Wow, you read our paper quite quickly. Impressive. You may have noticed the part where we described the paper as a "case study". I don't claim that we proved anything too generalizable with this work, although there are many such case studies in the literature that reach similar conclusions.
I also regret that space constraints precluded much of the reporting that you would have liked to have seen. Much of it was presented at the talk, but that is indeed insufficient. My apologies.
We resorted to the formal methods approach only after failing with most of the approaches you suggested. UML had nothing obvious to offer (according to my UML-skilled friends). The problem was a design problem: meditating over source code wasn't obviously helpful except to discover that a given design was flawed by observing it in the implementation. We were constrained to C in a POSIX environment, which was really at the heart of the problem, so changing languages or platforms wasn't possible. I was certainly paired with someone throughout the whole development process: sometimes an implementor, sometimes a formal-methods guy.
The effort the formal spec took us was about 4 weeks, most of it by me. The key insight actually only took a few days, the rest was just careful checking for other problems. Once we discovered our requirements were inconsistent, it took us another week or so to come up with relaxed requirements and a design that met them.
Ultimately, keep in mind that we applied formal methods to the problem only after failing multiple times to get a correct implementation using more conventional open source methods. You can quibble about whether it's science or not ("Computer Science : Science:: Plumbing : Hydrodynamics" --Strachey), but subjectively I solved a problem using Z that I and two other smart people working together hadn't solved without it even given a lot of effort. I'll mark this one in the success column.
As for your more general comments about Computer Science, let me observe that CS is pretty much the only science or engineering discipline without any generally-accepted formal descriptive and analytic notation. I'm using Z for this on occasion, and finding that I like it a lot: it helps me to clearly, succinctly, and unambiguously specify all sorts of things. I think that this is somewhat orthogonal from the "behavioral science" approach you seem to be advocating, but I think it is a legitimate need of the discipline. Your mileage, as always, may vary.
I'm sorry, I spoke imprecisely. I watched someone work around digitally-signed "security" code in NT 4 in the manner described. The target, however, was indeed not an audio driver. NT 4 did in fact have signing of some software subsystems.
The principle, of course, is the same whatever kind of signed code it happens to be. Change the code doing the check, and it becomes easy to supply a matching signature:-).
Thanks much, anonymous friend, for your gentle request for clarification. I hope that I have been of service.
Good point. Of course Microsoft and/or the major record labels have tried to shut down such sites in the past (e.g. DeCSS) with incredibly little success.
Under Windows, an encrypted audio stream may require that only signed audio output drivers may play the stream, and Microsoft will sign an audio driver only if it turns off cleartext digital outputs (such as the.wav redirection) at the request of the stream.
Until they get Palladium in (and even then there will inevitably be mod chips) you can modify Windows to quit checking the driver signature pretty easily. I've seen it done with NT 4. Then off you go again.
The sad thing about these Windows-only copy protection schemes is that they don't even work for Windows.
The image on my home-built teleprompter was -- of course -- backward. I tried to find a way to reverse the entire screen, but that was fruitless.
Note that with a modern version of the X server supporting Keith Packard's "Resize and Rotate" extension and utility, this could be easy. Just say "xrandr -x" to mirror the display left-to-right. (Unfortunately, this doesn't appear to work for all servers supporting the extension yet.)
For as long as there's been UNIX, there's been arguments that there needs to be a "one true UNIX", often phrased exactly the way parent phrased them. We've even tried it a couple of times. There was an era where it was pretty much all BSD, and another era where SVID compliance was pretty much mandatory.
Oddly, those eras have passed us by. Folks have found that a range of UNIX kernels and application suites, sharing some common function but tuned toward specific niches, is a better match for real work in the long run. Right now, there are a proliferation of Linux distros aimed at a few niches. They are slowly sorting themselves out, though. In the long run (over the next few years) UNIX/Linux as a whole will be strengthened by this competition and differentiation.
As a proud user of 2.9BSD and many of its successors, I'm happy to see the good parts of history repeat themselves once in a while.
Atlantis summary: The writing sucks. Every episode is transparent and stupid. The "humor" is unfunny, the "drama" boring, the "adventure" tame. The casting sucks. A pack of identical-looking 20-somethings led by a (gasp) 30-something boss: all with an "ideal" ethnic and gender mix inspired in no way by their time, place, or roles. The acting sucks. The actors have an emotional range somewhere between Pia Zadora and Keanu Reeves. The bad guys suck. The authors have talked explicitly about how they came up with the Wraith: a villain that has and uses a brain in interesting ways would have been much more fun. Having the Wraith star in pretty much every episode is also a big mistake IMHO.
Of course, all of this is based on the roughly four episodes I could stand to watch. I'll stick with my beloved SG-1, even as it jumps the shark.
I concur 100%. "Difficult to ignore?" I'm ignoring the little pile of slurs now!
I thought it was interesting how theoretical his argument about Wikipedia's accuracy is. I have faith in Wikipedia's accuracy because I use it all the time, checking its facts against other sources. On the points I've cared about, it fares very well indeed.
It is interesting that he is so exercised by the birthdate of Alexander Hamilton. The article about Hamilton is in general quite consonant with published biographical information. Amusingly, the birthdate has also already been corrected. Try achieving a 24-hour correction cycle sometime, Britannica!
I also find it fascinating how the/. crowd and mods have jumped on the "Wikipedia sucks" bandwagon for this article. These are presumably many of the same people who, like myself, use and contribute to it regularly. Just goes to show, IMHO, that the temptation to appear "intellectual" in public is very strong.
There will always be a place in the world for professionally-edited knowledge sources. For me, however, the cost of Britannica means that I will use it rarely, and at the library. For most of my actual reference needs, Wikipedia works just fine, today.
I don't necessarily think that folks are automatically less credible if they use an alias.
I think that using an alias to publish on "one side" of an issue (e.g. "scientific astrophysics is broken") while using a real name to publish on the "other side" of the same issue (e.g. "my research results in scientific astrophysics") is at least a tiny bit intellectually dishonest, and reduces credibility. I also think that for works where knowledge of the author's background might reflect on his credibility (one way or the other) it is somewhat intellectually dishonest not to make it clear that the name attached to the article is an alias. (I wasted a lot of time looking for the academic background and publication record of "Robert Newton", and was surprised to find that his actual record actually increased his scientific credibility.)
In brief, I just think that using an alias makes it harder to estimate credibility. If there's any pejoration in an alias, it is simply a residue of the fact that many folks who use an alias, especially in scientific publishing, do indeed do so to hide their lack of proper credentials. But then again, Thunderstruck, why would you take po8's word for it?:-)
IMHO, not really. The other major failings of the x86 insn set are its complex and expensive to decode insns, its lack of sufficient visible register names, and its difficult dynamic scheduling constraints. Much work has gone into ameliorating these issues in x86.
The problem with IA-64 is that the architects seemed to have this view that if one extreme was bad, the other must be good. Having 128 GP registers turns out to be as bad as having a small handful. Having 128-bit-wide instructions turns out to have its own set of problems compared to x86's hyper-compressed ones. Uber-static scheduling turns out to be as hard as not being able to do so well at all.
IOW, I (and others I know) think there was probably room in the market at the time for an x86-killer. But IA-64 wasn't it.
Identifying THREE parallel instructions at compile time, ALL THE TIME, is damn hard, and normally the compilers fail. Hence slow.
Actually, one of my MS students and I did some work, later extended in a MS thesis by Svante Arvedahl, that showed that it is pretty straightforward to produce decently-scheduled code for the IA-64 on a JIT basis using combinatorial search techniques and related heuristics. The cool part about this is that you can then use HotSpot(TM)-type techniques to get your instruction-level parallelism way up.
If the IA-64 hadn't tanked so badly in the marketplace, I'd still be working on this stuff...
Hey, didn't you read the attribution of the article? The author, Robert Newton, is a doctoral student in astrophysics, at an accredited university in the United States. An accredited university! I dunno who you are, but I'll bet you're not no doctoral student in astrophysics at an accredited university!
Of course, "Robert Newton" is not his real name: it appears to be Jason Lisle. He is actually now Dr. Jason Lisle, although apparently he wasn't in August. He should be congratulated on his freshly-minted doctoral degree from a reasonable-looking astrophysics program at the University of Colorado at Boulder. He should also be congratulated on his publication record. It includes a number of publications in creationist journals, and at least a few academic research publications
in various journals.
Funny guy: can't figure him out. Seems to be doing some good research on solar and planetary stuff, but to be somewhat out of his league in cosmology-land. His scientific credibility is certainly not enhanced by his use of an alias for much of his work, nor by his flippant dismissal of the cosmic microwave background as evidence for time scales of universal expansion.
Hope he finds what he is looking for. Maybe I'll learn something from him someday.
I screwed that post up badly. Sorry: hit Submit instead of Preview by mistake. Wish I could delete my own posts. Oh well, let's try again...
If you're serious about this, drop me a line at lawyer@po8.org. I can think of two projects I'm heavily involved with offhand that could use some pro bono legal help from somebody with the right credentials.
Yeah, I was reading the article (!) and imagining the horror of bibliophiles everywhere at taking food or drink (much less coffee) in the presence of a rare first edition. Hopefully it's just a fictional embellishment...
"I thought he was a rather sympathetic character. The wanton killing of earlier heroes in developing his machines was a bit much..."
Whaaa?? "A bit much"?? OK, so the man cruelly murdered dozens of innocent superheros in barbarous experiments. I wouldn't call that "a bit much". Only to be expected. If he had brutally tortured every human on earth and sent the planet to a fiery death, that would have been "a bit much". As it was, he was being perfectly reasonable. He had all my sympathy: heck, I wish he'd won.
Good point. In fact, I think we should extend copyright to include conversations. I can't count the number of times I've told someone something, and they've gone off and used it to make money and not compensated me. I think that we need a "conversation copyright" such that anything I tell you in conversation you can't use without my permission.
Sure, such a law is completely infeasible, and would complete inundate the court in frivolous lawsuits. But we are already part way there, aren't we?
Its ironic that you chose to talk about sharing recipes. Recipes aren't, in point of fact, copyrightable: go look it up. If you want to not share your recipe, you have a simple option: don't tell it to anyone. Why is this hard to swallow?
Copyright is a socially constructed concept. Basically, copyrightholders are entitled to a monopoly of sorts for a limited time on their work. Most people agree that the primary reason for this is to encourage more creation of works.
The interesting thing is that we've run a few little controlled experiments over the years in whether this is needed: the obvious examples are fonts and recipes. Neither fonts nor recipes, for obscure reasons, are copyrightable. (Recipe books and software representations of fonts are copyrightable, however. The effect of this is hard to estimate, but it appears to be small: see the giant collections of both artifacts freely and legally available online.) The results of these little experiments may surprise some: there's no shortage of either. In fact, skilled folks still work very hard at producing them!
The theory that folks won't produce quality intellectual artifacts without financial compensation is just that: a theory. The open source movement is another powerful example of counter-evidence to that theory. Many economists believe that better theories are needed at this point.
I believe that copyright can still play a valuable role in our society, but not because it inspires the creation of intellectual works in any broad sense. Copyright serves a useful role by ensuring that the authors of work are properly credited. It also attempts to ensure that any profits that do happen to be generated as a result of using the work commercially are properly allocated. However, copyright law is massively overconstructed for the first of these goals, and generally fails miserably at the second (c.f. music).
I thus support major changes in the copyright law, including limits on assignability of profits under copyright, dramatic shortening of the time period of the copying provisions of copyright, and extension of the crediting provisions of copyright to perpetuity.
Seems to me that Jon Stewart decided that he agreed with your analysis: there's no way to get an audience for a political show of the type he wants. So he went and got on a comedy show which talks about political topics.
Stewart never said not to "do as I do". He was just very clear about this: If you're doing entertainment, such as theatre or comedy, you should be labeling it theatre or comedy rather than journalism. Otherwise, you're doing a grave disservice to the surprising number of Americans who can't tell the difference.
NPR has manipulated laws and its public "competition" to the point where it has a near-monopoly on non-profit radio in the United States. For example, they have consistently sided with commercial broadcasters against allowing low-power (and thus low-cost) FM radio stations. Some college radio stations were driven off the air when NPR successfully lobbied the FCC to kill their licenses soon after it was formed.
Keep in mind that NPR is a medium-sized corporation: it pays salaries to quite a few people, owns infrastructure and facilities, etc. It has about the same set of concerns as any (privately-held) broadcasting corporation, including increasing market share and revenue.
In addition, as you observe, NPR is funded directly by the same large corporations that fund
the Democratic and Republican parties. While I'm skeptical that there's explicit tying of donations to content, I'm sure that NPR is careful to keep its overall format fundable.
Hence Jon Stewart's comment: "At least I can sleep at night." It's clear that if Stewart wanted to, he could put together a political wrestling federation show that would blow Crossfire and everything else off the map. But he doesn't, because he's not evil. The fact that the news networks can't make as much money if they're not evil doesn't excuse them from responsibility. "The people want to be lied to" is a lousy excuse, and Stewart called the Crossfire folks on this. They need to do some kind of theatre less harmful to our country.
$1.70 for packaging and distribution
$1.77 to the musician / artist (split among author, performers, and union)
$4.69 to the retailer
$7.83 to the publisher / label
Draw your own conclusions.
Slashdot Mirror
Wow, you read our paper quite quickly. Impressive. You may have noticed the part where we described the paper as a "case study". I don't claim that we proved anything too generalizable with this work, although there are many such case studies in the literature that reach similar conclusions.
I also regret that space constraints precluded much of the reporting that you would have liked to have seen. Much of it was presented at the talk, but that is indeed insufficient. My apologies.
We resorted to the formal methods approach only after failing with most of the approaches you suggested. UML had nothing obvious to offer (according to my UML-skilled friends). The problem was a design problem: meditating over source code wasn't obviously helpful except to discover that a given design was flawed by observing it in the implementation. We were constrained to C in a POSIX environment, which was really at the heart of the problem, so changing languages or platforms wasn't possible. I was certainly paired with someone throughout the whole development process: sometimes an implementor, sometimes a formal-methods guy.
The effort the formal spec took us was about 4 weeks, most of it by me. The key insight actually only took a few days, the rest was just careful checking for other problems. Once we discovered our requirements were inconsistent, it took us another week or so to come up with relaxed requirements and a design that met them.
Ultimately, keep in mind that we applied formal methods to the problem only after failing multiple times to get a correct implementation using more conventional open source methods. You can quibble about whether it's science or not ("Computer Science : Science :: Plumbing : Hydrodynamics" --Strachey), but subjectively I solved a problem using Z that I and two other smart people working together hadn't solved without it even given a lot of effort. I'll mark this one in the success column.
As for your more general comments about Computer Science, let me observe that CS is pretty much the only science or engineering discipline without any generally-accepted formal descriptive and analytic notation. I'm using Z for this on occasion, and finding that I like it a lot: it helps me to clearly, succinctly, and unambiguously specify all sorts of things. I think that this is somewhat orthogonal from the "behavioral science" approach you seem to be advocating, but I think it is a legitimate need of the discipline. Your mileage, as always, may vary.
See our Freenix 2002 paper for one example of applying formal methods to open source development. Worked great for us!
I'm sorry, I spoke imprecisely. I watched someone work around digitally-signed "security" code in NT 4 in the manner described. The target, however, was indeed not an audio driver. NT 4 did in fact have signing of some software subsystems.
The principle, of course, is the same whatever kind of signed code it happens to be. Change the code doing the check, and it becomes easy to supply a matching signature :-).
Thanks much, anonymous friend, for your gentle request for clarification. I hope that I have been of service.
Good point. Of course Microsoft and/or the major record labels have tried to shut down such sites in the past (e.g. DeCSS) with incredibly little success.
Under Windows, an encrypted audio stream may require that only signed audio output drivers may play the stream, and Microsoft will sign an audio driver only if it turns off cleartext digital outputs (such as the .wav redirection) at the request of the stream.
Until they get Palladium in (and even then there will inevitably be mod chips) you can modify Windows to quit checking the driver signature pretty easily. I've seen it done with NT 4. Then off you go again.
The sad thing about these Windows-only copy protection schemes is that they don't even work for Windows.
The image on my home-built teleprompter was -- of course -- backward. I tried to find a way to reverse the entire screen, but that was fruitless.
Note that with a modern version of the X server supporting Keith Packard's "Resize and Rotate" extension and utility, this could be easy. Just say "xrandr -x" to mirror the display left-to-right. (Unfortunately, this doesn't appear to work for all servers supporting the extension yet.)
For as long as there's been UNIX, there's been arguments that there needs to be a "one true UNIX", often phrased exactly the way parent phrased them. We've even tried it a couple of times. There was an era where it was pretty much all BSD, and another era where SVID compliance was pretty much mandatory.
Oddly, those eras have passed us by. Folks have found that a range of UNIX kernels and application suites, sharing some common function but tuned toward specific niches, is a better match for real work in the long run. Right now, there are a proliferation of Linux distros aimed at a few niches. They are slowly sorting themselves out, though. In the long run (over the next few years) UNIX/Linux as a whole will be strengthened by this competition and differentiation.
As a proud user of 2.9BSD and many of its successors, I'm happy to see the good parts of history repeat themselves once in a while.
Atlantis summary: The writing sucks. Every episode is transparent and stupid. The "humor" is unfunny, the "drama" boring, the "adventure" tame. The casting sucks. A pack of identical-looking 20-somethings led by a (gasp) 30-something boss: all with an "ideal" ethnic and gender mix inspired in no way by their time, place, or roles. The acting sucks. The actors have an emotional range somewhere between Pia Zadora and Keanu Reeves. The bad guys suck. The authors have talked explicitly about how they came up with the Wraith: a villain that has and uses a brain in interesting ways would have been much more fun. Having the Wraith star in pretty much every episode is also a big mistake IMHO.
Of course, all of this is based on the roughly four episodes I could stand to watch. I'll stick with my beloved SG-1, even as it jumps the shark.
I concur 100%. "Difficult to ignore?" I'm ignoring the little pile of slurs now!
I thought it was interesting how theoretical his argument about Wikipedia's accuracy is. I have faith in Wikipedia's accuracy because I use it all the time, checking its facts against other sources. On the points I've cared about, it fares very well indeed.
It is interesting that he is so exercised by the birthdate of Alexander Hamilton. The article about Hamilton is in general quite consonant with published biographical information. Amusingly, the birthdate has also already been corrected. Try achieving a 24-hour correction cycle sometime, Britannica!
I also find it fascinating how the /. crowd and mods have jumped on the "Wikipedia sucks" bandwagon for this article. These are presumably many of the same people who, like myself, use and contribute to it regularly. Just goes to show, IMHO, that the temptation to appear "intellectual" in public is very strong.
There will always be a place in the world for professionally-edited knowledge sources. For me, however, the cost of Britannica means that I will use it rarely, and at the library. For most of my actual reference needs, Wikipedia works just fine, today.
I don't necessarily think that folks are automatically less credible if they use an alias.
I think that using an alias to publish on "one side" of an issue (e.g. "scientific astrophysics is broken") while using a real name to publish on the "other side" of the same issue (e.g. "my research results in scientific astrophysics") is at least a tiny bit intellectually dishonest, and reduces credibility. I also think that for works where knowledge of the author's background might reflect on his credibility (one way or the other) it is somewhat intellectually dishonest not to make it clear that the name attached to the article is an alias. (I wasted a lot of time looking for the academic background and publication record of "Robert Newton", and was surprised to find that his actual record actually increased his scientific credibility.)
In brief, I just think that using an alias makes it harder to estimate credibility. If there's any pejoration in an alias, it is simply a residue of the fact that many folks who use an alias, especially in scientific publishing, do indeed do so to hide their lack of proper credentials. But then again, Thunderstruck, why would you take po8's word for it? :-)
Hope to hear from you soon.
IMHO, not really. The other major failings of the x86 insn set are its complex and expensive to decode insns, its lack of sufficient visible register names, and its difficult dynamic scheduling constraints. Much work has gone into ameliorating these issues in x86.
The problem with IA-64 is that the architects seemed to have this view that if one extreme was bad, the other must be good. Having 128 GP registers turns out to be as bad as having a small handful. Having 128-bit-wide instructions turns out to have its own set of problems compared to x86's hyper-compressed ones. Uber-static scheduling turns out to be as hard as not being able to do so well at all.
IOW, I (and others I know) think there was probably room in the market at the time for an x86-killer. But IA-64 wasn't it.
Identifying THREE parallel instructions at compile time, ALL THE TIME, is damn hard, and normally the compilers fail. Hence slow.
Actually, one of my MS students and I did some work, later extended in a MS thesis by Svante Arvedahl, that showed that it is pretty straightforward to produce decently-scheduled code for the IA-64 on a JIT basis using combinatorial search techniques and related heuristics. The cool part about this is that you can then use HotSpot(TM)-type techniques to get your instruction-level parallelism way up.
If the IA-64 hadn't tanked so badly in the marketplace, I'd still be working on this stuff...
Hey, didn't you read the attribution of the article? The author, Robert Newton, is a doctoral student in astrophysics, at an accredited university in the United States. An accredited university! I dunno who you are, but I'll bet you're not no doctoral student in astrophysics at an accredited university!
Of course, "Robert Newton" is not his real name: it appears to be Jason Lisle. He is actually now Dr. Jason Lisle, although apparently he wasn't in August. He should be congratulated on his freshly-minted doctoral degree from a reasonable-looking astrophysics program at the University of Colorado at Boulder. He should also be congratulated on his publication record. It includes a number of publications in creationist journals, and at least a few academic research publications in various journals.
Funny guy: can't figure him out. Seems to be doing some good research on solar and planetary stuff, but to be somewhat out of his league in cosmology-land. His scientific credibility is certainly not enhanced by his use of an alias for much of his work, nor by his flippant dismissal of the cosmic microwave background as evidence for time scales of universal expansion.
Hope he finds what he is looking for. Maybe I'll learn something from him someday.
I screwed that post up badly. Sorry: hit Submit instead of Preview by mistake. Wish I could delete my own posts. Oh well, let's try again...
If you're serious about this, drop me a line at lawyer@po8.org. I can think of two projects I'm heavily involved with offhand that could use some pro bono legal help from somebody with the right credentials.
If you're serious about this, drop me a line at . I can think of two projects I'm heavily involved with offhand that could use some pro bono legal help from somebody with the right credentials.
Yeah, I was reading the article (!) and imagining the horror of bibliophiles everywhere at taking food or drink (much less coffee) in the presence of a rare first edition. Hopefully it's just a fictional embellishment...
"I thought he was a rather sympathetic character. The wanton killing of earlier heroes in developing his machines was a bit much..."
Whaaa?? "A bit much"?? OK, so the man cruelly murdered dozens of innocent superheros in barbarous experiments. I wouldn't call that "a bit much". Only to be expected. If he had brutally tortured every human on earth and sent the planet to a fiery death, that would have been "a bit much". As it was, he was being perfectly reasonable. He had all my sympathy: heck, I wish he'd won.
Or something.
Portland State Aerospace Society
Good point. In fact, I think we should extend copyright to include conversations. I can't count the number of times I've told someone something, and they've gone off and used it to make money and not compensated me. I think that we need a "conversation copyright" such that anything I tell you in conversation you can't use without my permission.
Sure, such a law is completely infeasible, and would complete inundate the court in frivolous lawsuits. But we are already part way there, aren't we?
Its ironic that you chose to talk about sharing recipes. Recipes aren't, in point of fact, copyrightable: go look it up. If you want to not share your recipe, you have a simple option: don't tell it to anyone. Why is this hard to swallow?
Copyright is a socially constructed concept. Basically, copyrightholders are entitled to a monopoly of sorts for a limited time on their work. Most people agree that the primary reason for this is to encourage more creation of works.
The interesting thing is that we've run a few little controlled experiments over the years in whether this is needed: the obvious examples are fonts and recipes. Neither fonts nor recipes, for obscure reasons, are copyrightable. (Recipe books and software representations of fonts are copyrightable, however. The effect of this is hard to estimate, but it appears to be small: see the giant collections of both artifacts freely and legally available online.) The results of these little experiments may surprise some: there's no shortage of either. In fact, skilled folks still work very hard at producing them!
The theory that folks won't produce quality intellectual artifacts without financial compensation is just that: a theory. The open source movement is another powerful example of counter-evidence to that theory. Many economists believe that better theories are needed at this point.
I believe that copyright can still play a valuable role in our society, but not because it inspires the creation of intellectual works in any broad sense. Copyright serves a useful role by ensuring that the authors of work are properly credited. It also attempts to ensure that any profits that do happen to be generated as a result of using the work commercially are properly allocated. However, copyright law is massively overconstructed for the first of these goals, and generally fails miserably at the second (c.f. music).
I thus support major changes in the copyright law, including limits on assignability of profits under copyright, dramatic shortening of the time period of the copying provisions of copyright, and extension of the crediting provisions of copyright to perpetuity.
Seems to me that Jon Stewart decided that he agreed with your analysis: there's no way to get an audience for a political show of the type he wants. So he went and got on a comedy show which talks about political topics.
Stewart never said not to "do as I do". He was just very clear about this: If you're doing entertainment, such as theatre or comedy, you should be labeling it theatre or comedy rather than journalism. Otherwise, you're doing a grave disservice to the surprising number of Americans who can't tell the difference.
NPR has manipulated laws and its public "competition" to the point where it has a near-monopoly on non-profit radio in the United States. For example, they have consistently sided with commercial broadcasters against allowing low-power (and thus low-cost) FM radio stations. Some college radio stations were driven off the air when NPR successfully lobbied the FCC to kill their licenses soon after it was formed.
Keep in mind that NPR is a medium-sized corporation: it pays salaries to quite a few people, owns infrastructure and facilities, etc. It has about the same set of concerns as any (privately-held) broadcasting corporation, including increasing market share and revenue.
In addition, as you observe, NPR is funded directly by the same large corporations that fund the Democratic and Republican parties. While I'm skeptical that there's explicit tying of donations to content, I'm sure that NPR is careful to keep its overall format fundable.
Hence Jon Stewart's comment: "At least I can sleep at night." It's clear that if Stewart wanted to, he could put together a political wrestling federation show that would blow Crossfire and everything else off the map. But he doesn't, because he's not evil. The fact that the news networks can't make as much money if they're not evil doesn't excuse them from responsibility. "The people want to be lied to" is a lousy excuse, and Stewart called the Crossfire folks on this. They need to do some kind of theatre less harmful to our country.
If you have TV and a cable, you need a TiVo---trust me on this one. I never miss the Daily Show.
A simpler view of the same data...
$1.70 for packaging and distribution $1.77 to the musician / artist (split among author, performers, and union) $4.69 to the retailer $7.83 to the publisher / label Draw your own conclusions.