32-bit floating could be a valid issue (if you need more accuracy), but I don't see how IEEE-754 comes into that considering that even on current CPUs, it's not possible to have two FFTs (or even the same FFT compiled with different options) give exactly the same result. In any case, you've got rounding that's decided based on what goes in registers vs. memory. So, as long as the precision in the GPU is about the same as the 32-bit IEEE spec, I don't think anyone can really see the difference, especially considering that an FFT cannot "create" NaNs/Infs if there weren't any in the input.
See, the thing is you shouldn't have to go to a "top university" to get a decent education. I was also comparing to my own university, which was public and wasn't considered a "top university" either.
I did one semester at a US university (UConn) as part of an echange program. That was the only time in my life I felt I was increasing the class average just by myself. Funny thing is I've heard a few people there tell me "oh, you're going to the US to get a good education...". Yeah right!
Remember they were discussing differences in file formats
I do understand that. However, for a file format to be really useful, it must have good application support. So far, the app that supports it best is OO.o, but it still needs to be improved a lot (which is what I was trying to point out).
Can you clarify what you mean by [sound is] "buggy?" Or at least give examples?
I mean that 1) it's a bit of a pain to insert sounds in the first place and 2) On both my machines, sounds don't even work when I'm in presentation mode (have to go back to editor mode).
I think this [video] requrires Java, but I do know you can use video in OOo impress.
I've tried it. Gave up on it after spending an hour trying to get it to work.
OK, maybe I didn't use the right comparison and the Word equation editor sucks more than I thought. Still, I've been using LyX since 1997 and even back then (version 0.10.x IIRC), the equation editor was way ahead of the one in OO.o. It just seems to be that none of the developers actually care about it, which is a bit sad.
I actually do use LyX as a LaTeX frontend and I wouldn't switch to OO.o even if the equation editor was good. However, I use OO.o for the graphics because it's just so much faster than xfig and tgif. Problem is when I need equations in there -- the OO math editor is just awful. I haven't used the MS Word one either, but I don't think it's nearly as bad. BTW, I always find it funny that you can recognize a paper written with Word just because the equations don't look "quite right".
There are many areas where OO.o is still lacking badly. One of them is the math editor. I still think writing a scientific document in MS Word (and not LaTeX) is a dumb idea in the first place, but OO.o is far worse than even MS Word. Support for sound (works but buggy) and video (inexistant AFAIK) in presentations is another example. These are "just" implementation issues (not ODF-related), they will need to be fixed if OO is to compete with MS Office at the feature level (I refuse to us MS Office because the format is closed, but not everyone care about that).
At the same time though... this does conceivably give more power to Redmond as there is now less incentive for MA to leave the Windows/Office platform.
Not necessarily. I may also make the transition easier. First everyone just save to ODF, then the switch is easy to make because you don't need to get everyone to make the transition at the same time.
It's may not be bold, but at least it's realistic. The industry has shown they can do without Mercury and Cadmium, now they'll have to remove them from everything. It would be useless to say "no oil in cars anymore" if there isn't a real (practical, proven) alternative. However, once (e.g.) half the cars are conferted to cleaner stuff, they *could* do such a thing (not saying they will).
I wouldn't be so sure about that. Assume for a moment the F-150 is 3x heavier (don't know if it is, but the reasoning is the same). In the separate crash against a wall, the F-150 takes exactly 3x the impact energy that the Mini takes. Now, if they go head-on, they will in the best case split the impact energy in two (in the worst case, the Mini will take more). In other words, in a head-on, the F-150 takes about 2/3 of what it takes when hitting a wall, while the Mini takes twice as much. That's a *big* difference. It's like comparing an F-150 crashing in a wall at 81 km/h to a Mini crashing at 141 km/h. *This* is why I think these heavy vehicles are a danger to other people.
The went head-on one with the other or they hit a wall separately. There's a big difference. With a wall, I'd vote for the small, well-built car. With a collision, the lightest car takes most of the impact, which is why heavier cars/trucks are dangerous. I don't care if you kill yourself in your new Hummer. I do care if you kill me (or others) in collisions.
Most of the SUVs on the road aren't that bad. The full-sized SUVs, built on truck frames, are the ones to watch out for. My 2004 Durango is full-sized, and heavy, but built around collapsible tubes, rather than rigid I-beams, which is why it has a five-star crash rating.
I guess I should have been more precise... SUVs are a real danger for others on the roads.
Not quite. While I agree that this class of vehicle is more dangerous in itself (not colliding with other cars), the thing is that in a collision between a Hummer and a car, the car will still suffer most of the damage. Vehicles like that and even "regular" SUVs are a real danger on the roads. The whole thing has become a sort of arms race about whoever can feel the safest by knowing they'll "win" in the event of a collision.
I'd be terrified of being smushed by a truck while driving one.
Which is why every American feels the need to own a car that's bigger than the median car on the road. I say seatbelts should be forbidden on SUVs, while Hummer-type vehicles should be required to have a blade coming out of the steering wheel. Now, *that* would even the chance in case of a crash.
I think it depends more on the actual card than on the manufacturer. I've got an ATI card in my (Dell D600) laptop that works fine with only open-source drivers and I even get accelerated 3D. It's probably not as fast as the close-source ATI driver, but way enough for my needs (I never actually use the 3D stuff).
The only sensible definition for a planet is an object that is spherical due to its own gravity, orbits a star, and is not itself a star
OK, so any mass of liquified gas orgiting the sun would qualify, right? How about comets (if they are spherical)?. Oh and how about we say the moon is orbiting the sun and the Earth is just going around it?
Actually, if you want to limit the range of that laser, all you would have to do is make the short-term trajectory unpredictable. It would probably take several milli-seconds (if not more) to fire the laser so as long as you can change trajectory (by a meter or so) faster than that, you're OK.
Slashdot Mirror
32-bit floating could be a valid issue (if you need more accuracy), but I don't see how IEEE-754 comes into that considering that even on current CPUs, it's not possible to have two FFTs (or even the same FFT compiled with different options) give exactly the same result. In any case, you've got rounding that's decided based on what goes in registers vs. memory. So, as long as the precision in the GPU is about the same as the 32-bit IEEE spec, I don't think anyone can really see the difference, especially considering that an FFT cannot "create" NaNs/Infs if there weren't any in the input.
See, the thing is you shouldn't have to go to a "top university" to get a decent education. I was also comparing to my own university, which was public and wasn't considered a "top university" either.
I did one semester at a US university (UConn) as part of an echange program. That was the only time in my life I felt I was increasing the class average just by myself. Funny thing is I've heard a few people there tell me "oh, you're going to the US to get a good education...". Yeah right!
Remember they were discussing differences in file formats
I do understand that. However, for a file format to be really useful, it must have good application support. So far, the app that supports it best is OO.o, but it still needs to be improved a lot (which is what I was trying to point out).
Can you clarify what you mean by [sound is] "buggy?" Or at least give examples?
I mean that 1) it's a bit of a pain to insert sounds in the first place and 2) On both my machines, sounds don't even work when I'm in presentation mode (have to go back to editor mode).
I think this [video] requrires Java, but I do know you can use video in OOo impress.
I've tried it. Gave up on it after spending an hour trying to get it to work.
OK, maybe I didn't use the right comparison and the Word equation editor sucks more than I thought. Still, I've been using LyX since 1997 and even back then (version 0.10.x IIRC), the equation editor was way ahead of the one in OO.o. It just seems to be that none of the developers actually care about it, which is a bit sad.
I actually do use LyX as a LaTeX frontend and I wouldn't switch to OO.o even if the equation editor was good. However, I use OO.o for the graphics because it's just so much faster than xfig and tgif. Problem is when I need equations in there -- the OO math editor is just awful. I haven't used the MS Word one either, but I don't think it's nearly as bad. BTW, I always find it funny that you can recognize a paper written with Word just because the equations don't look "quite right".
There are many areas where OO.o is still lacking badly. One of them is the math editor. I still think writing a scientific document in MS Word (and not LaTeX) is a dumb idea in the first place, but OO.o is far worse than even MS Word. Support for sound (works but buggy) and video (inexistant AFAIK) in presentations is another example. These are "just" implementation issues (not ODF-related), they will need to be fixed if OO is to compete with MS Office at the feature level (I refuse to us MS Office because the format is closed, but not everyone care about that).
At the same time though... this does conceivably give more power to Redmond as there is now less incentive for MA to leave the Windows/Office platform.
Not necessarily. I may also make the transition easier. First everyone just save to ODF, then the switch is easy to make because you don't need to get everyone to make the transition at the same time.
It's may not be bold, but at least it's realistic. The industry has shown they can do without Mercury and Cadmium, now they'll have to remove them from everything. It would be useless to say "no oil in cars anymore" if there isn't a real (practical, proven) alternative. However, once (e.g.) half the cars are conferted to cleaner stuff, they *could* do such a thing (not saying they will).
So what you're saying is that you'd rather be treated for cancer in the US than not having cancer in the first place?
I wouldn't be so sure about that. Assume for a moment the F-150 is 3x heavier (don't know if it is, but the reasoning is the same). In the separate crash against a wall, the F-150 takes exactly 3x the impact energy that the Mini takes. Now, if they go head-on, they will in the best case split the impact energy in two (in the worst case, the Mini will take more). In other words, in a head-on, the F-150 takes about 2/3 of what it takes when hitting a wall, while the Mini takes twice as much. That's a *big* difference. It's like comparing an F-150 crashing in a wall at 81 km/h to a Mini crashing at 141 km/h. *This* is why I think these heavy vehicles are a danger to other people.
The went head-on one with the other or they hit a wall separately. There's a big difference. With a wall, I'd vote for the small, well-built car. With a collision, the lightest car takes most of the impact, which is why heavier cars/trucks are dangerous. I don't care if you kill yourself in your new Hummer. I do care if you kill me (or others) in collisions.
Most of the SUVs on the road aren't that bad. The full-sized SUVs, built on truck frames, are the ones to watch out for. My 2004 Durango is full-sized, and heavy, but built around collapsible tubes, rather than rigid I-beams, which is why it has a five-star crash rating.
I guess I should have been more precise... SUVs are a real danger for others on the roads.
That is already installed.
Not quite. While I agree that this class of vehicle is more dangerous in itself (not colliding with other cars), the thing is that in a collision between a Hummer and a car, the car will still suffer most of the damage. Vehicles like that and even "regular" SUVs are a real danger on the roads. The whole thing has become a sort of arms race about whoever can feel the safest by knowing they'll "win" in the event of a collision.
I'd be terrified of being smushed by a truck while driving one.
Which is why every American feels the need to own a car that's bigger than the median car on the road. I say seatbelts should be forbidden on SUVs, while Hummer-type vehicles should be required to have a blade coming out of the steering wheel. Now, *that* would even the chance in case of a crash.
Operating System dug deep into new computers being sold
You mean having all your OS buffer overflows built in the hardware?
My guess is that they would mean crashing the simulation or crashing the model, in other words: "And the answer is... NaN".
It's more about having someone to blame than actually extracting money out of them.
:-)
Got no problem, I always blame Linus anyway
Ask for support != sue. You can ask your Linux distro vendor for support too. I have yet to see any successful lawsuit over a Windows fault.
Just curious, who do you usually sue when Windows breaks?
I think it depends more on the actual card than on the manufacturer. I've got an ATI card in my (Dell D600) laptop that works fine with only open-source drivers and I even get accelerated 3D. It's probably not as fast as the close-source ATI driver, but way enough for my needs (I never actually use the 3D stuff).
The only sensible definition for a planet is an object that is spherical due to its own gravity, orbits a star, and is not itself a star
OK, so any mass of liquified gas orgiting the sun would qualify, right? How about comets (if they are spherical)?. Oh and how about we say the moon is orbiting the sun and the Earth is just going around it?
Actually, if you want to limit the range of that laser, all you would have to do is make the short-term trajectory unpredictable. It would probably take several milli-seconds (if not more) to fire the laser so as long as you can change trajectory (by a meter or so) faster than that, you're OK.
It's OK, they have a label stating "for internal use only".
Of course, there's a justice department. There's also a Ministry of Truth, a Ministry of Love, a The Ministry of Peace and a The Ministry of Plenty.