You say that until the power grid fails and your generator fails to kick on, leaving you with only battery backup in place.
That's why you have a plan of regular testing for your generator. You do have one of those and follow it, yes? (And remember to keep it fueled too, lest you be visited by the facepalm moment of clarity...)
lucky i didn't say supercomputing then or else i'd have really looked like a fool
Yes, but it was still precisely irrelevant to an article on supercomputing. The requirements there are really quite different to both home/office computing and also to general servers. For one thing, supercomputing is far more likely to be CPU-bound or interconnect-bound than the other cases I've just listed, which are virtually always either external-IO-bound or (if you're unlucky) memory-bound. This gives rise to a very different set of challenges: in particular, heat management is a massive issue because supercomputers are always packed as tightly together as possible to minimize interconnect communication delays. (Damn you, speed of light!)
I've got to take issue with this statement. Anything that takes over a couple years probably should not be started on new silicon as it doesnt make sense to start them yet due to Moores law. The guy that starts the same project a year from now using the same amount of money that you used will beat you to the final calculation and get the hookers and blow that you thought that you deserved.
He's not talking about the time to do one run of the calculation, but rather the time to develop the code and the time to let it be used by the target userbase for a while doing various things while supporting them. In that situation, there's no reason to not get started and to write portable code: you just transfer to new hardware when it becomes available and pick up all the Moore's Law improvements when you upgrade the kit.
What kind of war does Spain anticipate fighting in which a submarine would play a useful role?
Any war involving significant naval action is likely to involve submarines. You build the subs ahead of time because you don't really want to wait until the war starts to begin you defense procurement. (What war? I don't know, nor does anyone else, but with the sorts of long lead times involved you really can't wait for exactitude on that sort of thing.)
That would be the last place to start, as it would cost a fortune to replace all of the highway signs. Not only that, but also all of the mile markers, for which most states have every 1/10 of a mile.
Putting a mile marker every 0.1 miles seems like a waste of money to me. Every half mile would be enough, and would quite a bit less. No, if you were going to start changing, the first thing to do would make the informative signs giving distances to major destinations (large cities, international borders) along interstates also give the distance in km, while not saying anything about changing the more local signs giving distance to the next exit and other basic info like that.
Road signs are still mph, horse races are still miles and furlongs and beer is sold by the pint so I think we're happily confused on matters of units this side of the pond.
Almost all bottled beer is sold in 500ml quantities, with 330ml and 250ml also occurring, if fairly rarely. A few premium beers are in pint bottles, but I'm guessing that they'll convert eventually (when their bottle supplier decides to call it a day). Canned beer is virtually all in 330ml quantities. Only beer sold for immediate consumption in bars is in pint-denominated quantities (mostly because we've got very strict laws on that which would have to be altered) and since our pint is actually now defined in terms of metric units anyway, that's not a big problem.
I'd like a car with a metric speedometer/odometer. The only version of the model I want equipped this way is the Euro spec one. The White House says I can have it. NHTSA, the EPA and my state's DOL can go f*ck themselves.
Mechanical speedometers in Britain are typically dual-marked with both mph and km/h. Mechanical odometers tend to not be, but they're not needed by drivers nearly so frequently. Digital displays tend to be switchable (because that's trivial to do in software).
No, you MUST have laws forcing supermarkets to use specific weights or measures. Otherwise, Shady-Joe's meat market could just shave down their scales and sell you 14 oz of beef instead of a full pound. This is the entire reason we have standardized weights and measures in the first place.
Actually, you have laws that specify the clear meaning of weights and measures, and require that stores properly mark on quantities and so on. The problem isn't that you're getting 14 oz of beef; the problem is that you asked for 16 oz and the store is claiming that that's what you should pay for.
FWIW, this became a legal requirement in the UK 20+(?) years ago when we joined the EU and we have just about assimilated metric measures of volume and weight when it comes to consumables.
With the exception of alcohol sold in bars (which has its own set of regulations) and a few other goods, the quantity has to be marked in metric at reasonable accuracy, but can be sold in any amount and additionally marked with other measures too. Most things are now sold in metric quantities, but nothing in the regulations forces that.
Ever since I was in the Army, I've always written my dates as 12-FEB-09
Where I have a choice, I write dates using a format that produces "12-Feb-2009" as nobody (not even computers) misunderstands that. Failing that, ISO 8601 because it sorts properly without teaching that particular bit of software about time.
Sure, apart from Virgin laying cable to every door. OH WAIT.
They didn't lay cable to every door; Virgin have never sought to provide a universal service. But they did lay cable to mine, and my service is now good enough that the bottleneck is often elsewhere (e.g., at the web server end). Life is sometimes rather good...
Make each team produce a working robot by the end of a 48 hour period.
As long as you accept that some people are going to say "fuck that shit" and do something else. If you've got a bunch of very bright people, they might decide that there is a better priority thing to work on together than any goal you might try to impose. If they do this, you've won as cool and/or necessary stuff will be done. They're probably right about the prioritization, especially if multiple people agree on it, and you've got no chance of changing their mind by waving prizes about or shouting at them or whatever motivational technique you prefer. (Heck, you probably couldn't do it even if you were out-and-out employing them.)
The best part of when they cut a perk and lose programmers is how many of the management seem to think that the pathetic losers quit because they took away the free drinks or some such. Then they get angry when they realize how development has screeched to a halt when the only 3 competent programmers just took off. I have even heard accusations of sabotage.
So... How come you weren't one of the 3 competent programmers who took off so that you could hear the accusations? Or was management so incompetent that they let the accusations be heard outside the company (a good way to end up with a lawsuit, so demonstrating even more thoroughly the incompetence of management)?
Any company that does NOT do everything within the law to minimize their tax burden is both not doing the right thing by their shareholders, and handing their competitors a competitive advantage.
Except that many companies are also not paying dividends either, so there's no tax paid there either. The more you look at the detail of this, the crosser you will feel. I advise stopping before you cause yourself heart trouble...
I'd have thought that, with a name like that, anti-zombie techniques would be more effective. Let's be careful with which type of undead we are killing here!
But corporations are people. Why can't I pay income taxes on my net rather than gross?
Because that's what corporations bought with their "campaign contributions".
To be fair, there are some businesses that simply could not work at all — any kind of deposit-taking institution, such as a credit union or bank, even if not operated by Wall Street — if every piece of money that crosses the doors was taxed as income. Furthermore, there is a general benefit with having companies generally economically active rather than leaving money sitting there doing nothing: taxing net income after allowable outgoings is reasonable. The real problem is that someone found a way (well, several ways) to make income that should be taxed be actually untaxed, and a number of corporations have been using these tricks on a colossal scale. When tax income is generally well up, that's not seen by governments as too big a problem, but when tax income is down, the order comes down from on high to make sure that those who should owe actually pay up. The screaming as the true bills start to be charged will be very loud and shrill indeed.
Ordinary citizens aren't allowed to claim only net income because that would make overall tax income too low; a majority of people have net incomes so low that the taxes levied would be more expensive to collect than what would be earned. It would also make it far easier for the rich to dodge taxes, given that avoidance techniques are now well advanced. (Right now, they have to do games with corporations and loans and tricks like that.) I fear that the net result of all this pain going to be an increased tendency to extremism and an elevated chance of revolution: those are factors that are so random that who knows who'll come out on top?
Also, in reference to the UK, there's a stronger separation between people and companies than in the US; the tax system differentiates so that personal taxes aren't done using the same set of rules at all as companies are taxed under. This greatly reduces the amount of annoyance experienced by individual citizens, since all the rules aimed at complex business arrangements are only experienced by those who are actually running a business: people still grumble about how high taxes are (of course!) but not nearly so much about the complexity.
The value of gold as an intrinsic is based on its nearly unique property of very slow oxidation: it does not readily rust. Which means, if gold were plentiful, it would still have tremendous value: we would start plating cars with the stuff!
It's suitable for use in microelectronics as part of the electrical bonding of chips to their packaging. That means it has real intrinsic value: you can make something with it. It's also not reactive with the body, and so suitable for use in certain types of prosthetics (e.g., teeth) and quite pretty, and so suitable for making adornments and other decorations. Yep, definite intrinsic value there.
Which isn't to say that its current market value is comprised in the majority by its intrinsic value. It's also used as a repository of wealth, which is mostly not related to intrinsic value (heck, silicon dioxide of specified purity could be used for that too; gold's use there is merely conventional) and also for speculation purposes because of that repository use, and that speculation is virtually entirely unconnected to the intrinsic value. Just because gold has some intrinsic value, unlike a fiat currency where value is entirely driven by agreement, it doesn't mean that holding gold bought at a specific price is a good idea. It is not guaranteed to be inflation-proof, and certainly not so when in the throes of an investment bubble.
You now have a number that is divisible by none of the primes, which therefore must be a prime number.
Or the number is divisible by a prime that wasn't in you initial set.
The operation itself is guaranteed to give you a number that is coprime with the initial set. However, if you were to believe that there were a finite set of prime numbers and were then to use that finite set as the input into the coprime generator, you'd get something that is coprime with "all" prime numbers, which would therefore consequently show that there must be at least one prime number that is not in that set, and establish the result as a candidate for the missing prime. (If you previously believed that there are no prime numbers, the product-over-set-and-add-one operation produces 2.) The assumption that must have been false, given that everything else is a basic mathematical or logical operation, is that there is a finite set of prime numbers; there must be an infinite number of primes.
But TFA wasn't talking about this. It was talking about the number of pairs of primes where the difference between the pair is 2, and that's a very non-trivial property.
Now, if Tesla intends to move more mainstream into the market where people are looking for "a commuter car" instead of "a Tesla" - they'll absolutely have trouble by not having a local presence.
They simply don't need to solve that problem in North Carolina yet. There's millions of other potential customers elsewhere. Once they're big enough, they can deal with NC in any way that makes sense (e.g., by delegating the problem to someone who focuses on providing whatever level of support is required in that area) but that doesn't mean that they need to do it right now. And you know what? They won't deal with this right now.
So if Tesla started making wine, it would have to be bought buy a distributor before a liquor store here in MA could buy it and offer it for sale.
I think you'll find that people are far more prepared to travel to purchase an expensive car than they are to get a crate of beer and a bottle of wine for the weekend. (Assuming someone who is actively interested in purchasing both and who has the financial means to do so. Tesla aren't selling cars to street bums.)
Systems biology modeling of cells will require exascale computing
No, it won't because we won't be modeling objects as large as cells at the atomic level. Instead, we will use lots of coupled coarser models, saving the finer ones for parts where "interesting" things are happening (e.g., at membrane interfaces). People are already doing this sort of thing, but at a very coarse scale and with only very limited numbers of fine simulations.
Of course, I happen to think that the really interesting things happen when you scale up to modeling a whole tissue, or a whole organ, or even a whole system. That's where you stand a chance of going from academic pussyfooting around to something useful to ordinary people.
I think it took me 1-2 days to get up to speed from C# from C++. Not sure how long it'd take me to pick up java but I'd expect a week at most.
To read it? Yeah, with your background if it takes you more than a few days, you're slacking. Being able to write it, and write it well... that takes longer. In particular, learning the associated tooling and common libraries can take a lot of work.
FWIW, there are other OO languages that are considerably more different than that trio (those three have a pretty clear historical relationship to each other, and so share a lot of assumptions). For example, being able to subclass the class of classes makes you reconsider how lots of OO patterns that you think you know work.
Well in a large one where you are in classes of 400 plus students, I would say that individual professors matter less that one where you are in classes of 20.
But there's no reason in principle why a single professor should give tutorials to the entire year, especially at undergraduate level where there are often multiple people in a large department who can teach the same course module. (Lectures can scale up much larger than tutorials do, but the skills for giving a lecture aren't the same as those for running a laboratory session or giving a tutorial.)
Slashdot Mirror
You say that until the power grid fails and your generator fails to kick on, leaving you with only battery backup in place.
That's why you have a plan of regular testing for your generator. You do have one of those and follow it, yes? (And remember to keep it fueled too, lest you be visited by the facepalm moment of clarity...)
lucky i didn't say supercomputing then or else i'd have really looked like a fool
Yes, but it was still precisely irrelevant to an article on supercomputing. The requirements there are really quite different to both home/office computing and also to general servers. For one thing, supercomputing is far more likely to be CPU-bound or interconnect-bound than the other cases I've just listed, which are virtually always either external-IO-bound or (if you're unlucky) memory-bound. This gives rise to a very different set of challenges: in particular, heat management is a massive issue because supercomputers are always packed as tightly together as possible to minimize interconnect communication delays. (Damn you, speed of light!)
I've got to take issue with this statement. Anything that takes over a couple years probably should not be started on new silicon as it doesnt make sense to start them yet due to Moores law. The guy that starts the same project a year from now using the same amount of money that you used will beat you to the final calculation and get the hookers and blow that you thought that you deserved.
He's not talking about the time to do one run of the calculation, but rather the time to develop the code and the time to let it be used by the target userbase for a while doing various things while supporting them. In that situation, there's no reason to not get started and to write portable code: you just transfer to new hardware when it becomes available and pick up all the Moore's Law improvements when you upgrade the kit.
to begin your defense procurement
Meh. I did look at the preview, honest!
What kind of war does Spain anticipate fighting in which a submarine would play a useful role?
Any war involving significant naval action is likely to involve submarines. You build the subs ahead of time because you don't really want to wait until the war starts to begin you defense procurement. (What war? I don't know, nor does anyone else, but with the sorts of long lead times involved you really can't wait for exactitude on that sort of thing.)
That would be the last place to start, as it would cost a fortune to replace all of the highway signs. Not only that, but also all of the mile markers, for which most states have every 1/10 of a mile.
Putting a mile marker every 0.1 miles seems like a waste of money to me. Every half mile would be enough, and would quite a bit less. No, if you were going to start changing, the first thing to do would make the informative signs giving distances to major destinations (large cities, international borders) along interstates also give the distance in km, while not saying anything about changing the more local signs giving distance to the next exit and other basic info like that.
Road signs are still mph, horse races are still miles and furlongs and beer is sold by the pint so I think we're happily confused on matters of units this side of the pond.
Almost all bottled beer is sold in 500ml quantities, with 330ml and 250ml also occurring, if fairly rarely. A few premium beers are in pint bottles, but I'm guessing that they'll convert eventually (when their bottle supplier decides to call it a day). Canned beer is virtually all in 330ml quantities. Only beer sold for immediate consumption in bars is in pint-denominated quantities (mostly because we've got very strict laws on that which would have to be altered) and since our pint is actually now defined in terms of metric units anyway, that's not a big problem.
I'd like a car with a metric speedometer/odometer. The only version of the model I want equipped this way is the Euro spec one. The White House says I can have it. NHTSA, the EPA and my state's DOL can go f*ck themselves.
Mechanical speedometers in Britain are typically dual-marked with both mph and km/h. Mechanical odometers tend to not be, but they're not needed by drivers nearly so frequently. Digital displays tend to be switchable (because that's trivial to do in software).
No, you MUST have laws forcing supermarkets to use specific weights or measures. Otherwise, Shady-Joe's meat market could just shave down their scales and sell you 14 oz of beef instead of a full pound. This is the entire reason we have standardized weights and measures in the first place.
Actually, you have laws that specify the clear meaning of weights and measures, and require that stores properly mark on quantities and so on. The problem isn't that you're getting 14 oz of beef; the problem is that you asked for 16 oz and the store is claiming that that's what you should pay for.
Nobody's kilogram is exactly 1 kilogram, either.
Except for the international standard kilogram (held in Paris), and then I'm not sure about that.
FWIW, this became a legal requirement in the UK 20+(?) years ago when we joined the EU and we have just about assimilated metric measures of volume and weight when it comes to consumables.
With the exception of alcohol sold in bars (which has its own set of regulations) and a few other goods, the quantity has to be marked in metric at reasonable accuracy, but can be sold in any amount and additionally marked with other measures too. Most things are now sold in metric quantities, but nothing in the regulations forces that.
Ever since I was in the Army, I've always written my dates as 12-FEB-09
Where I have a choice, I write dates using a format that produces "12-Feb-2009" as nobody (not even computers) misunderstands that. Failing that, ISO 8601 because it sorts properly without teaching that particular bit of software about time.
Sure, apart from Virgin laying cable to every door. OH WAIT.
They didn't lay cable to every door; Virgin have never sought to provide a universal service. But they did lay cable to mine, and my service is now good enough that the bottleneck is often elsewhere (e.g., at the web server end). Life is sometimes rather good...
Make each team produce a working robot by the end of a 48 hour period.
As long as you accept that some people are going to say "fuck that shit" and do something else. If you've got a bunch of very bright people, they might decide that there is a better priority thing to work on together than any goal you might try to impose. If they do this, you've won as cool and/or necessary stuff will be done. They're probably right about the prioritization, especially if multiple people agree on it, and you've got no chance of changing their mind by waving prizes about or shouting at them or whatever motivational technique you prefer. (Heck, you probably couldn't do it even if you were out-and-out employing them.)
The best part of when they cut a perk and lose programmers is how many of the management seem to think that the pathetic losers quit because they took away the free drinks or some such. Then they get angry when they realize how development has screeched to a halt when the only 3 competent programmers just took off. I have even heard accusations of sabotage.
So... How come you weren't one of the 3 competent programmers who took off so that you could hear the accusations? Or was management so incompetent that they let the accusations be heard outside the company (a good way to end up with a lawsuit, so demonstrating even more thoroughly the incompetence of management)?
Any company that does NOT do everything within the law to minimize their tax burden is both not doing the right thing by their shareholders, and handing their competitors a competitive advantage.
Except that many companies are also not paying dividends either, so there's no tax paid there either. The more you look at the detail of this, the crosser you will feel. I advise stopping before you cause yourself heart trouble...
I'd have thought that, with a name like that, anti-zombie techniques would be more effective. Let's be careful with which type of undead we are killing here!
But corporations are people. Why can't I pay income taxes on my net rather than gross?
Because that's what corporations bought with their "campaign contributions".
To be fair, there are some businesses that simply could not work at all — any kind of deposit-taking institution, such as a credit union or bank, even if not operated by Wall Street — if every piece of money that crosses the doors was taxed as income. Furthermore, there is a general benefit with having companies generally economically active rather than leaving money sitting there doing nothing: taxing net income after allowable outgoings is reasonable. The real problem is that someone found a way (well, several ways) to make income that should be taxed be actually untaxed, and a number of corporations have been using these tricks on a colossal scale. When tax income is generally well up, that's not seen by governments as too big a problem, but when tax income is down, the order comes down from on high to make sure that those who should owe actually pay up. The screaming as the true bills start to be charged will be very loud and shrill indeed.
Ordinary citizens aren't allowed to claim only net income because that would make overall tax income too low; a majority of people have net incomes so low that the taxes levied would be more expensive to collect than what would be earned. It would also make it far easier for the rich to dodge taxes, given that avoidance techniques are now well advanced. (Right now, they have to do games with corporations and loans and tricks like that.) I fear that the net result of all this pain going to be an increased tendency to extremism and an elevated chance of revolution: those are factors that are so random that who knows who'll come out on top?
Also, in reference to the UK, there's a stronger separation between people and companies than in the US; the tax system differentiates so that personal taxes aren't done using the same set of rules at all as companies are taxed under. This greatly reduces the amount of annoyance experienced by individual citizens, since all the rules aimed at complex business arrangements are only experienced by those who are actually running a business: people still grumble about how high taxes are (of course!) but not nearly so much about the complexity.
The value of gold as an intrinsic is based on its nearly unique property of very slow oxidation: it does not readily rust. Which means, if gold were plentiful, it would still have tremendous value: we would start plating cars with the stuff!
It's suitable for use in microelectronics as part of the electrical bonding of chips to their packaging. That means it has real intrinsic value: you can make something with it. It's also not reactive with the body, and so suitable for use in certain types of prosthetics (e.g., teeth) and quite pretty, and so suitable for making adornments and other decorations. Yep, definite intrinsic value there.
Which isn't to say that its current market value is comprised in the majority by its intrinsic value. It's also used as a repository of wealth, which is mostly not related to intrinsic value (heck, silicon dioxide of specified purity could be used for that too; gold's use there is merely conventional) and also for speculation purposes because of that repository use, and that speculation is virtually entirely unconnected to the intrinsic value. Just because gold has some intrinsic value, unlike a fiat currency where value is entirely driven by agreement, it doesn't mean that holding gold bought at a specific price is a good idea. It is not guaranteed to be inflation-proof, and certainly not so when in the throes of an investment bubble.
You now have a number that is divisible by none of the primes, which therefore must be a prime number.
Or the number is divisible by a prime that wasn't in you initial set.
The operation itself is guaranteed to give you a number that is coprime with the initial set. However, if you were to believe that there were a finite set of prime numbers and were then to use that finite set as the input into the coprime generator, you'd get something that is coprime with "all" prime numbers, which would therefore consequently show that there must be at least one prime number that is not in that set, and establish the result as a candidate for the missing prime. (If you previously believed that there are no prime numbers, the product-over-set-and-add-one operation produces 2.) The assumption that must have been false, given that everything else is a basic mathematical or logical operation, is that there is a finite set of prime numbers; there must be an infinite number of primes.
But TFA wasn't talking about this. It was talking about the number of pairs of primes where the difference between the pair is 2, and that's a very non-trivial property.
Now, if Tesla intends to move more mainstream into the market where people are looking for "a commuter car" instead of "a Tesla" - they'll absolutely have trouble by not having a local presence.
They simply don't need to solve that problem in North Carolina yet. There's millions of other potential customers elsewhere. Once they're big enough, they can deal with NC in any way that makes sense (e.g., by delegating the problem to someone who focuses on providing whatever level of support is required in that area) but that doesn't mean that they need to do it right now. And you know what? They won't deal with this right now.
So if Tesla started making wine, it would have to be bought buy a distributor before a liquor store here in MA could buy it and offer it for sale.
I think you'll find that people are far more prepared to travel to purchase an expensive car than they are to get a crate of beer and a bottle of wine for the weekend. (Assuming someone who is actively interested in purchasing both and who has the financial means to do so. Tesla aren't selling cars to street bums.)
Systems biology modeling of cells will require exascale computing
No, it won't because we won't be modeling objects as large as cells at the atomic level. Instead, we will use lots of coupled coarser models, saving the finer ones for parts where "interesting" things are happening (e.g., at membrane interfaces). People are already doing this sort of thing, but at a very coarse scale and with only very limited numbers of fine simulations.
Of course, I happen to think that the really interesting things happen when you scale up to modeling a whole tissue, or a whole organ, or even a whole system. That's where you stand a chance of going from academic pussyfooting around to something useful to ordinary people.
I think it took me 1-2 days to get up to speed from C# from C++. Not sure how long it'd take me to pick up java but I'd expect a week at most.
To read it? Yeah, with your background if it takes you more than a few days, you're slacking. Being able to write it, and write it well... that takes longer. In particular, learning the associated tooling and common libraries can take a lot of work.
FWIW, there are other OO languages that are considerably more different than that trio (those three have a pretty clear historical relationship to each other, and so share a lot of assumptions). For example, being able to subclass the class of classes makes you reconsider how lots of OO patterns that you think you know work.
Well in a large one where you are in classes of 400 plus students, I would say that individual professors matter less that one where you are in classes of 20.
But there's no reason in principle why a single professor should give tutorials to the entire year, especially at undergraduate level where there are often multiple people in a large department who can teach the same course module. (Lectures can scale up much larger than tutorials do, but the skills for giving a lecture aren't the same as those for running a laboratory session or giving a tutorial.)