Times are a-changin'. The syntactically sensible style has gradually been taking hold over the last half century.
First news site I looked at, top story: http://www.guardian.co.uk/prisons/story/0,,1785585 ,00.html
2nd website I looked at, top story: http://news.bbc.co.uk/2/hi/europe/5028918.stm
Both extensively use ``".'' unless they are quoting full sentences, in which case all of that sentence's markup is included within the quotes including the leading capital letter. (Literally quoting the whole sentence, again syntactically sensible, but missing a ``.'' to terminate its own sentence in this case, so not pedantic enough to satisfy a LISP programmer like me.)
Both of the above are less sensible about ``",'', prefering ``,"'', which proves that the style guides haven't evolved quite enough yet.
Of course, you could ditch FP, and use a NTT (Number-Theoretic Transform), where the limbs are integers modulo some suitably-chosen prime and not atomic (don't quote me on this, but a number like 2^96-2^32+1 is sometimes used). The FPU unit can still be used to compute integer maths, of course, by restricting the range of the numbers used. Modular reduction with such a prime is simply a task of moving numbers between columns, effectively free.
I have invented a hybrid technique which permits unusually large bases for the limbs (I think I reached 7*10^12), at very little extra cost. I've yet to publish a paper describing it, but ought to. My prime-hunting projects, PIES and Les GeneFermiers, use this algorithm extensively, and are quite successful.
However, at the end of the day good old-fashioned complex FP FFTs seem to be fastest implementations, but I think that's mostly due to CPU manufacturers dedicating more transistors to such operations. 20 years ago, with only an external 8087 FPU, I'm sure it wouldn't have been the case.
FFT a recursive implementation of the DFT. Used for performing convolutions of two length n signals in Theta(n log(n) loglog(n)) time.
Bignum multiplication is convolution with carry propagation - a lot of numerical distributed computing tasks (such as GIMPS) are based around fast large FFTs.
Some hardware/system companies have a small bunch of volunteers for this very task - firstly they select programmers which they believe have l33t programming skills, then they lend you a top-of-the-range model (and even are prepared to ship it all the way to Finland if the volunteer lives in Finland), and in return you promise to work on hand-optimising code for their platform, and publishing the results.
You do realise that that page is written by a crank, don't you? Just check out his Usenet history for examples. And check out other things on his website, such as: http://www.rebelscience.org/Crackpots/notorious.ht m
Mars Rover, mars schmover. For real quality code, you want to look at NASA's stuff from the 60s/70s. Voyager being possibly the ultimate in keeping going no matter what (very useful fallback behaviour when components fail, for example). And I believe they did release the source to that. I think my g/f said that she had seen a copy a few years back. (But maybe she's a leet haxor dudette, and snarfed it clandestinely?!)
If you've ever forwarded an image file to a friend who might forward it to other people, then you are a potential vector for malware. Sure it may look like a picture of a carrot that looks like Tom Hanks, but if it causes a buffer overrun that installs a rootkit, and one of the friends-of-a-friend wants to 'photoshop' out the logo in the corner, then someone's getting as 0wned as if they clicked "yes" after downloading an executable.
The moment you say that security doesn't matter in on place, that becomes the ideal place for attacks to be focussed.
And it's time to start re-labelling those insults. AOL's off Usenet, Something just as bad - googlegroups - is on.
Except the "Me too", or "LOL!!!", is now posted without _any_ context at all so you don't even know what it's a reply to, rather than being pasted below the whole unedited prior conversation.
""" Sinclair had commissioned GST Computer Systems to produce an operating system for the machine, but switched to QDOS, developed in-house, before launch. GST's OS, designed by Tim Ward,... """
Most Macs I encountered in those days (e.g. in art depts of universities) came furnished with a Zip drive. And someone to tell you how superior they were, of course.
Degrees of freedom are a mathematical concept. If you're that unaware of the tools without upon which science and engineering are based, then you can't be much of either.
However, like all anonymous cowards you can't be either, as you're a nobody.
My emphasis should have been more on "public", sorry. Deals with governments to supply things to schools, isn't public in the same way that you and I buying stuff from cdbaby is.
I'm not sure how they (anyone) expects to stop these things being stolen more than copper and aluminium cable are. And if they do have a value in the developed world, then the black market trade in them could be quite horrendous.
Oh come on - I extracted the sentence with the blooper in - how much more trimming did you want me to do?/Degrees of freedom/ map approximately onto dimensions, or coordinates. To specify a hinge position, you only need to specify one coordinate - an angle; hence it has one degree of freedom.
Monitors with a fixed base on the desk would have 2 degrees of freedom - a left/right angle, and an up/down angle. If it can move on the desktop it has 2 more degrees of freedom ((x,y) coordinates). If the desktop can be raised or lowered, it has an additional 5th degree of freedom.
The most a rigid body can have is 6 degrees of freedom - 3 position, and 3 direction. The one that the above monitor lacks is the ability to rotate in-plane - flipping from landscape to portrait mode, for example.
An anglepoise lamp with a fixed base will typically have 2 (base) + 1 (elbow) + 2 (wrist) degrees of freedom. More join joints means more degrees of freedom.
Well, the fact that they don't intend to sell them to the public at all should counter most of your worries. This pledge drive is intending to get them to change that. However, given that pledges are currently coming in from people like "I. P. Freely", I guess it'll probably get ignored.
On the 'worldchanging' link on the pledge page, quoth Ethan Zuckerman: """ I wonder if the hinges are going to be a problem - the current design requires a hinge for the gasket and a separate hinge that allows 340 degrees of freedom between the screen and the keyboard. """
Yeah, right. How about you learn what engineering terms mean before you use them in _completely the wrong way_.
Sheesh, and such blog journalism is the future, eh? ${DEITY} help us. FatPhil
RTFA: """ The suggestion has been made that he also offer it for sale for ~$300 to the rest of us so that we do have an interesting macnine and can help to support the cost computers for the developing world. """
The evaluation module comes with an 8-bit microcontroller and a UART. You can therefore trivially connect it to any machine with a UART, be that a PC, linux box, or maybe even a Mac (do they still have UARTs?). What you do with the readings is then entirely up to you. However, the raw data it sends you is self-explanatory {x,y,z} data, and it shouldn't be too hard to tell a slash from a jab.
"They are accelerometers, meaning they can be used to detect relative motion."
And here we have proof why the mislabelling as motion sensors is so bad. You probably knew what you wanted to say, that I don't doubt, but the above is just plain wrong. Changes in motion, yes (dv/dt); but not differences in motion (v1-v2).
This is the 2000s, I'd hope it was just a single multi-axis accelerometer.
Freescale has some funky electron micrographs in their accelerometer literature, and they're so detailed you can see precisely how they work. Works of art!
Slashdot Mirror
Times are a-changin'. The syntactically sensible style has gradually been taking hold over the last half century.
5 ,00.html
First news site I looked at, top story:
http://www.guardian.co.uk/prisons/story/0,,178558
2nd website I looked at, top story:
http://news.bbc.co.uk/2/hi/europe/5028918.stm
Both extensively use ``".'' unless they are quoting full sentences, in which case all of that sentence's markup is included within the quotes including the leading capital letter. (Literally quoting the whole sentence, again syntactically sensible, but missing a ``.'' to terminate its own sentence in this case, so not pedantic enough to satisfy a LISP programmer like me.)
Both of the above are less sensible about ``",'', prefering ``,"'', which proves that the style guides haven't evolved quite enough yet.
FatPhil
Of course, you could ditch FP, and use a NTT (Number-Theoretic Transform), where the limbs are integers modulo some suitably-chosen prime and not atomic (don't quote me on this, but a number like 2^96-2^32+1 is sometimes used). The FPU unit can still be used to compute integer maths, of course, by restricting the range of the numbers used. Modular reduction with such a prime is simply a task of moving numbers between columns, effectively free.
I have invented a hybrid technique which permits unusually large bases for the limbs (I think I reached 7*10^12), at very little extra cost. I've yet to publish a paper describing it, but ought to. My prime-hunting projects, PIES and Les GeneFermiers, use this algorithm extensively, and are quite successful.
However, at the end of the day good old-fashioned complex FP FFTs seem to be fastest implementations, but I think that's mostly due to CPU manufacturers dedicating more transistors to such operations. 20 years ago, with only an external 8087 FPU, I'm sure it wouldn't have been the case.
FatPhil
FFT a recursive implementation of the DFT.
Used for performing convolutions of two length n signals in Theta(n log(n) loglog(n)) time.
Bignum multiplication is convolution with carry propagation - a lot of numerical distributed computing tasks (such as GIMPS) are based around fast large FFTs.
FatPhil
Some hardware/system companies have a small bunch of volunteers for this very task - firstly they select programmers which they believe have l33t programming skills, then they lend you a top-of-the-range model (and even are prepared to ship it all the way to Finland if the volunteer lives in Finland), and in return you promise to work on hand-optimising code for their platform, and publishing the results.
;-)
FatPhil, in Finland.
Statistically errors are a random walk. Distance=sqrt(#steps).
So you have 13.5 bits of error on average rather than 27.
I've just worked out that I do 1 billion 8192-limb double-precision complex FFTs per day!
You do realise that that page is written by a crank, don't you?t m
Just check out his Usenet history for examples.
And check out other things on his website, such as:
http://www.rebelscience.org/Crackpots/notorious.h
You're a loon Louis Savain.
A Brit would tend to end the sentence with the full stop, i.e. ". not ."
FatPhil
Mars Rover, mars schmover. For real quality code, you want to look at NASA's stuff from the 60s/70s. Voyager being possibly the ultimate in keeping going no matter what (very useful fallback behaviour when components fail, for example).
And I believe they did release the source to that. I think my g/f said that she had seen a copy a few years back. (But maybe she's a leet haxor dudette, and snarfed it clandestinely?!)
FatPhil
A true Brit would /never/ use 'saber'.
If you've ever forwarded an image file to a friend who might forward it to other people, then you are a potential vector for malware. Sure it may look like a picture of a carrot that looks like Tom Hanks, but if it causes a buffer overrun that installs a rootkit, and one of the friends-of-a-friend wants to 'photoshop' out the logo in the corner, then someone's getting as 0wned as if they clicked "yes" after downloading an executable.
The moment you say that security doesn't matter in on place, that becomes the ideal place for attacks to be focussed.
FatPhil
And it's time to start re-labelling those insults.
AOL's off Usenet, Something just as bad - googlegroups - is on.
Except the "Me too", or "LOL!!!", is now posted without _any_ context at all so you don't even know what it's a reply to, rather than being pasted below the whole unedited prior conversation.
Should you not be declaring an interest?
...
"""
Sinclair had commissioned GST Computer Systems to produce an operating system for the machine, but switched to QDOS, developed in-house, before launch. GST's OS, designed by Tim Ward,
"""
"Scrotes", we used to call it (UK)
Most Macs I encountered in those days (e.g. in art depts of universities) came furnished with a Zip drive. And someone to tell you how superior they were, of course.
And remember that Fogent labs had nothing to do with the patent.
They bought it.
Compression Labs Inc. filed the patent. So the original blame for
the situation being as it is lies with the USPTO and them, not Fogent.
Of course, Fogent are culpable for their slimy use of thumbscrews
after buying the patent.
FatPhil
Degrees of freedom are a mathematical concept. If you're that unaware of the tools without upon which science and engineering are based, then you can't be much of either.
However, like all anonymous cowards you can't be either, as you're a nobody.
My emphasis should have been more on "public", sorry. Deals with governments to supply things to schools, isn't public in the same way that you and I buying stuff from cdbaby is.
I'm not sure how they (anyone) expects to stop these things being stolen more than copper and aluminium cable are. And if they do have a value in the developed world, then the black market trade in them could be quite horrendous.
FatPhil
Oh come on - I extracted the sentence with the blooper in - how much more trimming did you want me to do? /Degrees of freedom/ map approximately onto dimensions, or coordinates.
To specify a hinge position, you only need to specify one coordinate - an angle; hence it has one degree of freedom.
Monitors with a fixed base on the desk would have 2 degrees of freedom - a left/right angle, and an up/down angle. If it can move on the desktop it has 2 more degrees of freedom ((x,y) coordinates). If the desktop can be raised or lowered, it has an additional 5th degree of freedom.
The most a rigid body can have is 6 degrees of freedom - 3 position, and 3 direction. The one that the above monitor lacks is the ability to rotate in-plane - flipping from landscape to portrait mode, for example.
An anglepoise lamp with a fixed base will typically have 2 (base) + 1 (elbow) + 2 (wrist) degrees of freedom. More join joints means more degrees of freedom.
Hence the absurdity of 340 degrees of freedom.
FatPhil
Well, the fact that they don't intend to sell them to the public at all should counter most of your worries. This pledge drive is intending to get them to change that. However, given that pledges are currently coming in from people like "I. P. Freely", I guess it'll probably get ignored.
It's all in TFA, if you'd cared to R it.
FatPhil
On the 'worldchanging' link on the pledge page, quoth Ethan Zuckerman:
"""
I wonder if the hinges are going to be a problem - the current design requires a hinge for the gasket and a separate hinge that allows 340 degrees of freedom between the screen and the keyboard.
"""
Yeah, right. How about you learn what engineering terms mean before you use them in _completely the wrong way_.
Sheesh, and such blog journalism is the future, eh? ${DEITY} help us.
FatPhil
RTFA:
"""
The suggestion has been made that he also offer it for sale for ~$300 to the rest of us so that we do have an interesting macnine and can help to support the cost computers for the developing world.
"""
What bit of "also" do you not understand?
FatPhil
The evaluation module comes with an 8-bit microcontroller and a UART.
You can therefore trivially connect it to any machine with a UART, be that a PC, linux box, or maybe even a Mac (do they still have UARTs?). What you do with the readings is then entirely up to you. However, the raw data it sends you is self-explanatory {x,y,z} data, and it shouldn't be too hard to tell a slash from a jab.
"They are accelerometers, meaning they can be used to detect relative motion."
And here we have proof why the mislabelling as motion sensors is so bad.
You probably knew what you wanted to say, that I don't doubt, but the above is just plain wrong. Changes in motion, yes (dv/dt); but not differences in motion (v1-v2).
This is the 2000s, I'd hope it was just a single multi-axis accelerometer.
Freescale has some funky electron micrographs in their accelerometer literature,
and they're so detailed you can see precisely how they work. Works of art!
http://freescale.com/
Yes, I work for them.
But wouldn't the other 2 be a waste of HD space, if that were to be your attitude?
OT - why do people think that these things are "motion" sensors? Such a thing would defy the laws of physics.
FatPhil