The only problem might be the cost (it's never cheap)
Grad school isn't really that expensive - most schools will give stipends, and it's not too hard to get fellowships. C.S. is actually one of the best fields to be in for this - fewer C.S. majors proportionally go to grad school, so there's less competition. In addition, the fairly generous National Science Foundation fellowships are limited to US citizens, and C.S. has a higher proportion of non-US-citizen grad students than other fields, so there's even less competition for those, if you happen to be a US citizen.
Yes, in the business world you're correct, but the "unless they went into academia" is a significant "unless." A far lower percentage of C.S. people than most other fields go on to get PhD's in the first place, because a lot of C.S. majors are more into the coding than the research/theoretical. This was especially worsened by the good tech economy of the past few years (why go to grad school when you can make $50k starting salary), so there's a fairly severe shortage of qualified C.S. faculty (most schools require professors to have PhD's in their fields, and few C.S. PhD's means few C.S. profs). Last I heard UCLA has been looking to hire six C.S. professors for two years now. Compare that to say, History, where you're lucky if a single good position opens up every few years, and have to end up teaching at community college while waiting.
So if you think going into academia is an option you'd like, C.S. is definitely one of the best fields to be in, competition-wise.
Even if we limit ourselves to fiction, sci-fi will be a *very* small portion of the fiction that survives to be well-read 50 years from now. Hell, I haven't even heard of half the authors Cliff mentioned NOW, and either have most other people. Authors who will stand the test of time are more timeless authors - ones like J.D. Salinger (Catcher in the Rye), George Orwell (1984, Animal Farm), John Steinbeck (Grapes of Wrath), etc. Does anyone really think in 50 years anyone will have heard of Orson Scott Card, much less place him beside authors like Steinbeck?
The 20th century authors still read 50 years from now will mostly be non-sci-fi authors. Sure, a few sci-fi authors will still be read - J.R.R. Tolkien, Douglas Adams, Isaac Asimov, etc. But the others you mentioned - hell I haven't even heard of half of them NOW. In 50 years the 20th century literary giants that will still be read will be authors like: George Orwell, John Steinbeck, Maya Angelou, J.D. Salinger, etc.
I don't think anyone would seriously argue that the sci-fi authors you mentioned are more well-read even now than someone like John Steinbeck is, let alone in 50 years.
Maybe it's just that I have a 17" rather than 21" or whatever you crazy people have, but overlapping windows always seem to me to be a pain in the ass. You can never see enough information in each window if it's not maximized. So I keep everything maximized, and switch between windows with tabs or hotkeys (either in X or on Windows). Occasionally I have a need to have two windows on the screen at once, but this is very rare. Usually it's faster to keep them both maximized and switch between them (for example, just hit alt-tab in windows repeatedly to switch back and forth).
This is how it is at UCR atleast... I hope its different somewhere else:) I've often wondered -- when there would be an emergency engineering situation where neither calculators nor books are avaliable (a situation that coresponds to testing).
Certainly I wouldn't put it past Microsoft to do something like that, but they do retain backwards compatibility with old NT SMB servers, so at the very least SAMBA is not yet a perfect clone of NT SMB servers. Not that this is necessarily the SAMBA developers' fault, as their code emulated all necessary features of SMB up until now.
The point remains that UIs are an important part of software design (as something great that has a shitty interface is still shitty for the user), so someone who calls themselves a "professional software developer" should be familiar with at least the basics of various UI types, of which the GUI is one. Certainly the Windows GUI is not the only or the only significant GUI, but it is one used by millions of people, which provides a large dataset from which to determine its successes and failures...that sort of insight is valuable if one wants to design GUIs, regardless of whether one wants to actually program Windows GUIs.
Have you tried networking with a SAMBA box? It blows.
If WinXP's network code is fast with all the supported uses (for example, using an NT SMB server rather than SAMBA), and is only slow when using SAMBA, don't you think perhaps the conclusion that SAMBA's SMB interoperability is broken is more reasonable than the conclusion that XP's is?
If code is speech, and I think that will be found to be legitmate sooner or later, can one not legitimately publish a Free DVD Player Suite as speech against the DVDCAA's restrictive licensing scheme?
Yes, but *only* source code has been ruled speech (and even then not universally, only by two courts AFAIK). Compiled code is not speech, and thus has no First Amendment protection. So the *most* that the court would uphold (if even that) is that you could publish the source code to such a suite legitimately, but you could not distribute compiled binaries of it (say in embedded devices) or implement it in hardware, as those wouldn't be forms of speech.
I'm not sure what they're worried about here; popular TV shows are already readily available on the internet for download, usually within hours of their airing. You don't need a ReplayTV to record them; you can do it with a tv-tuner card on your computer and video recording software. People do it all the time - this would make it easier, but you only need one person to record each show for everyone to be able to get it, so that's really not important. And there's far fewer TV shows aired than there are music CDs released, so you don't need nearly as many people to rip the stuff - the people with TV-tuner cards or video-in on their PC are enough.
So basically the stuff will be on the internet regardless of the outcome of this suit.
Those are done quite often as well, and while defects are much easier to spot in classical music, it's actually also much easier to encode, due to the lower frequency range (a single flute, for example, is much easier to encode accurately than a heavily distorted flute with white noise played over it). Sure, you probably can't hear a little distortion in the noisy music, but the encoder has a much harder difficulty deciding what's noise and what's not.
Which is why I mentioned rock music, because it's where LAME tends to be worst - with classical music even on the most insane VBR settings to ensure you get transparent encoding, you end up with around 160-180 kbps average.
While the audio test they're doing is good for "decent" audio, which is what Vorbis excels at, it's not a good test for what most of us are interested in - high (so-called "archival") quality audio compression. To do this testing LAME's -abr 134 vs. Vorbis's 128kbps is simply laughable; the current standards for LAME encoding are usually using either the "--r3mix" command line (as defined by www.r3mix.net) or the "--dm-preset standard" command-line. These typically average around 200-250kbps for rock music, and so far are indistinguishable from CD quality in the tests that have been done.
What would be interesting is if Vorbis can achieve these same results at lower bitrates; then I would agree it's better. If it can beat mp3 at 128kbps, then that's nice, but it's pretty irrelevant to me if it still sounds like crap (just not *as* crappy).
Thanks for the correction. I think it's still the case that the majority of NP-complete problems used in public-key encryption are reducible to each other in polynomial time - at least all the ones I've seen used are. And yes, a successful attack on RSA isn't the same thing as a successful attack on an NP-complete problem in the general case. I was speaking of the potential ramifications of a polynomial-time solution to the factoring problem in the general case; such a solution would most likely also mean all other public-key encryption systems would be breakable in polynomial time.
The problem is that if a method to quickly factor large primes were found, nearly all the alternative public key encryption techniques (i.e. connected graph problems and so on) would also be broken easily, as it's been shown that the NP-complete problems all reduce to each other. So some scheme other than relying on problems of NP-complete difficulty would have to be found (which are easy to verify solutions to, but hard to solve, as far as we know so far, which is what public-key encryption relies on).
The number he's quoting is still too high. I run an Athlon 1.33 Ghz with the stock fan and heatsink, and it idles at 49-51 C, maxing out at around 58 C.
I don't see how RMS has ever advocated forced labor. There is no forcing anywhere - you have a choice to labor or not. If you choose in your labors to use my code which I chose to license under the GPL, then you must follow my conditions. If these conditions are onerous or otherwise unacceptable to you, you are free to not use my code at all in your labors. There is no forcing, merely a quid pro quo offer of code in return for modifications.
Clearly that wasn't intended to be a pro-GPL and anti-MIT-license dig, but a comment wondering out loud whether this would be free in the free software way (regardless of which particular free software license) or "free" in the "source is available free of charge, but it's not free software" way that is becoming so common lately. Presumably the author was not aware of MIT's habit of using the MIT license for their software.
Which is what I mean when I say that the supporters of non-GPL free software licenses are a bit too defensive for their own good.
And that's, (imnsho) real freedom, not gpl freedom, as in "free as in speech so long as your opinion agrees with mine."
Why is it continually necessary to provide a GPL dig in every context, even where it's completely off-topic and irrelevant?
One would think the supporters of non-GPL free software licenses are a bit too defensive for their own good. RMS certainly never takes this many opportunities to berate the MIT and BSD licenses (in fact he considers them "free software").
Depends on your definition of "emulator." By your logic, UltraHLE ("Ultra High-Level Emulator," an N64 emulator), is not actually an emulator, as it merely translates calls to the N64 devkit API into Glide library calls - nearly identical to what WINE does.
Or to take it even further, there is no such thing as emulator, as even the lowest-level emulator is just an alternate implementation that translates the [processor name] interface into [e.g. x86] library calls.
Re:These guys have got the right idea.
on
Tiny Apps
·
· Score: 2
That's one thing I don't like about Office - it uses its own widget sets instead of the standard OS ones. Most apps (with a few exceptions) get it right and use the OS's widgets though, so it's generally more uniform.
Re:These guys have got the right idea.
on
Tiny Apps
·
· Score: 4, Insightful
Tell me that after installing a GUI desktop on your Linux box. Are you really claiming that X+GNOME or X+KDE and a plethora of widget sets in order to get a decent number of programs running is less-bloated and better-designed than Windows desktop?
Slashdot Mirror
The only problem might be the cost (it's never cheap)
Grad school isn't really that expensive - most schools will give stipends, and it's not too hard to get fellowships. C.S. is actually one of the best fields to be in for this - fewer C.S. majors proportionally go to grad school, so there's less competition. In addition, the fairly generous National Science Foundation fellowships are limited to US citizens, and C.S. has a higher proportion of non-US-citizen grad students than other fields, so there's even less competition for those, if you happen to be a US citizen.
Yes, in the business world you're correct, but the "unless they went into academia" is a significant "unless." A far lower percentage of C.S. people than most other fields go on to get PhD's in the first place, because a lot of C.S. majors are more into the coding than the research/theoretical. This was especially worsened by the good tech economy of the past few years (why go to grad school when you can make $50k starting salary), so there's a fairly severe shortage of qualified C.S. faculty (most schools require professors to have PhD's in their fields, and few C.S. PhD's means few C.S. profs). Last I heard UCLA has been looking to hire six C.S. professors for two years now. Compare that to say, History, where you're lucky if a single good position opens up every few years, and have to end up teaching at community college while waiting.
So if you think going into academia is an option you'd like, C.S. is definitely one of the best fields to be in, competition-wise.
Oh good, I was afraid only some measly states' attorneys general were against it. But to know that we have Samba on our side...
Or to get more recent, how about Wittgenstein?
Even if we limit ourselves to fiction, sci-fi will be a *very* small portion of the fiction that survives to be well-read 50 years from now. Hell, I haven't even heard of half the authors Cliff mentioned NOW, and either have most other people. Authors who will stand the test of time are more timeless authors - ones like J.D. Salinger (Catcher in the Rye), George Orwell (1984, Animal Farm), John Steinbeck (Grapes of Wrath), etc. Does anyone really think in 50 years anyone will have heard of Orson Scott Card, much less place him beside authors like Steinbeck?
The 20th century authors still read 50 years from now will mostly be non-sci-fi authors. Sure, a few sci-fi authors will still be read - J.R.R. Tolkien, Douglas Adams, Isaac Asimov, etc. But the others you mentioned - hell I haven't even heard of half of them NOW. In 50 years the 20th century literary giants that will still be read will be authors like: George Orwell, John Steinbeck, Maya Angelou, J.D. Salinger, etc.
I don't think anyone would seriously argue that the sci-fi authors you mentioned are more well-read even now than someone like John Steinbeck is, let alone in 50 years.
Maybe it's just that I have a 17" rather than 21" or whatever you crazy people have, but overlapping windows always seem to me to be a pain in the ass. You can never see enough information in each window if it's not maximized. So I keep everything maximized, and switch between windows with tabs or hotkeys (either in X or on Windows). Occasionally I have a need to have two windows on the screen at once, but this is very rare. Usually it's faster to keep them both maximized and switch between them (for example, just hit alt-tab in windows repeatedly to switch back and forth).
This is how it is at UCR atleast ... I hope its different somewhere else :) I've often wondered -- when there would be an emergency engineering situation where neither calculators nor books are avaliable (a situation that coresponds to testing).
Which is why at an increasing number of schools nearly all tests are open-note and open-book. At my school they're actually nearly all take-home as well, though they are timed (you're expected to follow the time requirement on your own, which surprisingly nearly everyone does). This, IMHO, corresponds well to a real-world situation - you have notes and computers/calculators available, but you also need to have enough knowledge to be able to solve the problems in a reasonable period of time without spending 10 hours reading books.
Certainly I wouldn't put it past Microsoft to do something like that, but they do retain backwards compatibility with old NT SMB servers, so at the very least SAMBA is not yet a perfect clone of NT SMB servers. Not that this is necessarily the SAMBA developers' fault, as their code emulated all necessary features of SMB up until now.
The point remains that UIs are an important part of software design (as something great that has a shitty interface is still shitty for the user), so someone who calls themselves a "professional software developer" should be familiar with at least the basics of various UI types, of which the GUI is one. Certainly the Windows GUI is not the only or the only significant GUI, but it is one used by millions of people, which provides a large dataset from which to determine its successes and failures...that sort of insight is valuable if one wants to design GUIs, regardless of whether one wants to actually program Windows GUIs.
Have you tried networking with a SAMBA box? It blows.
If WinXP's network code is fast with all the supported uses (for example, using an NT SMB server rather than SAMBA), and is only slow when using SAMBA, don't you think perhaps the conclusion that SAMBA's SMB interoperability is broken is more reasonable than the conclusion that XP's is?
This is yet another reason to use Linux.
UNIX has been doing bloat since before Windows was even released - remember Motif?
If code is speech, and I think that will be found to be legitmate sooner or later, can one not legitimately publish a Free DVD Player Suite as speech against the DVDCAA's restrictive licensing scheme?
Yes, but *only* source code has been ruled speech (and even then not universally, only by two courts AFAIK). Compiled code is not speech, and thus has no First Amendment protection. So the *most* that the court would uphold (if even that) is that you could publish the source code to such a suite legitimately, but you could not distribute compiled binaries of it (say in embedded devices) or implement it in hardware, as those wouldn't be forms of speech.
I'm not sure what they're worried about here; popular TV shows are already readily available on the internet for download, usually within hours of their airing. You don't need a ReplayTV to record them; you can do it with a tv-tuner card on your computer and video recording software. People do it all the time - this would make it easier, but you only need one person to record each show for everyone to be able to get it, so that's really not important. And there's far fewer TV shows aired than there are music CDs released, so you don't need nearly as many people to rip the stuff - the people with TV-tuner cards or video-in on their PC are enough.
So basically the stuff will be on the internet regardless of the outcome of this suit.
Those are done quite often as well, and while defects are much easier to spot in classical music, it's actually also much easier to encode, due to the lower frequency range (a single flute, for example, is much easier to encode accurately than a heavily distorted flute with white noise played over it). Sure, you probably can't hear a little distortion in the noisy music, but the encoder has a much harder difficulty deciding what's noise and what's not.
Which is why I mentioned rock music, because it's where LAME tends to be worst - with classical music even on the most insane VBR settings to ensure you get transparent encoding, you end up with around 160-180 kbps average.
While the audio test they're doing is good for "decent" audio, which is what Vorbis excels at, it's not a good test for what most of us are interested in - high (so-called "archival") quality audio compression. To do this testing LAME's -abr 134 vs. Vorbis's 128kbps is simply laughable; the current standards for LAME encoding are usually using either the "--r3mix" command line (as defined by www.r3mix.net) or the "--dm-preset standard" command-line. These typically average around 200-250kbps for rock music, and so far are indistinguishable from CD quality in the tests that have been done.
What would be interesting is if Vorbis can achieve these same results at lower bitrates; then I would agree it's better. If it can beat mp3 at 128kbps, then that's nice, but it's pretty irrelevant to me if it still sounds like crap (just not *as* crappy).
Thanks for the correction. I think it's still the case that the majority of NP-complete problems used in public-key encryption are reducible to each other in polynomial time - at least all the ones I've seen used are. And yes, a successful attack on RSA isn't the same thing as a successful attack on an NP-complete problem in the general case. I was speaking of the potential ramifications of a polynomial-time solution to the factoring problem in the general case; such a solution would most likely also mean all other public-key encryption systems would be breakable in polynomial time.
The problem is that if a method to quickly factor large primes were found, nearly all the alternative public key encryption techniques (i.e. connected graph problems and so on) would also be broken easily, as it's been shown that the NP-complete problems all reduce to each other. So some scheme other than relying on problems of NP-complete difficulty would have to be found (which are easy to verify solutions to, but hard to solve, as far as we know so far, which is what public-key encryption relies on).
The number he's quoting is still too high. I run an Athlon 1.33 Ghz with the stock fan and heatsink, and it idles at 49-51 C, maxing out at around 58 C.
I don't see how RMS has ever advocated forced labor. There is no forcing anywhere - you have a choice to labor or not. If you choose in your labors to use my code which I chose to license under the GPL, then you must follow my conditions. If these conditions are onerous or otherwise unacceptable to you, you are free to not use my code at all in your labors. There is no forcing, merely a quid pro quo offer of code in return for modifications.
Clearly that wasn't intended to be a pro-GPL and anti-MIT-license dig, but a comment wondering out loud whether this would be free in the free software way (regardless of which particular free software license) or "free" in the "source is available free of charge, but it's not free software" way that is becoming so common lately. Presumably the author was not aware of MIT's habit of using the MIT license for their software.
Which is what I mean when I say that the supporters of non-GPL free software licenses are a bit too defensive for their own good.
And that's, (imnsho) real freedom, not gpl freedom, as in "free as in speech so long as your opinion agrees with mine."
Why is it continually necessary to provide a GPL dig in every context, even where it's completely off-topic and irrelevant?
One would think the supporters of non-GPL free software licenses are a bit too defensive for their own good. RMS certainly never takes this many opportunities to berate the MIT and BSD licenses (in fact he considers them "free software").
Depends on your definition of "emulator." By your logic, UltraHLE ("Ultra High-Level Emulator," an N64 emulator), is not actually an emulator, as it merely translates calls to the N64 devkit API into Glide library calls - nearly identical to what WINE does.
Or to take it even further, there is no such thing as emulator, as even the lowest-level emulator is just an alternate implementation that translates the [processor name] interface into [e.g. x86] library calls.
That's one thing I don't like about Office - it uses its own widget sets instead of the standard OS ones. Most apps (with a few exceptions) get it right and use the OS's widgets though, so it's generally more uniform.
Tell me that after installing a GUI desktop on your Linux box. Are you really claiming that X+GNOME or X+KDE and a plethora of widget sets in order to get a decent number of programs running is less-bloated and better-designed than Windows desktop?