Yes, corporate policy to be good citizens of the Free Software community. The corporate policy most important for this press release is that OSDL only works on projects that have licensing compliant with the Open Source Definition, and it accepts projects from the community on a first-come first-served basis. There is, of course, some justification that one has to do to get the lab director to accept a project, I don't know if he'd want to test your IRC-bot:-).
Well, I was the one answering the reporter in this case, I think. And I said that Sun was much closer to Open Source than Microsoft. They have Linux on SPARC and they actually support it on Cobalt Micro systems, and they have freed OpenOffice, although they still have some big hurdles like Java licensing to jump over.
Of course, you always lose something in translation.
Certainly Sun would be welcome to participate in OSDL and Linux and Open Source in general.
There is some credible evidence that Mahlon Loomis invented radio about a quarter century before Marconi. For background, look here and here.
Thanks
Bruce
Re:Isn't this what Reiser FS is for?
on
MySQL FS
·
· Score: 1
You can manipulate tuples with a shell program. I once wrote a list-processing library in shell. OK, not the best way, but SQL is not the only way to extract a relation.
Bruce
Re:This project should help MySQL
on
MySQL FS
·
· Score: 2
Well, he's a bit confused but not too far off. The kernel would have been a heck of a lot less interesting if there hadn't already been the GNU system to run on top of it.
SQL is not the only way to extract a relation. There is room for improvement there. I don't particularly like SQL as a language. That's all I was trying to say.
Thanks
Bruce
Re:Isn't this what Reiser FS is for?
on
MySQL FS
·
· Score: 2
Yes, you don't have to tell me that:-) .
Re:Isn't this what Reiser FS is for?
on
MySQL FS
·
· Score: 1
Yes, but some would consider that an advantage:-)
Isn't this what Reiser FS is for?
on
MySQL FS
·
· Score: 4
Reiser FS is for building a database as a filesystem. See namesys.org.
You're right. But I doubt that msql can compete if they remain out of compliance with the Open Source Definition. There is mySql at the higher end, and sleepycat at the really-low end. So, maybe we'll see a change.
I audited this license while Apple was working on the draft. It's a good license. The only way in which it is GPL-incompatible is that it requires that you disclose some source code when the GPL would not. GPL only requires source code disclosure with distribution, APSL requires it if you use the code for your business, even if you do not distribute it.
This is of course speculation and won't be answered because of the fix he's in.
The matter is in the courts most likely means I sold an interest for stock, their stock is now worthless, by killing the issue they have devalued my own remaining interest, and they are keeping me from making money with my remaining interest. So, I took them to court.
The bottom line is that TPJ is an innocent bystander injured by a large train wreck.
I wrote a manager at Apple a long time ago, on behalf of the Freetype project, to ask about use of the patents in GPL software, and got no reply. I think there have been some staff changes since then, and it would make sense to ask again. But my understanding was that if they ever had to write out the patented algorithm, they could do so.
RF also follows the inverse square law. Just guessing, would not all non-coherent radiation?
Magnetic force is not important in itself unless it's so strong that it moves things. It's what is induced that is important. CMOS chips are not so bad - they have really short internal leads, so you need only concern yourself with what comes in on a wire. CMOS chips generally include diodes to power and ground at every I/O pin.
I've seen HIPOT testing with a spark gun. It crashed the workstation being tested, but did not damage it. We could probably have gotten rid of the crash, too. Nice big 1N4007 diodes to the power supply rails at all of the I/O pins might have done it.
It's pretty easy to suppress phone line EMI. Chokes, MOVs, gas tubes, all help. Good ones are too expensive for consumer equipment, but most modems I buy have two chokes, a few MOVs, and a gas tube that looks like a neon lamp.
EMP works by coupling a momentary AC electromagnetic signal to conductors that work as an antenna and deliver a high voltage to equipment. Most military aircraft are not vulnerable. I happen to have some military-surplus EMP suppressors on hand, and use them for lightning protection on my ham station. These are bulkhead-mounting devices that take the antenna cable from inside the metal aircraft body (think Faraday cage) to outside where the antenna is. They use a gas tube (which becomes conductive when a high voltage is imposed across it and shorts it out) an air gap (similar to the gas tube but a second line of defense) and a choke coil. For fixed-frequency radios, you can also use a 1/4 wave stub which appears as an open at the desired frequency and a short at many other frequencies (but is not as effective for lightning protection as people think because its impedance is too high).
My local area net uses shielded cables with only one end grounded (to avoid ground loops). This is to keep interference out of the radios, but protects the net from outside RF as well.
I have put choke coils (a toroid with a few turns of AC wire around it) in a lot of the electrical outlet boxes (again for interference reduction). But I could do a lot more work on that.
It's not a shielding experiment. The StrongARM processor architecture happens to be right for rad-hardness because it's fully static and it provides the most MIPS per watt of any modern CPU. The thing only uses 300 miliwatts at 230 MHz, vs. about 5 watts for any Pentium implementation. This is important both from a power budget,remember that power comes from solar panels and batteries for the dark periods, and we also have radio transmitters in that power budget, and a heat standpoint. A perfect black-body radiator in vaccumm can not efficiently radiate as much heat for its surface area as the typical CPU generates and you thus get into awkward active-cooling schemes.
Shielding is often ineffective because if you hit shielding with a high-energy particle the result is a bunch of low-energy particles that are even worse as far as memory and logic cells are concerned. The preferred shielding material is tantalum which is expensive and very difficult to machine, and adds mass to be lofted to orbit. Static rather than dynamic logic is more rad-hard and uses less power, so the StrongARM is a very good choice for a satellite.
No, it's not so big. If they haven't banked the 1802 address space, it's only able to address 64K. The environment they are programming in sounds like something in between 1802 assembler and Forth. The second flight computer has a StrongARM and actually addresses a reasonable size memory, but that computer is itself experimental.
It turns out the satellite has no boot ROMs. There is a hardware mode to DMA a program from the radio receiver into RAM and then execute it, and that's how the thing gets bootstrapped. They feel this is more reliable than having one byte of airborne code that could fail. Bootstrapping was also complicated that not all of the receivers were working, and the receiver that was working has a directional antenna that is not pointed toward the earth during part of its orbit. So, you needed a ham in the right place with S-band equipment (which you don't just buy at Ham Radio Outlet).
Why not put 8 processors behind one on-chip cache with write-back? That ends up looking like a single CPU off-chip. Aren't crossbars for explicitly-parallel algorithms?
Do you really have to chuck the whole die? Why not just blow the fuse to disable that CPU and sell a die with 7 CPUs instead of 8? My understanding is that memory yield-increasing technology works this way - they have "spare" rows that are production-time programmed to replace defective rows.
Thanks
Bruce
Bruce
Of course, you always lose something in translation.
Certainly Sun would be welcome to participate in OSDL and Linux and Open Source in general.
Thanks
Bruce
Thanks
Bruce
Bruce
Thanks
Bruce
Thanks
Bruce
Thanks
Bruce
Yes, you don't have to tell me that :-) .
Yes, but some would consider that an advantage :-)
Bruce
Thanks
Bruce
Thanks
Bruce
OK. Well, I won't push you any farther. Sorry.
Besides, most of the developers have already moved to K5.
Thanks
Bruce
The matter is in the courts most likely means I sold an interest for stock, their stock is now worthless, by killing the issue they have devalued my own remaining interest, and they are keeping me from making money with my remaining interest. So, I took them to court.
The bottom line is that TPJ is an innocent bystander injured by a large train wreck.
Thanks
Bruce
Thanks
Bruce
Magnetic force is not important in itself unless it's so strong that it moves things. It's what is induced that is important. CMOS chips are not so bad - they have really short internal leads, so you need only concern yourself with what comes in on a wire. CMOS chips generally include diodes to power and ground at every I/O pin.
I've seen HIPOT testing with a spark gun. It crashed the workstation being tested, but did not damage it. We could probably have gotten rid of the crash, too. Nice big 1N4007 diodes to the power supply rails at all of the I/O pins might have done it.
It's pretty easy to suppress phone line EMI. Chokes, MOVs, gas tubes, all help. Good ones are too expensive for consumer equipment, but most modems I buy have two chokes, a few MOVs, and a gas tube that looks like a neon lamp.
Thanks
Bruce
My local area net uses shielded cables with only one end grounded (to avoid ground loops). This is to keep interference out of the radios, but protects the net from outside RF as well.
I have put choke coils (a toroid with a few turns of AC wire around it) in a lot of the electrical outlet boxes (again for interference reduction). But I could do a lot more work on that.
Shielding is often ineffective because if you hit shielding with a high-energy particle the result is a bunch of low-energy particles that are even worse as far as memory and logic cells are concerned. The preferred shielding material is tantalum which is expensive and very difficult to machine, and adds mass to be lofted to orbit. Static rather than dynamic logic is more rad-hard and uses less power, so the StrongARM is a very good choice for a satellite.
Thanks
Bruce
Thanks
Bruce
These guys make me feel like a piker.
Thanks
Bruce
Thanks
Bruce
Thanks
Bruce
Thanks
Bruce