that too. I provide the full specification of how it works, plus the CAD files for those who are interested, along with the manufacturing files and sample VHDL code to show how to use it as well - all under GPL.
The first two devices are ISA bus I/O cards for control of hardware devices. In this case, the cards themselves are programmable (in the VHDL language) to allow the hardware functions to be changed, without ever having to remove the I/O card from the computer. You want parallel I/O?
The current project is a GPLed control board using a Motorola ColdFire processor. Once again, all of the data is available right there on the openh.org site. The goal is to make it run the uClinux code with no customization actually required. This will be fully free software running on fully free hardware.
Now my projects are still available on my own site http://FreeIO.org/ but when Alan at the University of Cape Town approached me about hosting my projects on the openH.org site, I jumped at the opportunity. Just as with Linux or any software project which is developed in the open, there is the opportunity for others to benefit, and their feedback will help improve the project.
Contrary to the statements of some, developing hardware is not really harder than developing software, although the tools are different. Open source tools are not there yet, but free (beer) tools are available which work on smaller projects rather well. An example is the freeware version of the Eagle electronic CAD system http://www.cadsoft.de/ which handles schematic and printed wiring board designs very well, and runs natively under Linux. You do have to learn it to use it, but the same can be said or C/perl/python or whatever.
Free Hardware design? Absolutely.
Enjoy!
Marty
Several years ago I worked for an "engineering service contractor" for NASA, and had the opportunity to design the sensor electronics for a space-based sun-staring telescope, called Solar X-ray Imager. The task would not have been excessively difficult except that the radiation levels expected were quite high, and we had to design it to be very radiation resistant. This proved to be very expensive to do.
For certain electronic functions, there are radiation hardened integrated circuits. We used a set of CMOS devices manufactured by Harris (now Intersil) which are manufactured using a silicon-on-sapphire technology which makes them tough indeed. (The manufacturer claimed that they were radiation resistant up to "strategic levels" - to run through massive radiation doses and never glitch...) While they were very good parts, their cost was incredible. They were getting US$ 225 for a single 75hc00 equivalent device which usually costs 25 cents. Processors and memories had prices that were proportionally worse. Nobody but government bodies with the power of taxation can afford these things.
The other method of achieving radiation hardness is shielding. In our case, we had to use this in addition to the radiation hardened integrated circuits, because the telescope sensor itself was not radiation hardened (otherwise it would not have been able to see light either!) The shielding we used was made from tantalum, machined in shapes to cover the sensitive parts. Now tantalum is not inexpensive, it does not machine well, and is heavy. The weight is the limiting factor, since whatever you use you have to launch into space, and the cost per kg is very high. Shielding which is sufficient is also too heavy to launch on a secondary payload.
My thoughts? Amsat did the best they could within the budget, and the budget (raised from donations) did not provide enough money to use the premium radiation hardened parts everywhere.
Too bad...
73 de W4TI
When measurements of polar ice thickness are plotted since 1950, we usually assume that there are no local events which could cause such a thing. However, during that period of time there have been increasing numbers of intentional cases of breaking-up of the polar ice, in order to navigate the surface of the polar region.
Recently there were photos of a commercial expedition, using a Russian nuclear-power ice breaker, for those scientists, journalists, and tourists who travelled to the pole by ship. While this may be a comfortable way to get there, it has the direct effect of causing major damage to the ice pack, and increasing the surface area in contact with the ocean. Add to this the vast quantities of excess heat generated by the ship's nuclear reactors, and there is some reason to beileve that there will be localized dammge to the ice pack. Just as with permafrost regions, damage from human presence can be very long-lasting.
It is also worth noting that undersea measurements of polar ice thickness are taken from submarines. Now the overwhelming majority of those subs which cross those cold waters are nuclear-powered. This is a matter of necessity, since the diesel/electric boats are not suited to long runs under the ice. Those nuclear reactors generate vast quantities of heat, which is dissipated in-place, under the ice. Now no one transit by a single ship may cause any measurable change, but the thousands of transits, each emitting many megajoules of energy as heat, can in concert cause measurable melting.
Whether or not there are external effects caused by long-term climactic changes, there are most certainly local causes which must not be ignored.
The laptop computer is the existence proof that the silent computer exists.
Most systems are overcooled. This is due to the system design requirement that says that if a computer power supply can deliver, say, 200 watts, that it also must be able to remove that much as heat. The fans are sized accordingly.
The best solution is probably to design a cooling system that works on demand. Just like the fan on many autos, it runs when the temperature rises above a certain level. Make the fan with a concinuously variable fan, and it would be quiet most of the time.
L!
I have noted that the K6-III has disappeared from the stores, and this explains it. It is a matter of product positioning. The K6-III had more cache, and fit the SS7 socket, but had a larger die than the K6-2. Why spend acres of silicon in on line 25 in Austin making K6-III parts, with a larger margin on the cheaper K6-2, when a new middle line processor is in the works.
Ah, economics... The new part needs shelf space by itself, so kill off the neato part nearest it.
The lesson here is to be silent about your naming plans for your new site until you act, and then snag the big three, com, org, and net, all at once. That is exactly what we had to do... and our organizational name was not confirmed until we knew we had the three doman names locked up. Perhaps our final name of freeio.org was not our first choice, but all three were available, and it was descriptive of what we do - GPL hardware designs.
Live and Learn...
-- The easiest way to lose your freedom is to fail to exercise it! --
One of the joys of working near Nashville, Tennessee is that our alternative music is actually somewhat local. So consider such older groups as Flatt & Scruggs, Bill Monroe, Roy Acuff, or more modern ones such as Ricky Scaggs & Kentucky Thunder, Allison Krauss, and such.
Good stuff, all... and fine to code by.
Ah, the joys of rule by decree. They can always change the decree whenever they wish. Note that this is not a change in law but rather a change in regulation. Oh, joy...
I see no lasting good news here!