While you are correct, a biology degree (and thus biologist) could contribute to climate science in a constructive and meaningful way.
[long list of completely irrelevant stuff]
Not a single thing you list has anything whatsoever to do with climate science. Nothing.
List one single paper in which he contributes significant work to climate science. There aren't any. He's a colorful popularizer, but he's a biologist, not a climate scientist.
To the parent's point, the post referenced Lovelock's actual work, not the particular degree he held. Now if you want to argue that parent's point isn't true because Lovelock did indeed do climate research fine, but talking about climate science doesn't make you a climate scientist, regardless of which degree is held. Doing actual science work, and getting published in peer reviewed journals, makes you a climate scientist... you know, just like every other science field.
I came to the same conclusions as you re: communications. Unfortunately entanglement doesn't last forever and tends to decay with time/interactions (with, say the walls of the fiber optic cables). My instinct says that the entanglement will never last long enough that such that a message could be encoded in a minimum size bulk packet of photons (such that you have enough photons to have a statistical certainty of the packet's state of entanglement, and thus a representation of a bit), sent, decoded, and finally understood by the message capturing device. It takes time to generate photons, switch in and out the entanglement process, etc. All that latency might just add up to be the 'time to live' of the entanglement state.
"Or you are using a computer monitor. Or a laptop."
Sorry, but unless you have the eyesight of a bird of prey then you won't be seing the pixels on a 1080 computer monitor from normal working distance.
Wow. I don't know about you, but I have pretty bad eyes, and I STILL notice rampant and distracting pixelation on my 24 inch monitor that does 1920x1200 resolution. The pixel pitch is just the wrong side of visible even for sitting at a good distance back. Considering my 15 inch laptop had an SXGA+ screen (that's 1400x1050) the pixel pitch was just fine (meaning I couldn't see them). But 1080p is definately NOT enough for standard monitors if you get ones that are decently sized. My standard was was the height of a 4:3 19inch monitor, but widescreen, and at that simple and usable size, the given resolution is not high enough.
A small step up in resolution would be plenty at that monitor size, but the next step that's typically available is enough for HUGE monitors (up to something like 32-36inches). I consider this a problem, because smaller monitors are too small (i.e. viewable height is reduced from a typical 4:3 19inch monitor) for decent work, web/document viewing, etc.
Well, to be fair, that cheap hardware price comes in no small part from regulation and settlements forcing AMD and INTEL to share patents to be able to compete (intel anti-trust stuff)... ya know.. the gub'mint and all. Thanks for proving competition is good, but only works with proper oversight and regulation.
Sure sure, I never meant to imply that you don't need to learn any sequences at all, particularly to be competitive at a high level. But they feel different, less rote, particularly during play (although that may have more to do with the fact that I'm not a very good Go player at the moment). A Joseki covers a portion of the board, and you have to strategically apply the one you think will be most advantageous considering the overall board makeup. So even though there are some sequences that are fixed, the whole board is much more dynamic. I guess it boils down to the larger board space and the strategic vs tactical nature of Go vs Chess. I guess you could think of it has having four different chess boards going on simultaneously, and each game can influence it's neighbors, so now not only do you have to consider the local position, but you also have to plan how that affects the global situation.
In line with what Skydyr said, Joseki are opening sequences, but they aren't fixed, they can be inverted, played in any corner depending on the rest of the board. So even if you are are playing Joseki, you can deviate from them, or change them around with out it becoming an auto loss (unlike say, chess).
Would it have something to do with the character that was assigned that accent? Say if Jar Jar was not a bumbling fool, but a really smart character type that helped solve many of the main plot problems instead of the supposed comic relief and buffoon, would there have been as much outrage/disgust?
I'm not sure what this has to do with the original premise. Where do you draw the line? Does an employer have the right to inspect your wallet? To inspect your car? Your house? Interview friends and family members? All as preconditions to hire you? Invading your facebook account basically amounts to all of the above.
I don't know about the GP, but for me it's the teeny tiny portion of window space that I now have to scroll around instead of just click. I have a nice large monitor but I'm in a tiny box scrolling around looking for something. That adds 2-3 times as many button/menu clicks. Unless you know a hack to get that stupid scrolling window to actually fill the screen instead of being artificially compressed into a little window. fly-out vs folder view list is not even a consideration.
There's a whole raft of problems with your assumptions and proposals, not least of which is you are completely ignoring history. There are many very good reasons things changed from the way they used to be. Those reasons still exist.
The way senators get selected was changed for reasons of corruption. Look up Tammany Hall era politics for some examples. A poll tax was tried before, it's goal was to keep a certain kind of person from voting. It was unconstitutional (or are you throwing that away now?) and had the effect of keeping people from voting and thus maintaining the current power structure and wealth classes instead of allowing government to reflect the needs of the people. Any poll test would have a similar effect, the very corruption you are trying to attack would twist it such that it enables/maintains the corrupt policies and politicians. That's happened repeatedly throughout history.
In addition, anyone currently receiving some form of "entitlement" should not get to vote because what they're going to vote for is not difficult to guess and this situation is too exploitable and too dangerous for our long-term survival.
This has the same issues, that means all I have to do to keep you from voting is give you a tax break, or send you a check for $1. That's an effective way to illustrate the law of unintended consequences, but not good policy. The cure is worse than the disease.
What you really want is to enable more votes to matter and lower the ability of power structures to manipulate the system. Public funding for elections I completely agree with. Anyone who gets enough support gets a fixed sized bucket and fixed time window for campaigning prior to election. Also serious reform of the voting system, proportioned voting and/or ranked voting for would go a tremendous way towards reducing the problems we're seeing now. No need for mass disenfranchisement or returning to policies even worse than what we have today. There's a lot of people that are way too attached to the one person one vote idea for no good reason. Now I know all this is unlikely to happen, but they're good ideas for reform and don't have quite the problems as some of the other proposals mentioned.
I am an EE, and pretty much this. To get something that looks like a square wave, the bandwidth of your overall circuitry has to be MUCH higher than the frequency of the clock, since the square edge has infinite frequency content. Now that's not to say a good approximation of a square wave could NOT exist at 4 GHz or higher, but there's no need and no point since that would complicate and hinder other aspects of the design.
This right here is B.S. During the industrial revolution when there were no laws or government granted powers... what corporations did was way worse than this. Corporations do NOT need government granted powers to be tyrannical and abusive. It is their nature, human nature in fact, to empower them selves over others. The only entity that can push back against that is the collective of society, which more often than not is represented by government and perhaps public interest groups that have a legal framework to assert their will within.
Thank you for the thoughtful reply, but no, I don't assume that. I'm objecting to the "OMG the sky will fall if somebody doesn't get vaccinated" pseudo-argument, and the misuse of the concept of herd immunity in that context.
Anti-vaxxers are trying to skate on herd immunity, yes. Their presence does weaken the herd, if by "weaken the herd" you mean they increase the size of an existing vulnerable population. But they do not cause plagues, they do not cause healthy individuals to stop producing antibodies, and they do not increase the mutation rate of organisms.
Careful, while mutation rate doesn't increase, having lots and lots more organisms at the same mutation rate dramatically increases the actual number of mutations, and thus increases the chances of successful and viable mutations occurring. 10 organisms with a.5% mutation rate is different from 10 trillion organisms with a.5% mutation rate. And in a partially vaccinated population each of those mutations will have more opportunities to spread and thrive and test their mutations against vaccinated individuals.
There are several large non-vaccinating communities in my area (Old-Order Amish, among others) and they do endure regular, preventable epidemics. Those epidemics do not cause chain reactions into the vaccinated population. They just plain don't. These unvaccinated people interact with the rest of us all the time, and it does not destroy the herd or compromise the herd immunity. That's independently verifiable fact.
Again, be careful with your phrasing here. Herd immunity is precisely what prevents one vulnerable population from infecting other vulnerable populations. And there are researched immunity rates required to achieve effective barriers between vulnerable populations and individuals. Wikipedia has some numbers and references.
Considering that the success of any particular immunization is largely unknowable and the immunity rate needed to achieve effective herd immunity barriers is very vague, every individual that doesn't get immunized does but their large community at a very real risk.
Also consider those populations you talk about, particularly the Amish, are often very isolated. They do interact with the outside world, but those interactions area rare, and the chance of encountering someone else who is un-vaccinated while being sick themselves is very small. That's the very concept the idea of Herd Immunity is meant to convey. Once overall immunity drops to a certain point, that chance of spread increases dramatically. There's a number of neat animations that illustrate this, this one isn't the best I've seen but it was easy to find and it does its job.
Remember that even people that have been successfully immunized can and will become temporary members of the vulnerable population under many circumstances. Sick from other diseases that weaken the immune system, pregnancies, medical treatments that weaken the immune system, etc. There are also those people that do not have a choice (too young, allergies, other medical condition). So those people are the ones that are being put at risk when someone chooses not to vaccinate their children. It's especially bad at a doctor's office since that's where people take their kids when they get sick, and if they bump into kids waiting to get vaccinated, or who's vaccination didn't take, they've suddenly endangered more than just their own child with your decision. I'd be more OK with people refusing vaccines if they would voluntarily quarantine themselves when sick and there were criminal negligence charges available if they're decision affected anyone else.
I think everyone wants the vax/antivax argument to be discretely binary, but it's really more of a nuanced continuum tha
You assume that getting a vaccine means you get immunity. This is not true. Since vaccine efficacy is not 100%, herd immunity side effects are required to protect those that vaccine didn't work on (plus any vulnerable populations). Yes. A successful vaccine means immunity, but there's really no way to know if your vaccine was successful unless you get the disease or go through extensive extra tests. Also, immune systems are not binary, partial efficacy is possible as well, in which case herd immunity is an additional buffer to help protect you.
Finally, there's mutations and evolution of the diseases. Even if you're immune to a strain of disease, a population raging with the disease will quickly produce a strain that you are not immune to. Herd immunity helps to keep that spread and evolution of disease strains in check.
Well, they wouldn't be asked to assault US civilians. They would be told to assault crazy terrorists who, by the way, have been threatening their families and loved ones. The one thing we're good at is controlling information access for our troops. Sure there are a few that think the Iraq was a bad thing, but most are quite proud of what we've "accomplished" over there.
Remember Abu Ghraib, the only way that happens is significant informational and psychological conditioning to dehumanize the 'enemy'. That can happen for anyone they fight, anywhere on earth, including inside our borders.
It was a soldier who was ordered to kill a us citizen in a drone attack, and did so without hesitation.
Bah, look at the labor put downs during the industrial revolution. The army is good at one thing only, following orders. The soldiers themselves probably won't have any idea what's really going on until long after it's over (if they ever get over the misinformation they've been fed).
Just look at how our military, in general, treated the people of Iraq or Afghanistan. Lots of news stories about unethical behavior. In order to wage war you need to dehumanize the enemy, and then the soldiers do whatever they want to them. That will easily happen with any group inside the US that the army would be ordered to fight. They wouldn't be citizens they'd be "evil animals threatening your family, friends, and way of life, we must stop them now!"
All this aside, the executive has already ordered the assassination of American citizens so it's hard to imagine any group even being allowed to get to that stage of organization in the first place. (see: red scare and McCarthyism, WW II Japanese American camps, etc. for how these things are handled historically)
Natural physical monopolies like what you describe should be publicly owned or have line sharing requirements. Preferebly both. This goes for telecoms, power infrastructure, roads, airports, etc. (we got roads mostly right, but the same reasoning applies to all.)
Turns out doing a thousand lines of heavy DSP INSIDE the interrupt handler
I'm not a computer engineer, or EE, or signals person, but even I can tell you why that would fail. Sounds like the problem is not that people program general purpose chips, but that someone hired general purpose programmers to write DSP code. DSP is a beast, it's its own subset of programming that requires knowledge of physics (waves and nyquist) that some programmers don't know. Or maybe the problem really was that your programmers don't know what a bloody interrupt is, and shouldn't even be writing embedded programs.
As you say, it requires several skill sets at once. Embedded DSP requires extensive theory and mathematical knowledge so you KNOW the stuff, as well as low level hardware knowledge of the target device (either in an FPGA or DSP or GPP) so the code is efficient and fast enough to meet the needs of the platform.
There's more than a few of our signal processing gurus with PHd's that are absolutely brilliant, but have no idea about low level hardware. They can develop algorithms and model them in matlab to prove the ideas are possible. They've taught themselves C, or taken a class or two, so they can program stuff and hack it together. But when their test code tries to get used 'for real', the design problems show up.
Knowing some C does not an embedded programmer make.
Slashdot Mirror
wtf? this is not the post I replied to....
While you are correct, a biology degree (and thus biologist) could contribute to climate science in a constructive and meaningful way.
[long list of completely irrelevant stuff]
Not a single thing you list has anything whatsoever to do with climate science. Nothing.
List one single paper in which he contributes significant work to climate science. There aren't any. He's a colorful popularizer, but he's a biologist, not a climate scientist.
To the parent's point, the post referenced Lovelock's actual work, not the particular degree he held. Now if you want to argue that parent's point isn't true because Lovelock did indeed do climate research fine, but talking about climate science doesn't make you a climate scientist, regardless of which degree is held. Doing actual science work, and getting published in peer reviewed journals, makes you a climate scientist... you know, just like every other science field.
I came to the same conclusions as you re: communications. Unfortunately entanglement doesn't last forever and tends to decay with time/interactions (with, say the walls of the fiber optic cables). My instinct says that the entanglement will never last long enough that such that a message could be encoded in a minimum size bulk packet of photons (such that you have enough photons to have a statistical certainty of the packet's state of entanglement, and thus a representation of a bit), sent, decoded, and finally understood by the message capturing device. It takes time to generate photons, switch in and out the entanglement process, etc. All that latency might just add up to be the 'time to live' of the entanglement state.
"Or you are using a computer monitor. Or a laptop."
Sorry, but unless you have the eyesight of a bird of prey then you won't be seing the pixels on a 1080 computer monitor from normal working distance.
Wow. I don't know about you, but I have pretty bad eyes, and I STILL notice rampant and distracting pixelation on my 24 inch monitor that does 1920x1200 resolution. The pixel pitch is just the wrong side of visible even for sitting at a good distance back. Considering my 15 inch laptop had an SXGA+ screen (that's 1400x1050) the pixel pitch was just fine (meaning I couldn't see them). But 1080p is definately NOT enough for standard monitors if you get ones that are decently sized. My standard was was the height of a 4:3 19inch monitor, but widescreen, and at that simple and usable size, the given resolution is not high enough.
A small step up in resolution would be plenty at that monitor size, but the next step that's typically available is enough for HUGE monitors (up to something like 32-36inches). I consider this a problem, because smaller monitors are too small (i.e. viewable height is reduced from a typical 4:3 19inch monitor) for decent work, web/document viewing, etc.
Well, to be fair, that cheap hardware price comes in no small part from regulation and settlements forcing AMD and INTEL to share patents to be able to compete (intel anti-trust stuff)... ya know.. the gub'mint and all. Thanks for proving competition is good, but only works with proper oversight and regulation.
Sure sure, I never meant to imply that you don't need to learn any sequences at all, particularly to be competitive at a high level. But they feel different, less rote, particularly during play (although that may have more to do with the fact that I'm not a very good Go player at the moment). A Joseki covers a portion of the board, and you have to strategically apply the one you think will be most advantageous considering the overall board makeup. So even though there are some sequences that are fixed, the whole board is much more dynamic. I guess it boils down to the larger board space and the strategic vs tactical nature of Go vs Chess. I guess you could think of it has having four different chess boards going on simultaneously, and each game can influence it's neighbors, so now not only do you have to consider the local position, but you also have to plan how that affects the global situation.
In line with what Skydyr said, Joseki are opening sequences, but they aren't fixed, they can be inverted, played in any corner depending on the rest of the board. So even if you are are playing Joseki, you can deviate from them, or change them around with out it becoming an auto loss (unlike say, chess).
How was that computer at playing Go?
Would it have something to do with the character that was assigned that accent? Say if Jar Jar was not a bumbling fool, but a really smart character type that helped solve many of the main plot problems instead of the supposed comic relief and buffoon, would there have been as much outrage/disgust?
I'm so angry right now, I'm molting.
Did you seriously just quote Disney's Aladdin in the same breath as talking/raging about Michael Bay's rebooting of a beloved and classic tale?
-1 BURN!
That 70's Show
I'm not sure what this has to do with the original premise. Where do you draw the line? Does an employer have the right to inspect your wallet? To inspect your car? Your house? Interview friends and family members? All as preconditions to hire you? Invading your facebook account basically amounts to all of the above.
I don't know about the GP, but for me it's the teeny tiny portion of window space that I now have to scroll around instead of just click. I have a nice large monitor but I'm in a tiny box scrolling around looking for something. That adds 2-3 times as many button/menu clicks. Unless you know a hack to get that stupid scrolling window to actually fill the screen instead of being artificially compressed into a little window. fly-out vs folder view list is not even a consideration.
There's a whole raft of problems with your assumptions and proposals, not least of which is you are completely ignoring history. There are many very good reasons things changed from the way they used to be. Those reasons still exist.
The way senators get selected was changed for reasons of corruption. Look up Tammany Hall era politics for some examples. A poll tax was tried before, it's goal was to keep a certain kind of person from voting. It was unconstitutional (or are you throwing that away now?) and had the effect of keeping people from voting and thus maintaining the current power structure and wealth classes instead of allowing government to reflect the needs of the people. Any poll test would have a similar effect, the very corruption you are trying to attack would twist it such that it enables/maintains the corrupt policies and politicians. That's happened repeatedly throughout history.
In addition, anyone currently receiving some form of "entitlement" should not get to vote because what they're going to vote for is not difficult to guess and this situation is too exploitable and too dangerous for our long-term survival.
This has the same issues, that means all I have to do to keep you from voting is give you a tax break, or send you a check for $1. That's an effective way to illustrate the law of unintended consequences, but not good policy. The cure is worse than the disease.
What you really want is to enable more votes to matter and lower the ability of power structures to manipulate the system. Public funding for elections I completely agree with. Anyone who gets enough support gets a fixed sized bucket and fixed time window for campaigning prior to election. Also serious reform of the voting system, proportioned voting and/or ranked voting for would go a tremendous way towards reducing the problems we're seeing now. No need for mass disenfranchisement or returning to policies even worse than what we have today. There's a lot of people that are way too attached to the one person one vote idea for no good reason. Now I know all this is unlikely to happen, but they're good ideas for reform and don't have quite the problems as some of the other proposals mentioned.
I am an EE, and pretty much this. To get something that looks like a square wave, the bandwidth of your overall circuitry has to be MUCH higher than the frequency of the clock, since the square edge has infinite frequency content. Now that's not to say a good approximation of a square wave could NOT exist at 4 GHz or higher, but there's no need and no point since that would complicate and hinder other aspects of the design.
So where do you see the big difference between designing a huge CPU completely on transistor level and writing a huge program purely in assembler?
I'm not the AC but one is hardware the other is software. Don't confuse the two.
This right here is B.S. During the industrial revolution when there were no laws or government granted powers... what corporations did was way worse than this. Corporations do NOT need government granted powers to be tyrannical and abusive. It is their nature, human nature in fact, to empower them selves over others. The only entity that can push back against that is the collective of society, which more often than not is represented by government and perhaps public interest groups that have a legal framework to assert their will within.
Thank you for the thoughtful reply, but no, I don't assume that. I'm objecting to the "OMG the sky will fall if somebody doesn't get vaccinated" pseudo-argument, and the misuse of the concept of herd immunity in that context.
Anti-vaxxers are trying to skate on herd immunity, yes. Their presence does weaken the herd, if by "weaken the herd" you mean they increase the size of an existing vulnerable population. But they do not cause plagues, they do not cause healthy individuals to stop producing antibodies, and they do not increase the mutation rate of organisms.
Careful, while mutation rate doesn't increase, having lots and lots more organisms at the same mutation rate dramatically increases the actual number of mutations, and thus increases the chances of successful and viable mutations occurring. 10 organisms with a .5% mutation rate is different from 10 trillion organisms with a .5% mutation rate. And in a partially vaccinated population each of those mutations will have more opportunities to spread and thrive and test their mutations against vaccinated individuals.
There are several large non-vaccinating communities in my area (Old-Order Amish, among others) and they do endure regular, preventable epidemics. Those epidemics do not cause chain reactions into the vaccinated population. They just plain don't. These unvaccinated people interact with the rest of us all the time, and it does not destroy the herd or compromise the herd immunity. That's independently verifiable fact.
Again, be careful with your phrasing here. Herd immunity is precisely what prevents one vulnerable population from infecting other vulnerable populations. And there are researched immunity rates required to achieve effective barriers between vulnerable populations and individuals. Wikipedia has some numbers and references.
Considering that the success of any particular immunization is largely unknowable and the immunity rate needed to achieve effective herd immunity barriers is very vague, every individual that doesn't get immunized does but their large community at a very real risk.
Also consider those populations you talk about, particularly the Amish, are often very isolated. They do interact with the outside world, but those interactions area rare, and the chance of encountering someone else who is un-vaccinated while being sick themselves is very small. That's the very concept the idea of Herd Immunity is meant to convey. Once overall immunity drops to a certain point, that chance of spread increases dramatically. There's a number of neat animations that illustrate this, this one isn't the best I've seen but it was easy to find and it does its job.
Remember that even people that have been successfully immunized can and will become temporary members of the vulnerable population under many circumstances. Sick from other diseases that weaken the immune system, pregnancies, medical treatments that weaken the immune system, etc. There are also those people that do not have a choice (too young, allergies, other medical condition). So those people are the ones that are being put at risk when someone chooses not to vaccinate their children. It's especially bad at a doctor's office since that's where people take their kids when they get sick, and if they bump into kids waiting to get vaccinated, or who's vaccination didn't take, they've suddenly endangered more than just their own child with your decision. I'd be more OK with people refusing vaccines if they would voluntarily quarantine themselves when sick and there were criminal negligence charges available if they're decision affected anyone else.
I think everyone wants the vax/antivax argument to be discretely binary, but it's really more of a nuanced continuum tha
You assume that getting a vaccine means you get immunity. This is not true. Since vaccine efficacy is not 100%, herd immunity side effects are required to protect those that vaccine didn't work on (plus any vulnerable populations). Yes. A successful vaccine means immunity, but there's really no way to know if your vaccine was successful unless you get the disease or go through extensive extra tests. Also, immune systems are not binary, partial efficacy is possible as well, in which case herd immunity is an additional buffer to help protect you.
Finally, there's mutations and evolution of the diseases. Even if you're immune to a strain of disease, a population raging with the disease will quickly produce a strain that you are not immune to. Herd immunity helps to keep that spread and evolution of disease strains in check.
Always the brilliance right after I run out of mod points! Thank for such a well phrased and clear minded post.
Well, they wouldn't be asked to assault US civilians. They would be told to assault crazy terrorists who, by the way, have been threatening their families and loved ones. The one thing we're good at is controlling information access for our troops. Sure there are a few that think the Iraq was a bad thing, but most are quite proud of what we've "accomplished" over there.
Remember Abu Ghraib, the only way that happens is significant informational and psychological conditioning to dehumanize the 'enemy'. That can happen for anyone they fight, anywhere on earth, including inside our borders.
It was a soldier who was ordered to kill a us citizen in a drone attack, and did so without hesitation.
Good lord man, you have to post this gem while I have no mod points. Someone please mod this post up!
Bah, look at the labor put downs during the industrial revolution. The army is good at one thing only, following orders. The soldiers themselves probably won't have any idea what's really going on until long after it's over (if they ever get over the misinformation they've been fed).
Just look at how our military, in general, treated the people of Iraq or Afghanistan. Lots of news stories about unethical behavior. In order to wage war you need to dehumanize the enemy, and then the soldiers do whatever they want to them. That will easily happen with any group inside the US that the army would be ordered to fight. They wouldn't be citizens they'd be "evil animals threatening your family, friends, and way of life, we must stop them now!"
All this aside, the executive has already ordered the assassination of American citizens so it's hard to imagine any group even being allowed to get to that stage of organization in the first place. (see: red scare and McCarthyism, WW II Japanese American camps, etc. for how these things are handled historically)
Natural physical monopolies like what you describe should be publicly owned or have line sharing requirements. Preferebly both. This goes for telecoms, power infrastructure, roads, airports, etc. (we got roads mostly right, but the same reasoning applies to all.)
Turns out doing a thousand lines of heavy DSP INSIDE the interrupt handler
I'm not a computer engineer, or EE, or signals person, but even I can tell you why that would fail. Sounds like the problem is not that people program general purpose chips, but that someone hired general purpose programmers to write DSP code. DSP is a beast, it's its own subset of programming that requires knowledge of physics (waves and nyquist) that some programmers don't know. Or maybe the problem really was that your programmers don't know what a bloody interrupt is, and shouldn't even be writing embedded programs.
As you say, it requires several skill sets at once. Embedded DSP requires extensive theory and mathematical knowledge so you KNOW the stuff, as well as low level hardware knowledge of the target device (either in an FPGA or DSP or GPP) so the code is efficient and fast enough to meet the needs of the platform.
There's more than a few of our signal processing gurus with PHd's that are absolutely brilliant, but have no idea about low level hardware. They can develop algorithms and model them in matlab to prove the ideas are possible. They've taught themselves C, or taken a class or two, so they can program stuff and hack it together. But when their test code tries to get used 'for real', the design problems show up.
Knowing some C does not an embedded programmer make.