It doesn't matter if the translation happens after tokenization. A 'variable' token won't be the same as a 'keyword' token even if they have the same name.
Wishful thinking. In practice it leads to much more complexity in the translation (lexing/parsing) stage than you'd want to handle. I can guarantee you that the error messages your "naive" users will have to decipher when they make a mistake, and the parser derails, will vastly negate any small advantage from having to learn "jos" than "if".
And it's not like that particular idea hasn't been tried before (in PL/1 and others). It wasn't pretty...
As for Mosaddegh he was about to nationalize the oil fields that we developed and was dealing with the Russians so it was right to depose him at the time.
That's a gross generalisation. In fact what was later to become BP gave Iran a bum deal, and when Mossadeq tried to negotiate BP answered with a flat out "no". (Which made made Acheson quip "Never had so few lost so much so stupidly and so fast.")
Now, also he didn't really talk the the Russians, as much as that's what the British needed Eisenhower to believe to justify American involvement. The US by and large were not that keen to rake the British chestnuts out of the fire, especially as the fire was of their own making. But painting Mossadeq as "soft on communism" did the trick. Selling that is more of a British diplomatic/intelligence success against a sometimes friendly nation than anything else.
This is all fairly well documented. And given what happened later (the Shah turning bad with tacit US approval) it's not too much of a stretch to say that the US "made Iran into a theocracy. (Esp. when you consider the embarassing intelligence failure on the part of the CIA that completely missed the revolution and hence prevented any mitigating actions by the US herself.)
The cheap Windows laptop I'm typing this on has never required more than occasional reboots for updates or crankiness.
So it's like my Linux installation then (Ubuntu 10.04 LTS) only you have to reboot to fix "crankiness" and updates.
I don't. And that's worth it. In my experience it's Windows that doesn't "just work". Linux by and large does. And for the very rare instances where there's a problem, it can be fixed, and finding the fix is easy. (Usually just a Google away.)
My Windows woes always end up with a "Don't know, there's no way to find out, reinstall", or "That's a known problem with no fix available." (Like me recent Win7 slowly eating the SSD partition due to some "backup dlls-if-a-user-should-delete-the-original-one scheme" growing without bounds.) And that's just one.
The woman that he is alleged to have raped has tried to drop the case that she was persuaded to bring.
In Sweden, not for these crimes. They fall under "allmÃnt Ã¥tal", i.e. the prosection doesn't even have to *have* a victim. (Be a bit difficult to bring someone to justice for murder otherwise).
Certain other crimes, while still being criminal and not civil matters, can only proceed as long as the victim presses the charges. But violent crimes such as these (and most others) aren't.
So victims in Sweden don't "bring" a case. They report it. Then it's in the hands of the authorities.
You don't represent Swedish popular opinion any more than I represent British.
He's not claiming that. He's just reporting the news. I.e. the Assange case was already well covered in the news here, so public opinion can be (and was) judged.
Now the Egyptian case was interesting in that it would probably have caused a much bigger stink than it did, if the foreign minister responsible hadn't gotten herself murdered, and the prime minister was near the end anyway. If Anna Lind (former foreign minister) that was next in line to take over the party had lived, the Egyptian affair would have gotten much more play than it ultimately did.
The Assange case is of course quite different. We're not talking "Swedes" here, i.e. non-ethnic (i.e. "bloody foreigners") Swedish citizens that are suspected of terrorism (nobody *really* cares about those, truth be told, and it pains me to see it), but a white Australian (i.e. western) guy "like us".
To see him sent to the US. Won't happen. It'd be political suicide. While Americans are quite popular, or at least, not impopular in Sweden, we have a long tradition of being suspicious (if not down right hostile) against the USA as a nation.
Sorry, that just wooshed straight over my head. (P.S. And thanks for correcting my bad spelling, it's of course Milner with a capical 'M' and no extra characters).
Erlang, of course, is a language designed to be as reliable and fault-tolerant as possible. I didn't know they used it in routers, but apparently some people want to.
Well, it's been used in "routers" for some time.:-)
It's certainly different from your average OO language, but it's no more "a fucking nightmare" than other functional languages like haskell and ocaml
Well, in the words of the guy behind Erlang, Joe Armstrong, "Erlang is a so-so functional language with excelent concurrency and distributed concurrency model" (paraphrase).
As a pure functional language, it shows it's age IMHO - no Hindley mildner type system, uninspired syntax (based on Prolog), etc. etc. But as a systems language for distributed, soft real time, highly available systems, nothing else can currently touch it. (Yes, I used to work at Ericsson and build routers based on Erlang/C.)
Part of that law is the right to wave it if you wish. So if you wave the right to privacy while using workplace computer systems (which you probably did when signing a contract or a click through before login) you ave the company permission to monitor your use.
Again, in my country. No, and it's a general "no". There is a definite limit to what the company can monitor, and you saying "sure go ahead" doesn't make that limit go away. (And furthermore, if you self incriminate by giving access that you shouldn't have been forced/coerced/asked to give, the results can't be used against you.
Now, this has and is not tested in courts over and over, because, drumroll, we treat people like adults in the workplace. And lo and behold, we also act like it.
In many nations, my own included (Sweden). It is not. (This specific case has not been tested, but the general rule has. They can't open your outgoing mail, so why open your outgoing encrypted tunnels?).
You have a reasonable expectation of privacy even when you are at work, and even though you're using company equipment.
By law. And it's funny; It doesn't lead to all the problems that you seem to be plagued by "over there". If you treat someone like an adult, chances are they'll act like one. If you insist on treating them as children, however, that's what you'll get.
Something that was completely impossible to detect and deal with back in the day when management didn't have computers, and we didn't have the internet at work.
It gets modded up because it doesn't *matter* what your company policy says on this matter as it *cannot* trump the *law*. If company policy said that you would consent to being incarcerated, flogged or executed at their discretion that wouldn't be legally valid either.
Some of us live in countries that *do* have laws pertaining to privacy etc. That you don't live in a likewise enlightened nation, isn't our fault.
1. there is clear evidence of the UK being complicit to some extent in extraordinary rendition cases to the US, yet - to my knowledge - no evidence of Sweden being involved.
As a Swede I would very much like for that to be true, but unfortunately it isn't.
They've been stopped ever since 2006, at least as far as we know.
But yes, I agree with the sentiment. He should just come back to Sweden and be done with it. I can't see us just handing him over to the US for a host of reasons. (Even though we've been pushed over on numerous occasions before.)
That artists made any money from recordings was never really true, except for a few really big acts. Witness Roger McGuinn of the Byrds (testimony before the house judiciary committee) to name just one:
In 1973 my work with the Byrds ended. I embarked on a solo recording career
on Columbia Records, and recorded five albums. The only money I've received for
these albums was the modest advance paid prior to each recording.
In 1977 I recorded three albums for Capitol Records in the group "McGuinn Clark
and Hillman." Even though the song "Don't You Write Her Off" was a top 40 hit,
the only money I received from Capitol Records was in the form of a modest advance.
In 1989 I recorded a solo CD, "Back from Rio", for Arista Records. This CD sold
approximately 500,000 copies worldwide, and aside from a modest advance, I have
received no royalties from that project.
So there's nothing new there. Live gigs were always the life blood of any musician in the "recording era".
Any appeal to pure logic for why SI is better than any other system is like trying to justify a political position or for that matter more akin to theology than anything else.
That's the heart of the matter isn't it? And I don't agree. That's a bit like saying that "well, the Japanese seem to do well with their way of writing, so it*s obviously equal in ever way when it comes to performance compared to the alternatives. (Hint; It's worse in many ways). Or, "Well all sufficiently semantically advanced programming languages are Turing complete, and hence it doesn't matter whether we write this is Brainfuck or Haskell. They're equivalent". (No, the similarities when it comes to utility ends with Turing). Or "well it's just a matter of notation anyway, we might as well do arithmetic using roman numerals instead of the Arabic we're using now" (Nope, not by a long shot, even though you can say 4 = IV).
Do you want me to go on? If humans are supposed to use it, shit like this matters. Now, the SI was based on the experience we had had with the previous systems (going back centuries) and their shortcomings, so OF COURSE it's an improvement. It'd be a sorry state of affairs if it wasn't. Is it perfect? Nope, but it's good enough by far.
Now to say that the US is falling behind in STEM because of that, I find a bit rich myself. I see much bigger problems than litres vs. hogsheads. BUT, of course, using backwards notations doesn't help either. It's just one more straw for the camel's back.
Nope. Pound is not both a unit of weight and a unit of force. The two are as distinct as the Newton and the kg. That you use the same sounding name for both of them doesn't fundamentally alter that fact.
Now, there are two ways of specifying specific impulse, one is to base it on the mass of the fuel, then you end up with a Isp as a velocity. If you use weight, you get a time. In either case the conversion factor is 'g' regardless.
This is *exactly* the same regardless of which system of units you work in. That you happen to have a simpler conversion since you're prone to conflating force and weight (lbf/lbm) doesn't really change this. Incidentally, confounding the two leads to a lot of problems in understanding, since your calculations are prone to giving the same result, but conceptually the wrong unit. Slugs notwithstanding. I haven't met a single american trained engineer in industry that actually used those.
But OK, if you're argument is that this unit isn't officially blessed by the SI. Then sure. But we use lots of units that while they are *based* on SI units, aren't officially blessed (fuel consumption is one straightforward example, l/100km). But who cares? The beauty of the SI system is that if you use it as a base, building your own engineering units becomes easy, straight forward, and converting back and forth for various calculations becomes (in general) simple without lots of convoluted conversion factors. You can think in SI, since you can do calculations and conversion in your head.
P.S. And on the divisibility. Guess why a standard kitchen module is 60cm (really 600mm)... Talk about divisibility
I don't know of any major American institution that sticks with imperial units with perhaps the exception of rocket propulsion engineers, who still stick with ISP mesurements in seconds (being pound-seconds of thrust per pound of mass). Guess what... most rocket scientist outside of America report their thrust efficiency in seconds as well.
I don't understand your argument. The second is a perfectly valid SI unit. It's a base unit in fact. And you even use the same definition; those rocket scientists use the SI second for specific impulse. (As opposed to one of the other possibilities).
This is a better example of a unit you didn't manage to screw up than one where we show inconsistency, than anything else...
It's really a shame system. If you don't have enough gold stars or silver turds or whatever, you look bad
I remember from my stint in the telecoms sector, where there was constant whining that we didn't fill in our time report cards. "Very important, that's how we get paid!". (Yeah right, it all got lumped into one account statement when sent up the ladder anyway).
So, they decided it was time for some automated shaming. Every month an automatic email got sent that listed how much unreported time you had. And our boss would follow up with the inevitable; "We need to do this, please, pretty please". However, since his name was always pegged at the number two spot on the list (by a wide margin) only bested by *his* boss, it had the opposite effect. Quite a few of us realised that we didn't really need to do it weekly as our bosses and co-workers apparently couldn't be bothered to do it monthly even... Shaming fail.:-)
And your comparison to Google again doesn't work because you read the stories from those early days and its a lot like Apple in that most of them believed in Page and Brin and NOT were just jazzed about doing free search engine work. Also most believed that Brin and Page had a great idea for a company and wanted in on the ground floor so they were getting compensated in other ways.
I remember it differently. Page and Brin, acording to what I read, shopped the idea around SI-valley for a long time and tried to sell it but there were no takers, as a) it's just an idea and b) "search has been done".
Anyway, I basically agree. People won't do "shit" work for free. But not everyone agrees on what "shit" is. And that's the majority of my point. Now, you're pretty adamant that fixing bugs is and always will be "shit". I don't really have as strong an opinion on the matter as you obviously do, but it's not really my experience, at least not to that degree. (There's the issue of one-upmanship, for instance.)
Typically the "it has to look right" half wins the buying decision.
Oh, it's worse than that. The looks are evaluated based on how it looks when the light is OFF. That is, it has too look good during the day, when it doesn't have a purpose. Doesn't matter if it produces any actual *light* at all...
You can sugar coat it all you want friend but there is no way in hell to make bug hunts anything but long tedious shit jobs because that is what they are.
That sounded almost like a challenge!:-) What I'm saying that if someone manages to find a way to make bugfixing sexy and fun, bug fixing will start to happen. I'm tempted to think about this...
Incidentally, this has nothing to do with free as in beer vs. speech. I think Google managed to attract funding in no small part because they managed to make search fun. You invest in people not ideas after all. And also I would think it depends on what you mean by "do". They were after all grad students. I've left industry to pursue an academic career (such as it is) myself, and it wasn't for the money. Let me tell you...:-)
There'll be plenty of people out there willing to take on the 'shit' jobs (and who won't see them as shit), it's just a matter of breaking down the practical and cultural barriers to getting them involved.
I read somewhere that that's what put Page and Brin apart, and what made them succeed was that they put search into a new light by declaring that it was FUN, instead of this boring buy-as-many-harddisks as you can routine info-processing job.
That resonated with me, because I remember always feeling that Altavista, even though interesting, was boring as hell. You couldn't pay me to get interested in that crap. Now Google on the other hand... (I'm talking back in the day, now not so much).
Slashdot Mirror
It doesn't matter if the translation happens after tokenization. A 'variable' token won't be the same as a 'keyword' token even if they have the same name.
Wishful thinking. In practice it leads to much more complexity in the translation (lexing/parsing) stage than you'd want to handle. I can guarantee you that the error messages your "naive" users will have to decipher when they make a mistake, and the parser derails, will vastly negate any small advantage from having to learn "jos" than "if".
And it's not like that particular idea hasn't been tried before (in PL/1 and others). It wasn't pretty...
As for Mosaddegh he was about to nationalize the oil fields that we developed and was dealing with the Russians so it was right to depose him at the time.
That's a gross generalisation. In fact what was later to become BP gave Iran a bum deal, and when Mossadeq tried to negotiate BP answered with a flat out "no". (Which made made Acheson quip "Never had so few lost so much so stupidly and so fast.")
Now, also he didn't really talk the the Russians, as much as that's what the British needed Eisenhower to believe to justify American involvement. The US by and large were not that keen to rake the British chestnuts out of the fire, especially as the fire was of their own making. But painting Mossadeq as "soft on communism" did the trick. Selling that is more of a British diplomatic/intelligence success against a sometimes friendly nation than anything else.
This is all fairly well documented. And given what happened later (the Shah turning bad with tacit US approval) it's not too much of a stretch to say that the US "made Iran into a theocracy. (Esp. when you consider the embarassing intelligence failure on the part of the CIA that completely missed the revolution and hence prevented any mitigating actions by the US herself.)
So no. The US fucked up. Deal with it and learn.
The cheap Windows laptop I'm typing this on has never required more than occasional reboots for updates or crankiness.
So it's like my Linux installation then (Ubuntu 10.04 LTS) only you have to reboot to fix "crankiness" and updates.
I don't. And that's worth it. In my experience it's Windows that doesn't "just work". Linux by and large does. And for the very rare instances where there's a problem, it can be fixed, and finding the fix is easy. (Usually just a Google away.)
My Windows woes always end up with a "Don't know, there's no way to find out, reinstall", or "That's a known problem with no fix available." (Like me recent Win7 slowly eating the SSD partition due to some "backup dlls-if-a-user-should-delete-the-original-one scheme" growing without bounds.) And that's just one.
The woman that he is alleged to have raped has tried to drop the case that she was persuaded to bring.
In Sweden, not for these crimes. They fall under "allmÃnt Ã¥tal", i.e. the prosection doesn't even have to *have* a victim. (Be a bit difficult to bring someone to justice for murder otherwise).
Certain other crimes, while still being criminal and not civil matters, can only proceed as long as the victim presses the charges. But violent crimes such as these (and most others) aren't.
So victims in Sweden don't "bring" a case. They report it. Then it's in the hands of the authorities.
You don't represent Swedish popular opinion any more than I represent British.
He's not claiming that. He's just reporting the news. I.e. the Assange case was already well covered in the news here, so public opinion can be (and was) judged.
Now the Egyptian case was interesting in that it would probably have caused a much bigger stink than it did, if the foreign minister responsible hadn't gotten herself murdered, and the prime minister was near the end anyway. If Anna Lind (former foreign minister) that was next in line to take over the party had lived, the Egyptian affair would have gotten much more play than it ultimately did.
The Assange case is of course quite different. We're not talking "Swedes" here, i.e. non-ethnic (i.e. "bloody foreigners") Swedish citizens that are suspected of terrorism (nobody *really* cares about those, truth be told, and it pains me to see it), but a white Australian (i.e. western) guy "like us".
To see him sent to the US. Won't happen. It'd be political suicide. While Americans are quite popular, or at least, not impopular in Sweden, we have a long tradition of being suspicious (if not down right hostile) against the USA as a nation.
Sorry, that just wooshed straight over my head. (P.S. And thanks for correcting my bad spelling, it's of course Milner with a capical 'M' and no extra characters).
Erlang, of course, is a language designed to be as reliable and fault-tolerant as possible. I didn't know they used it in routers, but apparently some people want to.
Well, it's been used in "routers" for some time. :-)
It's certainly different from your average OO language, but it's no more "a fucking nightmare" than other functional languages like haskell and ocaml
Well, in the words of the guy behind Erlang, Joe Armstrong, "Erlang is a so-so functional language with excelent concurrency and distributed concurrency model" (paraphrase).
As a pure functional language, it shows it's age IMHO - no Hindley mildner type system, uninspired syntax (based on Prolog), etc. etc. But as a systems language for distributed, soft real time, highly available systems, nothing else can currently touch it. (Yes, I used to work at Ericsson and build routers based on Erlang/C.)
Part of that law is the right to wave it if you wish. So if you wave the right to privacy while using workplace computer systems (which you probably did when signing a contract or a click through before login) you ave the company permission to monitor your use.
Again, in my country. No, and it's a general "no". There is a definite limit to what the company can monitor, and you saying "sure go ahead" doesn't make that limit go away. (And furthermore, if you self incriminate by giving access that you shouldn't have been forced/coerced/asked to give, the results can't be used against you.
Now, this has and is not tested in courts over and over, because, drumroll, we treat people like adults in the workplace. And lo and behold, we also act like it.
Should it even be legal?
In many nations, my own included (Sweden). It is not. (This specific case has not been tested, but the general rule has. They can't open your outgoing mail, so why open your outgoing encrypted tunnels?).
You have a reasonable expectation of privacy even when you are at work, and even though you're using company equipment.
By law. And it's funny; It doesn't lead to all the problems that you seem to be plagued by "over there". If you treat someone like an adult, chances are they'll act like one. If you insist on treating them as children, however, that's what you'll get.
Something that was completely impossible to detect and deal with back in the day when management didn't have computers, and we didn't have the internet at work.
Not.
It gets modded up because it doesn't *matter* what your company policy says on this matter as it *cannot* trump the *law*. If company policy said that you would consent to being incarcerated, flogged or executed at their discretion that wouldn't be legally valid either.
Some of us live in countries that *do* have laws pertaining to privacy etc. That you don't live in a likewise enlightened nation, isn't our fault.
Because it would be against the *law*. And their policy obviously state that breaches of the law are to be reported to the police.
Well, not necessary, but a live human is hard to hold down when you try to cut away a piece of their major blood vessels.
That's what the anesthesiologist is for, silly.
1. there is clear evidence of the UK being complicit to some extent in extraordinary rendition cases to the US, yet - to my knowledge - no evidence of Sweden being involved.
As a Swede I would very much like for that to be true, but unfortunately it isn't.
They've been stopped ever since 2006, at least as far as we know.
But yes, I agree with the sentiment. He should just come back to Sweden and be done with it. I can't see us just handing him over to the US for a host of reasons. (Even though we've been pushed over on numerous occasions before.)
So there's nothing new there. Live gigs were always the life blood of any musician in the "recording era".
Any appeal to pure logic for why SI is better than any other system is like trying to justify a political position or for that matter more akin to theology than anything else.
That's the heart of the matter isn't it? And I don't agree. That's a bit like saying that "well, the Japanese seem to do well with their way of writing, so it*s obviously equal in ever way when it comes to performance compared to the alternatives. (Hint; It's worse in many ways). Or, "Well all sufficiently semantically advanced programming languages are Turing complete, and hence it doesn't matter whether we write this is Brainfuck or Haskell. They're equivalent". (No, the similarities when it comes to utility ends with Turing). Or "well it's just a matter of notation anyway, we might as well do arithmetic using roman numerals instead of the Arabic we're using now" (Nope, not by a long shot, even though you can say 4 = IV).
Do you want me to go on? If humans are supposed to use it, shit like this matters. Now, the SI was based on the experience we had had with the previous systems (going back centuries) and their shortcomings, so OF COURSE it's an improvement. It'd be a sorry state of affairs if it wasn't. Is it perfect? Nope, but it's good enough by far.
Now to say that the US is falling behind in STEM because of that, I find a bit rich myself. I see much bigger problems than litres vs. hogsheads. BUT, of course, using backwards notations doesn't help either. It's just one more straw for the camel's back.
Pound is not both a unit of weight and a unit of force
Sorry, typo. Meant "weight/force and mass"...
Nope. Pound is not both a unit of weight and a unit of force. The two are as distinct as the Newton and the kg. That you use the same sounding name for both of them doesn't fundamentally alter that fact.
Now, there are two ways of specifying specific impulse, one is to base it on the mass of the fuel, then you end up with a Isp as a velocity. If you use weight, you get a time. In either case the conversion factor is 'g' regardless.
This is *exactly* the same regardless of which system of units you work in. That you happen to have a simpler conversion since you're prone to conflating force and weight (lbf/lbm) doesn't really change this. Incidentally, confounding the two leads to a lot of problems in understanding, since your calculations are prone to giving the same result, but conceptually the wrong unit. Slugs notwithstanding. I haven't met a single american trained engineer in industry that actually used those.
But OK, if you're argument is that this unit isn't officially blessed by the SI. Then sure. But we use lots of units that while they are *based* on SI units, aren't officially blessed (fuel consumption is one straightforward example, l/100km). But who cares? The beauty of the SI system is that if you use it as a base, building your own engineering units becomes easy, straight forward, and converting back and forth for various calculations becomes (in general) simple without lots of convoluted conversion factors. You can think in SI, since you can do calculations and conversion in your head.
P.S. And on the divisibility. Guess why a standard kitchen module is 60cm (really 600mm)... Talk about divisibility
I don't know of any major American institution that sticks with imperial units with perhaps the exception of rocket propulsion engineers, who still stick with ISP mesurements in seconds (being pound-seconds of thrust per pound of mass). Guess what... most rocket scientist outside of America report their thrust efficiency in seconds as well.
I don't understand your argument. The second is a perfectly valid SI unit. It's a base unit in fact. And you even use the same definition; those rocket scientists use the SI second for specific impulse. (As opposed to one of the other possibilities).
This is a better example of a unit you didn't manage to screw up than one where we show inconsistency, than anything else...
It's really a shame system. If you don't have enough gold stars or silver turds or whatever, you look bad
I remember from my stint in the telecoms sector, where there was constant whining that we didn't fill in our time report cards. "Very important, that's how we get paid!". (Yeah right, it all got lumped into one account statement when sent up the ladder anyway).
So, they decided it was time for some automated shaming. Every month an automatic email got sent that listed how much unreported time you had. And our boss would follow up with the inevitable; "We need to do this, please, pretty please". However, since his name was always pegged at the number two spot on the list (by a wide margin) only bested by *his* boss, it had the opposite effect. Quite a few of us realised that we didn't really need to do it weekly as our bosses and co-workers apparently couldn't be bothered to do it monthly even... Shaming fail. :-)
And your comparison to Google again doesn't work because you read the stories from those early days and its a lot like Apple in that most of them believed in Page and Brin and NOT were just jazzed about doing free search engine work. Also most believed that Brin and Page had a great idea for a company and wanted in on the ground floor so they were getting compensated in other ways.
I remember it differently. Page and Brin, acording to what I read, shopped the idea around SI-valley for a long time and tried to sell it but there were no takers, as a) it's just an idea and b) "search has been done".
Anyway, I basically agree. People won't do "shit" work for free. But not everyone agrees on what "shit" is. And that's the majority of my point. Now, you're pretty adamant that fixing bugs is and always will be "shit". I don't really have as strong an opinion on the matter as you obviously do, but it's not really my experience, at least not to that degree. (There's the issue of one-upmanship, for instance.)
Typically the "it has to look right" half wins the buying decision.
Oh, it's worse than that. The looks are evaluated based on how it looks when the light is OFF. That is, it has too look good during the day, when it doesn't have a purpose. Doesn't matter if it produces any actual *light* at all...
You can sugar coat it all you want friend but there is no way in hell to make bug hunts anything but long tedious shit jobs because that is what they are.
That sounded almost like a challenge! :-) What I'm saying that if someone manages to find a way to make bugfixing sexy and fun, bug fixing will start to happen. I'm tempted to think about this...
Incidentally, this has nothing to do with free as in beer vs. speech. I think Google managed to attract funding in no small part because they managed to make search fun. You invest in people not ideas after all. And also I would think it depends on what you mean by "do". They were after all grad students. I've left industry to pursue an academic career (such as it is) myself, and it wasn't for the money. Let me tell you... :-)
There'll be plenty of people out there willing to take on the 'shit' jobs (and who won't see them as shit), it's just a matter of breaking down the practical and cultural barriers to getting them involved.
I read somewhere that that's what put Page and Brin apart, and what made them succeed was that they put search into a new light by declaring that it was FUN, instead of this boring buy-as-many-harddisks as you can routine info-processing job.
That resonated with me, because I remember always feeling that Altavista, even though interesting, was boring as hell. You couldn't pay me to get interested in that crap. Now Google on the other hand... (I'm talking back in the day, now not so much).