Also, saftey-critical code is requirements driven. That is, the requirements are defined, then the code is written to implement the requirements. Using this approach, even if you could find something in Windows that fit your requirements, by the time you removed everything else, it wouldn't be Windows anymore. There would likely be no requirement for the user interface, for example.
Well, I know for a fact that the major medical device manufacturers 'embed' the mainstream OSes into critical life support systems. So there's anectocal evidence that the FDA approves of Windows and OS/2 in such applications.
I was not a part of the auditing process, and in fact didn't work at the company during the time when they qualified the OS/2 they embedded. I was there during the transition to NT and know it was on-schedule, so not just marketing hype. And I know how tough the Verification-Test people were on the firmware that the team I worked with was coding. More anecdotal evidence than anything else.
Billion dollar critical care medical device companies use Windows NT and OS/2 and similar OSes in the console/external controllers of their implanted devices.
Thanks for the 'Government Computer News: US Navy Ship stalled because of Windows NT' Anecdote. That one sure gets recycled a lot.
Often there is a 'restrictor' in the shower head that you can easily remove. I put in a new shower head at the Townhouse I lived in last, and it was pitiful. Then I took the head off and noticed that there was an insert I could remove with a regular phillips screwdriver.
I think it's sort of an aptitude test. People who don't know how to use a screwdriver suffer. Also, I probably broke some sort of law by making my modification.
The "more eyes looking at it" is a good thing, since only takes one person to find a hole and report it.
That's true, as far as you take it. But the Open Source codebase history is full of examples that can be cited where there have been glaring holes that were only uncovered after years of use of the code. In those instances "it only takes one person to find a hole and keep it secret to exploit, until someone else finds it who 'tells.'" The knife cuts both ways.
And it's certainly easier for smart malevolent hackers, highly educated ones working for 'enemy' agencies, to review Open Source looking for exploits than it is to pound away at binaries with a debugger.
And then, if you want to connect it to a network, you have to introduce and audit what you add in. Integrators build on what is audited. What Microsoft did isn't as ridiculous as some people make it out to be.
Why assembly language? That just breeds lazyness, as you say, in the developers.
Use Machine Code, entered on a bank of toggle switches. You're not gonna have Lazy Programmers doing that! They'd probably become more physically fit than ever!
Your suggestion of huge forks of the code while maintaining rigorous binary compatability.... hmmm... maybe I'll let somebody else tackle that one...
It might actually be more capable across multiple architectures than Linux. Because it doesn't drag along as many 'universal' needs and applications, it can be lighter and more focused, and thus easier to port around. If it doesn't require a multi-user environment, it can do away with that whole layer of complexity, the way that BeOS did, with their POSIX-like design that wasn't bogged down with a multiuser model.
Diverse hardware support is completely irrelevant and unimportant in many embedded applications. It's nice if there is a code base out there to rely on as a starting point, but you're going to roll in exactly the only driver support that you need, and for a RTOS, you'll probably want to code it closer-to-the-hardware than it is in a general purpose OS. There's not going to be a lot of auto-seeking and device probing going on during bootup. Drivers are going to be hard coded to exactly what's in the target hardware. Loose auto-seek designs lower reliability, and introduce many, many variables that make testing the system much more complex for the VT team, thus much more expensive.
Windows could easily be qualified, by a team under NDA with a source license for Windows.
I've worked at a medical device company that produces implantable cardiac devices. They write their own code for the embedded devices, but all reading and control of the devices is done with licensed 'embedded' copies of off-the-shelf OSes in the 'Programmer' system. With Windows NT or OS/2, to be specific. In systems like that a 'narrow' approach is taken to qualify the design. A specific controlled code base is used, and a specific controlled hardware platform is chosen. It's rigorously tested and normally 'closed' components are audited.
You're almost certainly wrong in your assertion about Windows. Perhaps you work for a vendor who likes things that way, because they'd rather not build on a Windows platform. That's not an unreasonable approach to take, especially if you can make more money rolling your own and/or building in a codebase you've got your fingers deeper into.
It is worth adding to your point that the 'Cathederal Method' versus the 'Bazaar Method' is not an open versus closed source schism. Raymond wrote that essay as a criticism of the small-group closed way that the GNU Emacs team was doing their work. It's quite possible, and rather common for source-disclosed projects, even ones that are released with the GPL or BSD licenses, to be developed by small closed groups who don't actively solicit outside code.
and closed systems that crash vehicles into Mars due to metric to English conversion bugs.
Don't you think you're overreaching a little bit there? How does open or closedness affect something like that? You presented some good points but when you plop in a chestnut like that one, you make yourself look foolish.
Solaris x86 could still be positioned as a strong Linux and Windows desktop competitor if they chose to push it. It now has the Gnome desktop and with good Java support and OpenOffice it's a pretty good system. I'm running one box with it now and am pretty happy with how it performs.
1. They gave up on a new OS and just bought one from NeXT.
2. They've taken to promoting Scully's expensive sugar water (Pepsi) with 'winning bottlecaps' to download pop tunes from an Apple server and play them back with their overpriced Walkman clone.
That's not anything like the innovation Apple was doing a decade ago. I guess it's innovative in the mind of marketing types, though.
Websites aren't being taken offline. What is happening is that a default home page displays info about the issue, and a link is provided to access the website.
So you're suggesting Scotts should produce a type of grass that will make it easier for their customers to buy less of their chemicals and other lawn care products?
Let's see. Why would they possibly think that wasn't such a good idea....
I built a PAiA Theremin from kit form in 1979. It was a pretty neat instrument.
I attempted to build one from scratch a few years before that. But that was when I was a kid in H.S. and didn't have good soldering skills, etc. Plus that earlier-design theremin used 'coils.' I can't remember the number of times when I was a kid that I looked at the schematic for a project and was disappointed that it used a coil. Dunno why I was so afraid of winding coils, but probably because they're fussier to make than slapping in resistors, capacitors, and transistors.
My first electronic musical instrument was an electronic organ I built, using the schematic for the Code Practice Oscillator in a Boy Scout merit badge pamphlet. I build the 'oscillator' and then figured out 'I can change the resistor and it varies pitch' and then 'I can build eight or ten 'keys' for my code practice oscillator, all which put a different resistor value in the circuit. That was in fifth grade. Back in the 60's.
The IBM 5100, which my father brought home from work a few times when I was in High School for us to mess with, had a program that printed a bunch on the line printer, and the sounds the line printer made while printing this particular pattern was a fairly good rendition of the William Tell Overture. The rhythm pattern, at least.
That was in the medium-old days, before the PC. Not sure what the 5100 is classified as.
The RS/6000 that I ran NetBSD on was one of the PREP boxes. So it had a PowerPC processor, ISA and PCI slots, an S3 video chip, and used PS/2 keyboard and mouse. In all regards except for running on a PPC, it was a PC. I ran NetBSD/PREP on it. Not all RS/6000 boxes are as 'PC compatible' as that machine (a 7248 box). I now have an absolutely ancient box, one based on the Power1 processor, that I seem to be limited to running AIX on. Though I will be checking on that before long.
I bought what I thought was a retail-box version of MASM (Microsoft's Macro Assembler). It's the last version they ever sold of that product, so it's relatively 'current.' I bought it on eBay and paid a fairly stiff price for it.
Now, I've decided that I don't need it anymore. I found a textbook that has the same version of MASM on a CD-ROM, and I found a place to download all the user and reference manuals as PDF files. So I decided to relist it and try to recoup my losses on eBay (would make sense, since the reason I paid so much for it there was that people bid against me so there is a market for something as arcane as MASM.) Almost immediately after I listed it someone (whose bidding history shows that they buy and resell Microsoft and Borland development tools cheap on eBay preusmably for resale) asked me for details about it. I looked on the part number, and yes, there was an AE in it, indicating it's an Academic Edition. So I delisted it.
Effectively, you can get away with selling Microsoft products on eBay without them having 'clean' credentials, but you take a risk when doing so. And you take a risk when buying such software on eBay that you're buying something that you'll not be able to sell back.
But I am a boxed-software collector, so I've just put the MASM box and manuals and stuff up on the shelf. It's not as old as the IBM BASIC Compiler 1.0, PC-DOS 1.0, Windows 1.03, and DR-DOS/GEM boxes that I have in intact condition, but it's part of 'the collection' (sigh).
Not only that. It would mean that there would need to be some sort of unaccredited 'Open Source' alternative icon. Which would then be adopted by unscrupulous vendors of 'unclean' software.
This is a particular camel whose nose shouldn't be allowed in the tent.
Slashdot Mirror
Also, saftey-critical code is requirements driven. That is, the requirements are defined, then the code is written to implement the requirements. Using this approach, even if you could find something in Windows that fit your requirements, by the time you removed everything else, it wouldn't be Windows anymore. There would likely be no requirement for the user interface, for example.
Well, I know for a fact that the major medical device manufacturers 'embed' the mainstream OSes into critical life support systems. So there's anectocal evidence that the FDA approves of Windows and OS/2 in such applications.
I was not a part of the auditing process, and in fact didn't work at the company during the time when they qualified the OS/2 they embedded. I was there during the transition to NT and know it was on-schedule, so not just marketing hype. And I know how tough the Verification-Test people were on the firmware that the team I worked with was coding. More anecdotal evidence than anything else.
Billion dollar critical care medical device companies use Windows NT and OS/2 and similar OSes in the console/external controllers of their implanted devices.
Thanks for the 'Government Computer News: US Navy Ship stalled because of Windows NT' Anecdote. That one sure gets recycled a lot.
Often there is a 'restrictor' in the shower head that you can easily remove. I put in a new shower head at the Townhouse I lived in last, and it was pitiful. Then I took the head off and noticed that there was an insert I could remove with a regular phillips screwdriver.
I think it's sort of an aptitude test. People who don't know how to use a screwdriver suffer. Also, I probably broke some sort of law by making my modification.
Well, I would be surprised. You seem to be an outsider just speculating.
The "more eyes looking at it" is a good thing, since only takes one person to find a hole and report it.
That's true, as far as you take it. But the Open Source codebase history is full of examples that can be cited where there have been glaring holes that were only uncovered after years of use of the code. In those instances "it only takes one person to find a hole and keep it secret to exploit, until someone else finds it who 'tells.'" The knife cuts both ways.
And it's certainly easier for smart malevolent hackers, highly educated ones working for 'enemy' agencies, to review Open Source looking for exploits than it is to pound away at binaries with a debugger.
And then, if you want to connect it to a network, you have to introduce and audit what you add in. Integrators build on what is audited. What Microsoft did isn't as ridiculous as some people make it out to be.
Why assembly language? That just breeds lazyness, as you say, in the developers.
Use Machine Code, entered on a bank of toggle switches. You're not gonna have Lazy Programmers doing that! They'd probably become more physically fit than ever!
Your suggestion of huge forks of the code while maintaining rigorous binary compatability.... hmmm... maybe I'll let somebody else tackle that one...
Like that one were they ALLOW PEOPLE TO LOG IN AS ROOT. That is system admininstrator,
Whoah! You're throwing around the big heavy words now. Better, ummm... work on your spelling, though.
Back in the 19th century, frecklefaced adolescent boys wanted to be riverboat pilots.
In the early 20th centyr, the same boys wanted to be railroad engineers.
Now it appears that 'System Administrator' has filled that niche of boy wonderment for some.
It might actually be more capable across multiple architectures than Linux. Because it doesn't drag along as many 'universal' needs and applications, it can be lighter and more focused, and thus easier to port around. If it doesn't require a multi-user environment, it can do away with that whole layer of complexity, the way that BeOS did, with their POSIX-like design that wasn't bogged down with a multiuser model.
Diverse hardware support is completely irrelevant and unimportant in many embedded applications. It's nice if there is a code base out there to rely on as a starting point, but you're going to roll in exactly the only driver support that you need, and for a RTOS, you'll probably want to code it closer-to-the-hardware than it is in a general purpose OS. There's not going to be a lot of auto-seeking and device probing going on during bootup. Drivers are going to be hard coded to exactly what's in the target hardware. Loose auto-seek designs lower reliability, and introduce many, many variables that make testing the system much more complex for the VT team, thus much more expensive.
Windows could easily be qualified, by a team under NDA with a source license for Windows.
I've worked at a medical device company that produces implantable cardiac devices. They write their own code for the embedded devices, but all reading and control of the devices is done with licensed 'embedded' copies of off-the-shelf OSes in the 'Programmer' system. With Windows NT or OS/2, to be specific. In systems like that a 'narrow' approach is taken to qualify the design. A specific controlled code base is used, and a specific controlled hardware platform is chosen. It's rigorously tested and normally 'closed' components are audited.
You're almost certainly wrong in your assertion about Windows. Perhaps you work for a vendor who likes things that way, because they'd rather not build on a Windows platform. That's not an unreasonable approach to take, especially if you can make more money rolling your own and/or building in a codebase you've got your fingers deeper into.
It is worth adding to your point that the 'Cathederal Method' versus the 'Bazaar Method' is not an open versus closed source schism. Raymond wrote that essay as a criticism of the small-group closed way that the GNU Emacs team was doing their work. It's quite possible, and rather common for source-disclosed projects, even ones that are released with the GPL or BSD licenses, to be developed by small closed groups who don't actively solicit outside code.
and closed systems that crash vehicles into Mars due to metric to English conversion bugs.
Don't you think you're overreaching a little bit there? How does open or closedness affect something like that? You presented some good points but when you plop in a chestnut like that one, you make yourself look foolish.
Solaris x86 could still be positioned as a strong Linux and Windows desktop competitor if they chose to push it. It now has the Gnome desktop and with good Java support and OpenOffice it's a pretty good system. I'm running one box with it now and am pretty happy with how it performs.
Apple keeps innovating?
I thought they quit that awhile ago. Lately:
1. They gave up on a new OS and just bought one from NeXT.
2. They've taken to promoting Scully's expensive sugar water (Pepsi) with 'winning bottlecaps' to download pop tunes from an Apple server and play them back with their overpriced Walkman clone.
That's not anything like the innovation Apple was doing a decade ago. I guess it's innovative in the mind of marketing types, though.
Websites aren't being taken offline. What is happening is that a default home page displays info about the issue, and a link is provided to access the website.
Look once before you rant, please.
So you're suggesting Scotts should produce a type of grass that will make it easier for their customers to buy less of their chemicals and other lawn care products?
Let's see. Why would they possibly think that wasn't such a good idea....
Oh, puhlease! Cowboy Neal has better taste than that! He drives the Slashdot Cruiser!
Are you saying AOL is a legislative body or law enforcement agency?
That will be a scary day, incidentally. They own a big chunk of 'land' in cyberspace already...
The slashdot editors are spamming us with this story.
Ummm, what kind of car does Cowboy Neal drive??
Raffle, anybody?
I thought 'Industrial' started with Throbbing Gristle, in the U.K.
I built a PAiA Theremin from kit form in 1979. It was a pretty neat instrument.
I attempted to build one from scratch a few years before that. But that was when I was a kid in H.S. and didn't have good soldering skills, etc. Plus that earlier-design theremin used 'coils.' I can't remember the number of times when I was a kid that I looked at the schematic for a project and was disappointed that it used a coil. Dunno why I was so afraid of winding coils, but probably because they're fussier to make than slapping in resistors, capacitors, and transistors.
My first electronic musical instrument was an electronic organ I built, using the schematic for the Code Practice Oscillator in a Boy Scout merit badge pamphlet. I build the 'oscillator' and then figured out 'I can change the resistor and it varies pitch' and then 'I can build eight or ten 'keys' for my code practice oscillator, all which put a different resistor value in the circuit. That was in fifth grade. Back in the 60's.
The IBM 5100, which my father brought home from work a few times when I was in High School for us to mess with, had a program that printed a bunch on the line printer, and the sounds the line printer made while printing this particular pattern was a fairly good rendition of the William Tell Overture. The rhythm pattern, at least.
That was in the medium-old days, before the PC. Not sure what the 5100 is classified as.
The RS/6000 that I ran NetBSD on was one of the PREP boxes. So it had a PowerPC processor, ISA and PCI slots, an S3 video chip, and used PS/2 keyboard and mouse. In all regards except for running on a PPC, it was a PC. I ran NetBSD/PREP on it. Not all RS/6000 boxes are as 'PC compatible' as that machine (a 7248 box). I now have an absolutely ancient box, one based on the Power1 processor, that I seem to be limited to running AIX on. Though I will be checking on that before long.
I'm sorry. The idea of a 'software police' with the kind of enforcement powers that the FDA have just frightens me.
I bought what I thought was a retail-box version of MASM (Microsoft's Macro Assembler). It's the last version they ever sold of that product, so it's relatively 'current.' I bought it on eBay and paid a fairly stiff price for it.
Now, I've decided that I don't need it anymore. I found a textbook that has the same version of MASM on a CD-ROM, and I found a place to download all the user and reference manuals as PDF files. So I decided to relist it and try to recoup my losses on eBay (would make sense, since the reason I paid so much for it there was that people bid against me so there is a market for something as arcane as MASM.) Almost immediately after I listed it someone (whose bidding history shows that they buy and resell Microsoft and Borland development tools cheap on eBay preusmably for resale) asked me for details about it. I looked on the part number, and yes, there was an AE in it, indicating it's an Academic Edition. So I delisted it.
Effectively, you can get away with selling Microsoft products on eBay without them having 'clean' credentials, but you take a risk when doing so. And you take a risk when buying such software on eBay that you're buying something that you'll not be able to sell back.
But I am a boxed-software collector, so I've just put the MASM box and manuals and stuff up on the shelf. It's not as old as the IBM BASIC Compiler 1.0, PC-DOS 1.0, Windows 1.03, and DR-DOS/GEM boxes that I have in intact condition, but it's part of 'the collection' (sigh).
Not only that. It would mean that there would need to be some sort of unaccredited 'Open Source' alternative icon. Which would then be adopted by unscrupulous vendors of 'unclean' software.
This is a particular camel whose nose shouldn't be allowed in the tent.