The bullets were still dangerous at that range because they had not been fired at an angle steep enough to cause them to expend all their energy fighting gravity Fighting gravity uses up time, not energy (or at least, you get the energy back). If you fire at a high angle, the bullet spends enough time in the air that it gets slowed by air-friction. In the absence of an atmosphere, the bullet would hit with the full muzzle velocity no matter what angle it was fired at (unless the target is at a higher altitude than the gun).
No, MP3 only allows a predefined list of sampling frequencies, and the highest one on the list is 48kHz. You could use more bits per sample (it's internally floating-point), but most mp3 decoders I know of will only output 16 bits. Both of those issues are fixed if you use Vorbis instead.
You didn't misread the page, but the page is out of date. As of a few months ago, there is an open source implementation of WMV3 aka VC1 in libavcodec. WMA2 has been implemented for a while. That video plays just fine on my 64bit linux without Microsoft's binary codecs.
Who said anything about transcoding? Just copy the dvd, bit for bit. The GP's point is that you don't need the permission of the media store program to burn multiple copies: if you can burn one copy, then any generic cd/dvd duplicator can do the rest.
That just amounts to, "make sure someone else is on the bottom rung of the pyramid scheme". The fact remains that not everyone can have a ratio >1. If your tracker requires ratio >1, then the only possible steady state is to have a churn of accounts getting banned -- if there are leechers, they will be among the banned people, but the system will guarantee that some people get banned whether or not they anyone is trying to leech. Luckily at least some tracker administrators understand that, and the sites I use typically have the hard threshold around 0.5, not 1.
The Ogg bitstream is also hated by just about everyone who knows anything about media file formats. If you're going to propose a new standard Free media format, please use something like Matroska or NUT instead. And the answer to your question is: no, Snow and Dirac can't be easily inserted into Ogg. Ogg's inability to handle arbitrary codecs is one of the reasons I don't recommend Ogg.
Signed drivers would be ok if I get to decide whose signatures to trust. What's wrong is when that decision is reserved for Microsoft, thus taking freedom away from the user.
Right. There are already conventional key exchange algorithms that are as good as conventional encryption. If you're worried enough about conventional algorithms being broken to switch to quantum, then you have to use it for the data transfer too, not just for key exchange. (Unless your conventional algorithm is a one-time pad, in which case quantumly sending the key is enough. Which is actually how the standard formulation of QC works. But in terms of bandwidth usage, that's the same as sending the message over the quantum channel.)
QC requires that you have a direct fiber-optic connection between the two parties. In short, what it provides is a guarantee that you'll notice if someone splices into your fiber, and gives you a chance to abort communication before the eavesdropper gains any information.
If you use a dedicated line without the quantum part, that's proof against random people on the internet, but someone with a shovel and various optic hardware could still snoop on it (and then maybe break the conventional encryption). Too paranoid for you? sure, but some businesses/governments might not want to take even that risk.
Then again, it underclocks to 800MHz when unplugged because the manufacturer said so. What does the manufacturer have to do with anything? I also have a 2.2GHz A64 3400+ laptop, and mine underclocks to 800MHz when unplugged because/etc/cpufreqd.conf says so. But if I want a speed boost at some cost in battery life, I can change that.
Ok, I wrote some code to benchmark various versions: http://akuvian.org/src/invsqrt.tar.gz. invsqrt_q3 is the version from the article. It gets miscompiled unless I add -fno-strict-aliasing. invsqrt_union is the same except with a union instead of pointer casts. This is the more correct way to write it in C, and fixes the aliasing issue. invsqrt_c is the trivial 1.0/sqrtf(x). For the benchmark, all of these were inlined inside a loop, so the SSE and 3DNOW versions do not include any time spent moving the result to or from the general floating-point registers.
athlon64:
30.15 cycles in invsqrt_c
34.02 cycles in invsqrt_q3
34.02 cycles in invsqrt_union
3.02 cycles in invsqrt_sse
3.02 cycles in invsqrt_3dnow
pentium4:
46.21 cycles in invsqrt_c
70.03 cycles in invsqrt_q3
84.90 cycles in invsqrt_union
4.06 cycles in invsqrt_sse
So I'd say that, no, the hack is no longer a win on modern processors.
Clearer? I don't think so, but that's just an opinion. Faster? Not on x86. Integers and floats use different registers, and it's not possible to convert one to the other without going through memory anyway. (SSE can do it, but then SSE has a single instruction for InvSqrt so it wouldn't need the function at all.)
A 64 bit architecture wouldn't necessarily harm anything. The only non-portable assumptions made by the code are: int is 32 bits, and float is stored in IEEE-754 format. The first assumption is still true in 64 bit gcc, and even if it isn't you can just write int32_t instead.
Kind of like how it takes less CPU power to decode a WAV file than to decode an MP3 file. On a decent computer, decoding an MP3 takes some fraction of a percent of the cpu (here: Athlon64 2.4GHz, 0.4%). It is in fact faster to read and decode an MP3 than it is to read a WAV off my harddrive. That said, I understand the point you were trying to make, and I can't say which way the tradeoff goes for the procedural graphics being discussed.
Do you realize that an average 300+ watt machine running 24/7 costs you about $15-$20 a month in electricity?
300W * 24h * 30 days = 216 kWh / month. My electric bill is.059$/kWh, which would make a total of 12.74$/month. Except that my laptop draws a max of 90W and idles at 30W, and I leave it on 24/7 for convenience and various background processes. So the difference between running Folding@home vs leaving it mostly idle would cost me all of 2.55$/month.
Since the criteria for entry say that any new submission must beat the current record, it's no surprise that only 3 people are listed. You're not seeing any of the people who didn't win.
If you use LAME, set your Q to 9. A 320kbps MP3 with Q=1 and 320kbps mp3 with Q=9 are WILDLY different, while both the same bitrate and same size.
Huh? Lame's manpage says that -q 0 is the slowest and highest quality, while -q 9 is the fastest and lowest quality. Do the win32 frontends remap that range or something?
There is a serious ambiguity - if that person adds or changes some of the code, can that person convert the entire module to V3 or is the module now fragmented? And if the latter, how is anyone supposed to keep track of what statement of code is under what license?
Good question, but the issue is not new to GPLv3. The same question arises if you want to combine some GPLv2 code with some LGPLv2 code into one program. They're certainly compatible, and the compiled program as a whole is covered by GPL. But are you obliged to maintain notices on each piece of code saying that this part is GPL and that part is LGPL, in case someone wants to extract the LGPL pieces for their own LGPL program?
FWIW, I have done this. And what I chose to do (without knowing whether it's legal) was to keep an LGPL notice on any whole files I imported, but for smaller chunks of LGPL code that I pasted into an otherwise GPL file, I just creditted the author and didn't specify any difference in license.
Either your computer is really slow, or you're using a small block size. The time needed to generate the parity files is proportional to (size of data files) * (size of parity files) / (parity block size). I generally use a block size of 1 MiB if I'm filling a DVD with 90% data and 10% PAR2, and it's not ridiculously slow.
Flash write cycles are limited per block, not for the whole drive. Do you overwrite your whole harddrive once every 5 seconds? It's not even capable of writing that fast.
Consider a 100GB drive, writing continuously at 50MB/s for 5 years. With perfect defect management, that would only use up 80,000 write cycles.
The GPL has an answer for that: "The source code for a work means the preferred form of the work for making modifications to it."
There's no special exemption just because final product you distribute can be called "data" rather than "code". It still has a source. For example, I could GPL a TeX document, and then if someone wanted to modify and redistribute it, they would have to distribute the TeX form, and not the rendered postscipt/pdf/dvi form.
And how would that differ from any other trojan? Sure, you'd want to infect something that's supposed to connect to the internet, so as to avoid outbound firewalls. But I don't see how VOIP software makes any better target than, say, a web browser.
Civ IV does come with a poster insert of the tech tree, as did Civ III and Civ II That's not the point. Of course players can know the tech tree, whether it comes in the manual or they have to look it up on the web. But it have an option like in Alpha Centauri, where you don't research "laser guns", you research "military tech" and have a chance of getting lasers, or a better jet engine, or whatever.
Yes, the notice does say that the results were censored.
But that's not the only explanation for the difference: All localized versions of Google always bias results to display the local sites first. Tiananmen is a place, and a massacre isn't the only thing that ever happened there. It stands to reason that a majority of Chinese sites would mention it for other reasons.
Slashdot Mirror
No, MP3 only allows a predefined list of sampling frequencies, and the highest one on the list is 48kHz.
You could use more bits per sample (it's internally floating-point), but most mp3 decoders I know of will only output 16 bits.
Both of those issues are fixed if you use Vorbis instead.
You didn't misread the page, but the page is out of date. As of a few months ago, there is an open source implementation of WMV3 aka VC1 in libavcodec. WMA2 has been implemented for a while. That video plays just fine on my 64bit linux without Microsoft's binary codecs.
Who said anything about transcoding? Just copy the dvd, bit for bit. The GP's point is that you don't need the permission of the media store program to burn multiple copies: if you can burn one copy, then any generic cd/dvd duplicator can do the rest.
That just amounts to, "make sure someone else is on the bottom rung of the pyramid scheme".
The fact remains that not everyone can have a ratio >1. If your tracker requires ratio >1, then the only possible steady state is to have a churn of accounts getting banned -- if there are leechers, they will be among the banned people, but the system will guarantee that some people get banned whether or not they anyone is trying to leech.
Luckily at least some tracker administrators understand that, and the sites I use typically have the hard threshold around 0.5, not 1.
The Ogg bitstream is also hated by just about everyone who knows anything about media file formats. If you're going to propose a new standard Free media format, please use something like Matroska or NUT instead.
And the answer to your question is: no, Snow and Dirac can't be easily inserted into Ogg. Ogg's inability to handle arbitrary codecs is one of the reasons I don't recommend Ogg.
Signed drivers would be ok if I get to decide whose signatures to trust. What's wrong is when that decision is reserved for Microsoft, thus taking freedom away from the user.
Right. There are already conventional key exchange algorithms that are as good as conventional encryption. If you're worried enough about conventional algorithms being broken to switch to quantum, then you have to use it for the data transfer too, not just for key exchange. (Unless your conventional algorithm is a one-time pad, in which case quantumly sending the key is enough. Which is actually how the standard formulation of QC works. But in terms of bandwidth usage, that's the same as sending the message over the quantum channel.)
QC requires that you have a direct fiber-optic connection between the two parties. In short, what it provides is a guarantee that you'll notice if someone splices into your fiber, and gives you a chance to abort communication before the eavesdropper gains any information.
If you use a dedicated line without the quantum part, that's proof against random people on the internet, but someone with a shovel and various optic hardware could still snoop on it (and then maybe break the conventional encryption). Too paranoid for you? sure, but some businesses/governments might not want to take even that risk.
(yes, I am a quantum physicist)
invsqrt_q3 is the version from the article. It gets miscompiled unless I add -fno-strict-aliasing.
invsqrt_union is the same except with a union instead of pointer casts. This is the more correct way to write it in C, and fixes the aliasing issue.
invsqrt_c is the trivial 1.0/sqrtf(x).
For the benchmark, all of these were inlined inside a loop, so the SSE and 3DNOW versions do not include any time spent moving the result to or from the general floating-point registers.
So I'd say that, no, the hack is no longer a win on modern processors.
Clearer? I don't think so, but that's just an opinion.
Faster? Not on x86. Integers and floats use different registers, and it's not possible to convert one to the other without going through memory anyway. (SSE can do it, but then SSE has a single instruction for InvSqrt so it wouldn't need the function at all.)
A 64 bit architecture wouldn't necessarily harm anything.
The only non-portable assumptions made by the code are: int is 32 bits, and float is stored in IEEE-754 format.
The first assumption is still true in 64 bit gcc, and even if it isn't you can just write int32_t instead.
Kind of like how it takes less CPU power to decode a WAV file than to decode an MP3 file.
On a decent computer, decoding an MP3 takes some fraction of a percent of the cpu (here: Athlon64 2.4GHz, 0.4%). It is in fact faster to read and decode an MP3 than it is to read a WAV off my harddrive.
That said, I understand the point you were trying to make, and I can't say which way the tradeoff goes for the procedural graphics being discussed.
Do you realize that an average 300+ watt machine running 24/7 costs you about $15-$20 a month in electricity?
.059$/kWh, which would make a total of 12.74$/month.
300W * 24h * 30 days = 216 kWh / month. My electric bill is
Except that my laptop draws a max of 90W and idles at 30W, and I leave it on 24/7 for convenience and various background processes. So the difference between running Folding@home vs leaving it mostly idle would cost me all of 2.55$/month.
Since the criteria for entry say that any new submission must beat the current record, it's no surprise that only 3 people are listed. You're not seeing any of the people who didn't win.
If you use LAME, set your Q to 9. A 320kbps MP3 with Q=1 and 320kbps mp3 with Q=9 are WILDLY different, while both the same bitrate and same size.
Huh? Lame's manpage says that -q 0 is the slowest and highest quality, while -q 9 is the fastest and lowest quality. Do the win32 frontends remap that range or something?
More likely "ee" vs "iie", though I seem to remember there was a pair even closer than that.
Or just consider the english equivalent: "u-huh" vs "uh-uh". Imo, those fit the bill of meaning "yes" and "no" and differing only in inflection.
Good question, but the issue is not new to GPLv3.
The same question arises if you want to combine some GPLv2 code with some LGPLv2 code into one program. They're certainly compatible, and the compiled program as a whole is covered by GPL. But are you obliged to maintain notices on each piece of code saying that this part is GPL and that part is LGPL, in case someone wants to extract the LGPL pieces for their own LGPL program?
FWIW, I have done this. And what I chose to do (without knowing whether it's legal) was to keep an LGPL notice on any whole files I imported, but for smaller chunks of LGPL code that I pasted into an otherwise GPL file, I just creditted the author and didn't specify any difference in license.
Either your computer is really slow, or you're using a small block size. The time needed to generate the parity files is proportional to (size of data files) * (size of parity files) / (parity block size).
I generally use a block size of 1 MiB if I'm filling a DVD with 90% data and 10% PAR2, and it's not ridiculously slow.
Flash write cycles are limited per block, not for the whole drive. Do you overwrite your whole harddrive once every 5 seconds? It's not even capable of writing that fast. Consider a 100GB drive, writing continuously at 50MB/s for 5 years. With perfect defect management, that would only use up 80,000 write cycles.
The GPL has an answer for that: "The source code for a work means the preferred form of the work for making modifications to it."
There's no special exemption just because final product you distribute can be called "data" rather than "code". It still has a source. For example, I could GPL a TeX document, and then if someone wanted to modify and redistribute it, they would have to distribute the TeX form, and not the rendered postscipt/pdf/dvi form.
And how would that differ from any other trojan?
Sure, you'd want to infect something that's supposed to connect to the internet, so as to avoid outbound firewalls. But I don't see how VOIP software makes any better target than, say, a web browser.
Civ IV does come with a poster insert of the tech tree, as did Civ III and Civ II
That's not the point. Of course players can know the tech tree, whether it comes in the manual or they have to look it up on the web.
But it have an option like in Alpha Centauri, where you don't research "laser guns", you research "military tech" and have a chance of getting lasers, or a better jet engine, or whatever.
Can someone say whether this chinese page has a similar notice? http://images.google.cn/images?q=tiananmen
whose results are dramatically different than results for the same search in the USA: http://images.google.com/images?q=tiananmen
Yes, the notice does say that the results were censored.
But that's not the only explanation for the difference:
All localized versions of Google always bias results to display the local sites first. Tiananmen is a place, and a massacre isn't the only thing that ever happened there. It stands to reason that a majority of Chinese sites would mention it for other reasons.
Yes, DivX and XviD are MPEG-4. But they use Avdanced Simple Profile whereas the iPod only supports Simple Profile.