This has nothing to do with greed, this has to do with expecting fair compensation for work performed.
A lot of companies stand to benefit financially from this work being done. Why wouldn't I work for a company that's willing to pay me fairly for my contribution instead of one that tries to lowball me? No matter how good the cause is, they should still pay the workers fair market rates. It's not like google is hurting for cash.
Dude. Undergrads need to eat and sleep, too.
Also: "Google defines a student as an individual enrolled in or accepted into an accredited institution including (but not necessarily limited to) colleges, universities, masters programs, PhD programs and undergraduate programs."
I would probably apply to this program, but the pay really really sucks, wow.
$4500 for the summer?
I'm a math PhD student, writing a dissertation in numerics. I made 5 times that much last summer *after tax*, and I imagine most other CS-type graduate students can get about the same. $4500 is chicken feed.
First, this post is basically an advertisement. Duly disregarded as such.
Second, young enough (or who went to few enough The Who concerts, space shuttle launches, etc) to still have an appreciable hearing range over 20kHz can hear the difference between AAC and CD or between CD and 192ksample 48bit digital, whatever. It's just a matter of knowing what to listen for (a sizzle ride in a jazz trio exposes almost everything, it's a good place to start). I hear the differences easily, but... I don't care. I'm listening to my records because I want to hear the music, not because I want to hear how pristine my stereo/computer/ipod is. Especially with an iPod, the ambient noise of the listening environment far exceeds the compression loss almost all the time.
In short, yeah, digital music sounds crappy. But no one cares. If they did, they wouldn't spend $50k on a home stereo, they'd spend $400k hiring a good string quartet to sit in their living room every evening all year and play on demand. Or, *gasp*, leave the house and go to the symphony/jazz club/hippie open mic night/metal show/random guy playing guitar on the street/whatever.
AFAIK, this works with all powerbooks, but I have the last rev 1.67 ghz g4, 128mb graphics. I sometimes use it in this mode to play games on my 24" LCD, though usually I just have both screens on so that I have more real estate.
You asked if it were possible to disable the LCD and use only an external monitor with a powerbook. I told you how to do so. (And yes, I do own a powerbook, and yes it does work). The powerbook will happily run with the lid shut, using an external monitor as the only display device. It will go to sleep when you initially close the lid, when you wake it from sleep via usb mouse or keyboard or by plugging in the monitor it will come on using the external monitor.
You asked a question, I gave you a straightforward answer, why all the hate?
The east still owns the big chunk of the news media because there has yet to be a newspaper worth reading produced west of Washington DC. I agree that the NYT et al leave a lot to be desired at times, but come on: LA times? Good once a week, if you're lucky. SF chronicle? Worthless rag. I'd almost prefer to read USA Today. I'd buy the NYT or WSJ at $10 before I would buy a west coast paper for $1.
You're arguing that Windows font rendering "looks better", grandparent is arguing that OS X's font rendering are more faithful to the "exact" representation of the abstract font. These are not opposing points of view. Why are you arguing?
FWIW, I have no difficulty reading the fonts on my powerbook. I can't argue with your preference for the Windows rendering - personal preference isn't up for discussion - but if you find the Apple font rendering blurry, perhaps you should schedule an eye exam?
Or maybe it's just that Apple wants to give people a consistent user experience? No worrying about which songs are or aren't DRM-ed? It's been a design goal for a long time.
I'm not going to claim that this is the reasoning, or that something more sinister isn't at work, but assuming the worst of a company doesn't mean you get to pass it off as fact.
Go to any reasonably large business, and they'll have a site license from Microsoft for sure, even though every single system they buy already comes preloaded with something. What about Apple? (I kid)
Yes, but the complex x86 instructions (and many simpler ones as well) take more than one cycle to execute. The relevant measure isn't the number of instructions required to accomplish a task, but the number of cycles required. You can easily concoct examples for which x86 requires fewer instructions but more cycles.
Depends on what you mean by "native". The english spoken in the US (especially in the mid-atlantic, more so in appalachia) is actually *more* similar to, say, elizabethan english than the english spoken in the UK.
Actually, I play WoW on a powerbook. I have a 5 button mouse. I don't use it. Ever. It's honestly faster for me to use the trackpad and modifier keys than to take my hand off the keyboard (I have a bazillion key bindings, sigh, engineering). If I played a melee character (rogue, really) I would probably use the mouse in pvp combat, but I play mage and warlock, and I find the keyboard + trackpad + 1button to be faster than anything else.
In fairness, I have big hands, so I can mouselock on the trackpad while using my other fingers to hit keyboard keybindings. ymmv.
Merom has something like 6 watts higher TDP than Yonah; that's a lot of extra heat for a small package to deal with. I don't think anyone is hamstringing anything.
There's more to ID than external form. Everything inside the case is ID, as well. The physical package is not just a pretty shell wrapped around whatever the engineers come up with - the whole enclosure is designed together from the get-go.
I agree with your points here almost 100%. That line of argument pisses me off, too.
That said, I do attach some weight to the fact that webkit is open, and MSHTML (to the best of my knowlede) is not. Microsoft could (note "could") choose to not expose more optimized code through the API, or play lots of other games to hinder third party apps.
Dear god, no. Yes, it has double precision, but it's only zomgwtffast in non IEEE-754 single precision. Holy rounding errors, Batman! Will there be obscenely fast LAPACK/fft/convolution benchmarks on the cell? Yes. Will those codes be usable for serious science? Not really.
you're right that that's the step i'm conceptually thinking about, but if you expect that that algorithm will compute any semblance of sin or cos, i'm afraid you're rather wrong - it returns 1 for any input. even if it weren't for that bug, it wouldn't give you a numerically good sin or cos. the error in the taylor series polynomials is concentrated at the ends of the interval - you can use a much lower dimensional polynomial if you choose one that approximates with roughly equal error throughout the interval.
secondly, your range reduction step is totally useless - first, it computes the result of dividing by 2pi repeatedly, not of subtracting off some multiples of 2pi. second, roundoff error is going to obliterate any accuracy you may have once had as soon as x gets "big".
check out k_rem_pio2.c in fdlibm for a good free (beer & speech) implementation of range reduction for these functions, and see how hard it really is. this is not, by any stretch of the imagination, easy coding.
medical imaging. the inverse integral transforms that are used in processing are often computed using ffts. do *you* want a PET/MRI/whatever scan to miss the piece of shrapnel in your chest because it was small enough to be obscured by the single-precision noise floor?
there's a bazillion other applications that need double (and higher) precision, but i'm pretty sure that's the most compelling one for the average person.
Slashdot Mirror
This has nothing to do with greed, this has to do with expecting fair compensation for work performed. A lot of companies stand to benefit financially from this work being done. Why wouldn't I work for a company that's willing to pay me fairly for my contribution instead of one that tries to lowball me? No matter how good the cause is, they should still pay the workers fair market rates. It's not like google is hurting for cash.
Dude. Undergrads need to eat and sleep, too. Also: "Google defines a student as an individual enrolled in or accepted into an accredited institution including (but not necessarily limited to) colleges, universities, masters programs, PhD programs and undergraduate programs."
I would probably apply to this program, but the pay really really sucks, wow.
$4500 for the summer?
I'm a math PhD student, writing a dissertation in numerics. I made 5 times that much last summer *after tax*, and I imagine most other CS-type graduate students can get about the same. $4500 is chicken feed.
First, this post is basically an advertisement. Duly disregarded as such.
Second, young enough (or who went to few enough The Who concerts, space shuttle launches, etc) to still have an appreciable hearing range over 20kHz can hear the difference between AAC and CD or between CD and 192ksample 48bit digital, whatever. It's just a matter of knowing what to listen for (a sizzle ride in a jazz trio exposes almost everything, it's a good place to start). I hear the differences easily, but... I don't care. I'm listening to my records because I want to hear the music, not because I want to hear how pristine my stereo/computer/ipod is. Especially with an iPod, the ambient noise of the listening environment far exceeds the compression loss almost all the time.
In short, yeah, digital music sounds crappy. But no one cares. If they did, they wouldn't spend $50k on a home stereo, they'd spend $400k hiring a good string quartet to sit in their living room every evening all year and play on demand. Or, *gasp*, leave the house and go to the symphony/jazz club/hippie open mic night/metal show/random guy playing guitar on the street/whatever.
Are there any ibooks that can run with the lid closed? I'm under the impression that they all go to sleep.
AFAIK, this works with all powerbooks, but I have the last rev 1.67 ghz g4, 128mb graphics. I sometimes use it in this mode to play games on my 24" LCD, though usually I just have both screens on so that I have more real estate.
You asked if it were possible to disable the LCD and use only an external monitor with a powerbook. I told you how to do so. (And yes, I do own a powerbook, and yes it does work). The powerbook will happily run with the lid shut, using an external monitor as the only display device. It will go to sleep when you initially close the lid, when you wake it from sleep via usb mouse or keyboard or by plugging in the monitor it will come on using the external monitor.
You asked a question, I gave you a straightforward answer, why all the hate?
plug in DVI monitor, USB keyboard and mouse, close powerbook lid, enjoy.
The east still owns the big chunk of the news media because there has yet to be a newspaper worth reading produced west of Washington DC. I agree that the NYT et al leave a lot to be desired at times, but come on: LA times? Good once a week, if you're lucky. SF chronicle? Worthless rag. I'd almost prefer to read USA Today. I'd buy the NYT or WSJ at $10 before I would buy a west coast paper for $1.
You're arguing that Windows font rendering "looks better", grandparent is arguing that OS X's font rendering are more faithful to the "exact" representation of the abstract font. These are not opposing points of view. Why are you arguing?
FWIW, I have no difficulty reading the fonts on my powerbook. I can't argue with your preference for the Windows rendering - personal preference isn't up for discussion - but if you find the Apple font rendering blurry, perhaps you should schedule an eye exam?
Or maybe it's just that Apple wants to give people a consistent user experience? No worrying about which songs are or aren't DRM-ed? It's been a design goal for a long time.
I'm not going to claim that this is the reasoning, or that something more sinister isn't at work, but assuming the worst of a company doesn't mean you get to pass it off as fact.
Yes, but the complex x86 instructions (and many simpler ones as well) take more than one cycle to execute. The relevant measure isn't the number of instructions required to accomplish a task, but the number of cycles required. You can easily concoct examples for which x86 requires fewer instructions but more cycles.
Depends on what you mean by "native". The english spoken in the US (especially in the mid-atlantic, more so in appalachia) is actually *more* similar to, say, elizabethan english than the english spoken in the UK.
Actually, I play WoW on a powerbook. I have a 5 button mouse. I don't use it. Ever. It's honestly faster for me to use the trackpad and modifier keys than to take my hand off the keyboard (I have a bazillion key bindings, sigh, engineering). If I played a melee character (rogue, really) I would probably use the mouse in pvp combat, but I play mage and warlock, and I find the keyboard + trackpad + 1button to be faster than anything else. In fairness, I have big hands, so I can mouselock on the trackpad while using my other fingers to hit keyboard keybindings. ymmv.
Merom has something like 6 watts higher TDP than Yonah; that's a lot of extra heat for a small package to deal with. I don't think anyone is hamstringing anything.
If the original failure was due to a manufacturing defect, not poor design, then it rather is fixing the problem.
There's more to ID than external form. Everything inside the case is ID, as well. The physical package is not just a pretty shell wrapped around whatever the engineers come up with - the whole enclosure is designed together from the get-go.
I agree with your points here almost 100%. That line of argument pisses me off, too.
That said, I do attach some weight to the fact that webkit is open, and MSHTML (to the best of my knowlede) is not. Microsoft could (note "could") choose to not expose more optimized code through the API, or play lots of other games to hinder third party apps.
what the hell do you think this has to do with the question that was asked?
No, but webkit is a published framework that third party developers can link against.
you're right that that's the step i'm conceptually thinking about, but if you expect that that algorithm will compute any semblance of sin or cos, i'm afraid you're rather wrong - it returns 1 for any input. even if it weren't for that bug, it wouldn't give you a numerically good sin or cos. the error in the taylor series polynomials is concentrated at the ends of the interval - you can use a much lower dimensional polynomial if you choose one that approximates with roughly equal error throughout the interval. secondly, your range reduction step is totally useless - first, it computes the result of dividing by 2pi repeatedly, not of subtracting off some multiples of 2pi. second, roundoff error is going to obliterate any accuracy you may have once had as soon as x gets "big". check out k_rem_pio2.c in fdlibm for a good free (beer & speech) implementation of range reduction for these functions, and see how hard it really is. this is not, by any stretch of the imagination, easy coding.
you might, but you'd be wrong. the inverse problem in question is ill-posed, and subject to a fairly high amount of numerical instability.
medical imaging. the inverse integral transforms that are used in processing are often computed using ffts. do *you* want a PET/MRI/whatever scan to miss the piece of shrapnel in your chest because it was small enough to be obscured by the single-precision noise floor?
there's a bazillion other applications that need double (and higher) precision, but i'm pretty sure that's the most compelling one for the average person.