I noticed that very few of the posts I've read (on the dupe thread at least) caught the real issue (from the 1st time this got posted).
The problem is that the computers on the ground and the space shuttle carry over to the new year differently. One goes to day 1 of the next year, the other goes to day 366 of the current year, but I forgot which side does which. This is a problem NOT because the calander is wrong, but because things that need to sync time between the two is off. This is for important things like starting operations like reenty, space station docking, etc. Also, they most likley use a date/time code that's based on days, and ignores months, and most likley ignores day of the week too. So Dec. 31st would just be Day 365 (unless it's a leap year).
This seems like an easy fix, until you take into account all of the systems that need to be fixed, vs. the relativly easy workaround of not flying when the year switch happens. It's not just the shuttle systems, but the tracking stations, mission control centers, possibly some code on the ISS, etc. Not to mention the verification side, and making sure you didn't break anything else....
Someone mentioned Daylight Savings Time switchovers and asked how the space systems deal with that but can't deal with a year change. The answer is, the systems do not deal with daylight savings time changeovers, and don't have to. All the time on every system is in Zulu (GMT) time, some displays might be in local, but that's just for convience, nothing is based off of local time.
So, what code do you review? And what kernel version? And what happens when there's the next big kernel version jump? How many people and how much time does it take to review all that code?
The advantage of an RTOS like LynxOS or Green Hills Integerity, or VxWorks, is that it just works, sure that price sounds pretty big, but compare it to the time of digging through all the code yourself. It is possible to get a version of Linux that could pass DO-178B certification, but it's something else entirerly to keep the kernel tree anywhere close to up to date while maintaing certification.
It's really easy to say something about the code being readily avilable. It's another thing for a given orginization to actaully review it, or even a small enough subset of it for the core kernel and a small number of drivers.
Other then the remote start, no. I didn't see any details on how he's doing that. I'd be kinda nervious about anything critical using *either* windows, or linux (or any other OS that's targeted more for desktops, servers, clusters, mainframes, set-top boxes....)
Then again, when I work on "mission critical systems", I'm talking about flight control hardware/software... But that's the same mentaility you *should* take.
OK, that I'll give you. Hell, usability in the professional software market isn't great either. I saw a program (for VLSI chip design) that crashed when you clicked on zoom in twice in a row. On a Sun, sometimes totally loging you out of the system, other times killing it off totally. For a $100K/user/year program. (Last year btw).
I hate to make you sadder, but I don't even really know what illustrator is (I basicly started reading the comments to see the ones on "wth does this program do"
And, btw, why does the fact that people don't know what Inkscape is sadden you? Maybe if this was an all art site... Should I be sadded that most people have no idea what I'm talking about when I rattle off the tools I use regularly (primarly digital hardware and printed circuit design)
I wasn't comparing it to a CS class, I was comparing it to a Computer Engineering class. I've seen CS majors try to design hardware, and it wasn't pretty most of the time. In Comptuer Engineering, chip design or FPGA design is a lot of what we do.
Granted, it was an oversimplification, but I've seen groups that were a mix of grad students and undergrads crank out hardware designs a hell of a lot more complicated then a CPU and have a fairly good implementation. (Granted, mostly targeting FPGAs. One example I know a lot about was speech processing/reconigition.)
Uhh, as far as I understand most of them are Japenese? I had read (granted in a book from 1987) that the big American ones were pretty much being killed off.
Given that the above may be wrong, then they learned the hard way. If it's right, well, Japenese companies always took a longer term view. Not as innovataive though, it'd be nice to have management with a longer-term view + the US culture of encoraging radical ideas in R&D.
From the article: "The move still shows that China is capable of designing complex microprocessors."
As is any junior level class in computer engineering... A bit more complex then the Hennasy and Patterson (The Classic!) book on computer arch and cpu design, but not by much. Given what (admittedly little) the article said about China's CPU fabs, I wonder if the newest Xilinx or Altera FPGA's could implement the design and run it faster..
Actaully, it's not for every game, just every game that charges to pay, for bigger LARPs its pretty common to chip in $1 to cover the costs of the people running the game.
Depends on the FPGA, you can get one from Xilinx in low quanity that'll run a soft-core processor for about $7 (Spartan-3 series, newest (90 nm) chip in their low cost/high volume line), in production quanity is a lot cheaper.
In response to the other posters, I have not seen much in terms of open source software for FPGA design, and 0 that worked for creating the hardware bitstream to dump into a chip. There are free versions of the tools avilable from Xilinx and Altera for their respective chips.
And, FPGAs are really useful as an implementation, not just for prototyping. It can cost up to $30M to set up an ASIC line, not counting engineering or the $100K+ for the tools...
It's simple, if you're a programmer, and you really are that good, you can do more with less memory and cpu cycles. I can write a program that runs on a $5 chip with 256 bytes of ram, and 16k flash, and have it controlling a robot. Go do that in java:)
Nothing can compare to having a dedicated test staff. At the last software place I worked (part-time, in school, while getting my engineering degree), we had 3-6 college students working on testing most of the time (we would also be given some small projects to work on).
Testing goes far beyond what any automated system can test, if you have a user in there somewhere. You also need to check things like "How easy is it to use?" and "Does this feature make sense?". We also suggested features that the program did not have, but from our experiance using it, thought that it should have.
I just want to say that I have no problem, in principle, with using open source for something like the RTOS in a flight control system, or some other hard real-time application.
My issue is with trying to make a hard RTOS out of a *nix base, I don't think it's the best approach to take. The only OSS RTOS I know of is XMK, which is nice, but only supports microcontrollers. Is there a true OSS counterpart for high end RTOS's yet?
The June issue had an article from GHS, and one from LynuxWorks. COTS Journal is published for developers (Systems, hardware, and software) of commerical off the shelf products, which is essentailyl a standard for military procurement where possible (You can get a computer, OS, etc, "off the shelf", but something like a sub or cruise missle is another story)
I never understood the point of his FUD, up until he started writing about this, I would not have considered Linux a viable alternative to Integrity (One of the RTOS's from GHS), for a varaity of reaons. Keep in mind that this is not a destop, workstation, or server OS, this is hard real-time. You can't use something like Integrity for a desktop or server (well, it might be possible, but I doubt that it would perform as well as some *nix or even NT/2k/XP). And you shouldn't use something like Linux or Windows Super Server of the Year Edition for flight control. He actaully has me wondering how well the latest "real time" patches for Linux actaully perform....
On a side note, I know a few people who work at places like Northrup and Lockheed, and from what I've heard, on a lot of projects, code is changed, documented, changes logged, and tested on almost a line by line basis (Obviously not counting a curly bracket or something like that as a seperate line). The versioning software keeps track of who changed what on what line, and when. That's not something you will find in Linux, any open source developers here up for that level of documentation? On the other hand, it is true that that's something you won't find from microsoft either.
Over long range, if you want to be able to hit a waypoint with any sort of accuracy, the easist way is to use GPS. Dead Reckoning requires updates based on landmarks, and a decent odomiter and compass as well. The alternative is to use an inertial system, a good one is pretty expensive. Both of the above also depend on knowing your starting position accuratly. Incremintal error *will* add up. There is a reason that the DoD created GPS. I'm working on a robot navigation system for my Senior Design Project ( http://www.pitt.edu/~mmdst23 has the old details, I plan on fixing that, and adding some source to our Sourceforge site soon, but only after we get things working properly.) We're using DR, but thats because it's better in an indoor enviroment, with the low cost of a Differential GPS receiver, there arn't many reasons to NOT use one, unless you're really paranoid about GPS jamming/spoofing (which is why the military GPS has an encrypted correction signal as well). That said, a backup DR system does appear to be needed, and using a compass is important anyway, GPS can give your corse over ground, which is where you're going, but it cant give your heading, which is where you're pointed at.
At one time I thought that as well. Currently I'm working part time for a company that does some ship navagation software in testing (I can schedule around classes), and I can say that that model really dosen't work.
The big problem is that you don't have to buy the charts from us, we're compadable with other charts as well, and there are standards for electronic charts that many vendors follow (I'm not sure, but I'd assume that street-level charts have a simular standard). So in this case, the company making the software could end up not making money at all. The give the software away for free method works best where people/companies would pay for services, the only way this could work out for the company would be if the software somehow could only use the maps they sell, and they charged a lot more money for them.
The cheapest FPGA I know of is the Altera Educational board, you can get it for about $150 if you are a student (I'm not sure what else, you might have to go to one of the schools (colleges) that participate in their program.) For $150, you get suprisingly little performance, I'd be suprised if you could get a Mips 3k (one of the first mips cpu's, 32 bits and really slow), or an early Sparc cpu in one. I'd be really suprised if you could get a fast 486 or early pentium in one.
Well, to be honest, I didn't quite make that point earlyier. Sure, in many cases, a game thats pure eye candy is not as much fun as a game thats got a lot of depth, but older graphics. But I'd say both are important, if you make a game with the proper mix of good graphics and good gameplay, you have a classic. Look at a games like System Shock 2, Deus Ex, or Max Payne, more powerfull 3d can be good for gamers. Then again, look at all the crappy 3d shooters that came out, and it seems bad. The key here is to buy games from a place that lets you return them if they suck, or at least try the demo and read a review or 2 first.
Re:Linux and Video
on
Loki Goes Postal
·
· Score: 3, Informative
Do you mean "most Windows users" or "most Windows gamers" ? Most gamers will have a decent 3d card, and most of the "major" cards (Nvidia, ATI, Matrox, 3dfx) will work, at least most of the time. The Kyro chips will probally have support
soon too.
As for a "low-quality" software rendering, what do you mean? Less detail, lower resolution? Its not really possible to have lower detail levels then whats in the game, and you can't run a game in less then the lowest resolution the game supports. Most of what you're looking for could only be done in the game's code itself, and to be honest, mean, cruel, blunt (whatever you want to call it), if you don't have a 3d card, you're pretty far off from the target audience for a 3d game. Games need to constantly be innovative and use new technology, and that needs more 3d power. Basicly, a 3d card thats 2 years old (or more) should be replaced if you want to play the latest 3d games. The good news is that the price of cards goes down as well, so you can get a still decent but 6 month old card for a lot less then what it would have cost new. If 3d is annoying you enough, you could (probally) find a decent 3d card for ~ $50-100, or a pretty good card for ~ $100-125.
Slashdot Mirror
I noticed that very few of the posts I've read (on the dupe thread at least) caught the real issue (from the 1st time this got posted).
The problem is that the computers on the ground and the space shuttle carry over to the new year differently. One goes to day 1 of the next year, the other goes to day 366 of the current year, but I forgot which side does which. This is a problem NOT because the calander is wrong, but because things that need to sync time between the two is off. This is for important things like starting operations like reenty, space station docking, etc. Also, they most likley use a date/time code that's based on days, and ignores months, and most likley ignores day of the week too. So Dec. 31st would just be Day 365 (unless it's a leap year).
This seems like an easy fix, until you take into account all of the systems that need to be fixed, vs. the relativly easy workaround of not flying when the year switch happens. It's not just the shuttle systems, but the tracking stations, mission control centers, possibly some code on the ISS, etc. Not to mention the verification side, and making sure you didn't break anything else....
Someone mentioned Daylight Savings Time switchovers and asked how the space systems deal with that but can't deal with a year change. The answer is, the systems do not deal with daylight savings time changeovers, and don't have to. All the time on every system is in Zulu (GMT) time, some displays might be in local, but that's just for convience, nothing is based off of local time.
Weird, what cert is that for? And why is it the opposite of the ratings for DO-178B, where A is the most critical level?
So, what code do you review? And what kernel version? And what happens when there's the next big kernel version jump? How many people and how much time does it take to review all that code?
The advantage of an RTOS like LynxOS or Green Hills Integerity, or VxWorks, is that it just works, sure that price sounds pretty big, but compare it to the time of digging through all the code yourself. It is possible to get a version of Linux that could pass DO-178B certification, but it's something else entirerly to keep the kernel tree anywhere close to up to date while maintaing certification.
It's really easy to say something about the code being readily avilable. It's another thing for a given orginization to actaully review it, or even a small enough subset of it for the core kernel and a small number of drivers.
No one would ever design a critical flight system for commerical aviation based on windows. Thats why they make DO-178B'able OSs :)
Other then the remote start, no. I didn't see any details on how he's doing that. I'd be kinda nervious about anything critical using *either* windows, or linux (or any other OS that's targeted more for desktops, servers, clusters, mainframes, set-top boxes ....)
... But that's the same mentaility you *should* take.
Then again, when I work on "mission critical systems", I'm talking about flight control hardware/software
OK, that I'll give you. Hell, usability in the professional software market isn't great either. I saw a program (for VLSI chip design) that crashed when you clicked on zoom in twice in a row. On a Sun, sometimes totally loging you out of the system, other times killing it off totally. For a $100K/user/year program. (Last year btw).
I hate to make you sadder, but I don't even really know what illustrator is (I basicly started reading the comments to see the ones on "wth does this program do"
... Should I be sadded that most people have no idea what I'm talking about when I rattle off the tools I use regularly (primarly digital hardware and printed circuit design)
And, btw, why does the fact that people don't know what Inkscape is sadden you? Maybe if this was an all art site
I wasn't comparing it to a CS class, I was comparing it to a Computer Engineering class. I've seen CS majors try to design hardware, and it wasn't pretty most of the time. In Comptuer Engineering, chip design or FPGA design is a lot of what we do.
Granted, it was an oversimplification, but I've seen groups that were a mix of grad students and undergrads crank out hardware designs a hell of a lot more complicated then a CPU and have a fairly good implementation. (Granted, mostly targeting FPGAs. One example I know a lot about was speech processing/reconigition.)
Uhh, as far as I understand most of them are Japenese? I had read (granted in a book from 1987) that the big American ones were pretty much being killed off.
Given that the above may be wrong, then they learned the hard way. If it's right, well, Japenese companies always took a longer term view. Not as innovataive though, it'd be nice to have management with a longer-term view + the US culture of encoraging radical ideas in R&D.
From the article:
... A bit more complex then the Hennasy and Patterson (The Classic!) book on computer arch and cpu design, but not by much. Given what (admittedly little) the article said about China's CPU fabs, I wonder if the newest Xilinx or Altera FPGA's could implement the design and run it faster ..
"The move still shows that China is capable of designing complex microprocessors."
As is any junior level class in computer engineering
Actaully, it's not for every game, just every game that charges to pay, for bigger LARPs its pretty common to chip in $1 to cover the costs of the people running the game.
Depends on the FPGA, you can get one from Xilinx in low quanity that'll run a soft-core processor for about $7 (Spartan-3 series, newest (90 nm) chip in their low cost/high volume line), in production quanity is a lot cheaper.
...
In response to the other posters, I have not seen much in terms of open source software for FPGA design, and 0 that worked for creating the hardware bitstream to dump into a chip. There are free versions of the tools avilable from Xilinx and Altera for their respective chips.
And, FPGAs are really useful as an implementation, not just for prototyping. It can cost up to $30M to set up an ASIC line, not counting engineering or the $100K+ for the tools
Only multi-million dollar projects? Bah :) ....
Look to C, Assembly, and (but deceasingly) Ada for the multi-billion dollar projects
It's simple, if you're a programmer, and you really are that good, you can do more with less memory and cpu cycles. I can write a program that runs on a $5 chip with 256 bytes of ram, and 16k flash, and have it controlling a robot. Go do that in java :)
Nothing can compare to having a dedicated test staff. At the last software place I worked (part-time, in school, while getting my engineering degree), we had 3-6 college students working on testing most of the time (we would also be given some small projects to work on).
Testing goes far beyond what any automated system can test, if you have a user in there somewhere. You also need to check things like "How easy is it to use?" and "Does this feature make sense?". We also suggested features that the program did not have, but from our experiance using it, thought that it should have.
I just want to say that I have no problem, in principle, with using open source for something like the RTOS in a flight control system, or some other hard real-time application.
My issue is with trying to make a hard RTOS out of a *nix base, I don't think it's the best approach to take. The only OSS RTOS I know of is XMK, which is nice, but only supports microcontrollers. Is there a true OSS counterpart for high end RTOS's yet?
The June issue had an article from GHS, and one from LynuxWorks. COTS Journal is published for developers (Systems, hardware, and software) of commerical off the shelf products, which is essentailyl a standard for military procurement where possible (You can get a computer, OS, etc, "off the shelf", but something like a sub or cruise missle is another story)
? id=100128
? id=100129
LynuxWorks: http://www.cotsjournalonline.com/home/article.php
GHS: http://www.cotsjournalonline.com/home/article.php
I never understood the point of his FUD, up until he started writing about this, I would not have considered Linux a viable alternative to Integrity (One of the RTOS's from GHS), for a varaity of reaons. ....
Keep in mind that this is not a destop, workstation, or server OS, this is hard real-time. You can't use something like Integrity for a desktop or server (well, it might be possible, but I doubt that it would perform as well as some *nix or even NT/2k/XP). And you shouldn't use something like Linux or Windows Super Server of the Year Edition for flight control. He actaully has me wondering how well the latest "real time" patches for Linux actaully perform
On a side note, I know a few people who work at places like Northrup and Lockheed, and from what I've heard, on a lot of projects, code is changed, documented, changes logged, and tested on almost a line by line basis (Obviously not counting a curly bracket or something like that as a seperate line). The versioning software keeps track of who changed what on what line, and when. That's not something you will find in Linux, any open source developers here up for that level of documentation? On the other hand, it is true that that's something you won't find from microsoft either.
Over long range, if you want to be able to hit a waypoint with any sort of accuracy, the easist way is to use GPS. Dead Reckoning requires updates based on landmarks, and a decent odomiter and compass as well. The alternative is to use an inertial system, a good one is pretty expensive. Both of the above also depend on knowing your starting position accuratly. Incremintal error *will* add up. There is a reason that the DoD created GPS.
I'm working on a robot navigation system for my Senior Design Project ( http://www.pitt.edu/~mmdst23 has the old details, I plan on fixing that, and adding some source to our Sourceforge site soon, but only after we get things working properly.)
We're using DR, but thats because it's better in an indoor enviroment, with the low cost of a Differential GPS receiver, there arn't many reasons to NOT use one, unless you're really paranoid about GPS jamming/spoofing (which is why the military GPS has an encrypted correction signal as well). That said, a backup DR system does appear to be needed, and using a compass is important anyway, GPS can give your corse over ground, which is where you're going, but it cant give your heading, which is where you're pointed at.
Read the title bar, it says it's VMware, so it could easily have been running under Linux.
It is a native port, but it hasn't been released yet.
At one time I thought that as well. Currently I'm working part time for a company that does some ship navagation software in testing (I can schedule around classes), and I can say that that model really dosen't work.
The big problem is that you don't have to buy the charts from us, we're compadable with other charts as well, and there are standards for electronic charts that many vendors follow (I'm not sure, but I'd assume that street-level charts have a simular standard). So in this case, the company making the software could end up not making money at all. The give the software away for free method works best where people/companies would pay for services, the only way this could work out for the company would be if the software somehow could only use the maps they sell, and they charged a lot more money for them.
The cheapest FPGA I know of is the Altera Educational board, you can get it for about $150 if you are a student (I'm not sure what else, you might have to go to one of the schools (colleges) that participate in their program.) For $150, you get suprisingly little performance, I'd be suprised if you could get a Mips 3k (one of the first mips cpu's, 32 bits and really slow), or an early Sparc cpu in one. I'd be really suprised if you could get a fast 486 or early pentium in one.
Well, to be honest, I didn't quite make that point earlyier. Sure, in many cases, a game thats pure eye candy is not as much fun as a game thats got a lot of depth, but older graphics. But I'd say both are important, if you make a game with the proper mix of good graphics and good gameplay, you have a classic. Look at a games like System Shock 2, Deus Ex, or Max Payne, more powerfull 3d can be good for gamers. Then again, look at all the crappy 3d shooters that came out, and it seems bad. The key here is to buy games from a place that lets you return them if they suck, or at least try the demo and read a review or 2 first.
Do you mean "most Windows users" or "most Windows gamers" ? Most gamers will have a decent 3d card, and most of the "major" cards (Nvidia, ATI, Matrox, 3dfx) will work, at least most of the time. The Kyro chips will probally have support
soon too.
As for a "low-quality" software rendering, what do you mean? Less detail, lower resolution? Its not really possible to have lower detail levels then whats in the game, and you can't run a game in less then the lowest resolution the game supports. Most of what you're looking for could only be done in the game's code itself, and to be honest, mean, cruel, blunt (whatever you want to call it), if you don't have a 3d card, you're pretty far off from the target audience for a 3d game. Games need to constantly be innovative and use new technology, and that needs more 3d power. Basicly, a 3d card thats 2 years old (or more) should be replaced if you want to play the latest 3d games. The good news is that the price of cards goes down as well, so you can get a still decent but 6 month old card for a lot less then what it would have cost new. If 3d is annoying you enough, you could (probally) find a decent 3d card for ~ $50-100, or a pretty good card for ~ $100-125.