MySQL was not designed as a robust relational database, and its creators didn't seem to be intent on making it so, or else they'd have designed it differently. It was designed as a very quick and quite dirty SQL frontend/ISAM backend system to support small, informal databases (or so it seems)
On MySQL doing only "informal" databases, you have a point: it is not good on points like referential integrity. But on "small", I think you are unfair. It scales very well to very large numbers of records/gigabytes, and performs well when doing so. If your challenge is sheer numbers of records, MySQL is good. It is still OK if you have very few developers who understand the entire database and its constraints. If you have many tables with complex relationships and many developers, some of indifferent quality, you need a lot more enforcement than MySQL gives you.
Leaning over to take corners at speed is cool unless it throws everyones coffee into the isle.
On the contrary, tilting over stops the passengers coffeee spilling. Like leaning into the corner on a motorcycle or banking a plane - if you get it right, everything feels perfectly normal to those inside.
The problems come if you fail to tilt when you should - then everybody's coffee does get thrown around, and you have to slow down to non-tilt speed, making you very late. Or more dangerously, if you fail to untilt or tilt when you shouldn't, becasue then you could be tilting into the other guys track just as he comes towards you tilting into your space. It really has to be very, very reliable - and it wasn't.
The "life sized map" goes back at least to Lewis Carroll, in Sylvie and Bruno, back in the mid 19th century, and has been take up by other writers since. I think it enough of a cliche not to need attribution.
Broadcasters would use it. If they could afford it, they would like to record e very frame of every camera shot that they record, and every frame of every channel they transmit. That is too expensive, even on tape - but at the prices quoted, tehy would use it.
Mind you, I have to say that this sounds too good to be true. Million-fold increases in capacity don't come in one jump. They may well come in 20 doublings over 20 years, or perhaps a little faster with the occasional quadrupling.
And I don't think I would trust one single electron with my bit. The little buggers are to slippery. Current generation storage works on a kind of "electon mob rule" you line up a few hundred of the blighters, and if one of them gets out of line, the others soon pull him back. Even if the guy is right about nudging single electrons, You need a lot more thqan that to make some storage.
All the fuss about nanobots is overrated. Two classes of nanobots have existed since time immemorial - non-fueled types (viruses) and the much larger fuelled types (bacteria). If we manage to build nanobots, it will be reinventing something that nature has done before - sonar, cameras (eyes), levers, packaging (eggs) and so on.
If you're talking about project managers, I think they ought to know more than their underlings.
I disagree. I work for a company which produces hardware. The Project Manager for whom I am working is usually a hardware engineer. Often a good hardware engineer. And one who can cut a fair bit of software - in these days of programmable devices, hardware looks a lot like software. But he has not got, and knows he has not got, my software skills. He cannot do my job, and would be wrong to try to.
In multi-discipliniary teams, it is unreasonable to expect anyone to be able to master the all the skills used in the project. Managers have to have only enough skill to judge outcomes, not to do the work themselves.
I was hoping that this would provide some real evidence for Dark Matter. I have a problem with something so massive which, as far as I can see, is invented to explain a single fact: the anomalously fast rotation of galaxies.
But this article doesn't do that. It says, as I understand it, if the rotation of galaxies is caused by dark matter then dark matter has these properties. If the unexpected rotation is caused by something else, then this is just a curious kind of meta-measurement,
It is a bit like the phlogiston theory. If fire were caused by the release of phlogiston, you could measure the mass of phlogiston - and come out with a negative mass. Which is perfectly logical, but counter-intuitive. Further investigation then makes the phlogiston theory even less attractive - but in the short term, the theory can be patched to work.
Unfortunately, I think the blame lies at least in large part with the consumer. As long as people put up with programming errors in a $500 software suite that they would never accept in an $80 DVD player, we will continue to have these problems.
Another blame is featuritis. In a competitive marketplace, Marketing and hence Management want the now features now. Which means that you tnd to start implementing stuff without doing a full design, and implement enough to get a demo version running quickly, and then are not allowed to go back and rewrite before shipping.
In many cases, what marketing want is not defined enough to write a specification in Z. "It must be fast and user friendly" is too soft. So you knock something up and say "how about that" - and they ship it. Formal specifications alone would be a big gain.
I have all my private domains forwarding immediately to GMail. Nobody sees my gmail address, but I use it to read most mail. GMail's spam filter is pretty good, so that, while I get 1-200 spams a day, only about 3-4 make it to the inbox.
For me, the switch to using GMail has been pretty positive. And if I have to change to GoogleMail, I only have to repoint my three private domains.
They did factor in the cost of manufacturing - $8. The others, being fixed rather than per-unit, depend upon the volume sold. If they sell hundreds of thousands, that would be high; if they sell tens of millions, that would be low. They are working only BOM*2 rather than the *4 you remember - which is only viable if you expect huge volumes.
Or, to put it another way, if it cost $8 million to develop (design, software etc. - just a guess) and they are making $80 per unit, they need to sell 100,000 units to break even. 200,000 units gives a 100% ROI, whcih is the sort of return Apple needs to make off *every* product if it is to live. A million units is the sort of figure they need to keep their stock looking cheerful. Given the buzz they have, this doesn't look unlikely.
You're jumping to the conclusion that this is a retail operation, which was not in the original question. I could phone up a company to talk to
My wife/lover/golf partner
The purchasing manager who gave me a good price yesterday
The lawyer who is handling my case
The guy who designed the module I think I have just found a bug in (he wrote it for me)
The building supervisor to tell him there is a hole in his fence
the graphic designer working on my commissionm
My doctor/therapist
All these are cases when phone calls are made to the company which need to get to a particular person who may be on another call, away from their desk etc, so I will have to be put on hold. In none of those cases could somebody else be of any help, except in taking a message if I get tired of waiting.
That presupposes that the people on the end are interchangeable. If I want to talk to *the* guy who knows what I want to akk, bu he is on another call, I would rather be put on hold with decent hold music than routed to another person who may be very nice to me but cannot answer the question/threat/complain/offer I want to deliver. You want the person with your file on their desk, not just any bod.
To answer the question: popular classics, or popular pop (e.g. Beatles). But turn of the damned voice that breaks in every 15 seconds to say how important your call is. If the music is bad, peopel will hang up enyway. If the music is good, don't break into it just as they are beginning to enjoy it.
Wrong, I think. They index the databases using hashes, because otherwise a linear search of the whole database would take forever. But if they get a hit, they pull up the raw data to let a human have an opinion of the quality of the match. It doesn'rt matter during the detection stage, but courts won't take a machine's word for it on the match - they insist on a human experts opinion on the match between suspect and scene-of-crime.
Thirty years ago, the parent of a friend wanted to solve the same problem before the days of small computers. He had a nice house by the river, with a big square lawn reaching down to the waters edge. He hammered a stake into the centre of the lawn, tied a rope to the it, and tied the other end to the mower. Jam the mower throttle open, and the mower goes round in ever-decreasing circles as the rope winds round the stake. It works, for a couple of orbits, so he goes into the house to get the camera to get evidence of his cleverness. Unfortunately, the rope provides a neat rocking pressure on the stake guaranteed to maximise its chances of pulling out of the ground. When he comes back with the camera, there is a nice neat mowed line leading to the water's edge, and a pair of mower handles sticking out of the water.
They don't go into the physics. If, as you say, it only reflects certain frequencise, it can approach black - it just depends upon what you define as the cutoff albedo for black: 10%? 2%? 0.1%?
On the other hand, it might be an angle thing. It absorbs light coming from all angles except a special one, and reflects spherically light coming from that angle. In which case, from all other angles it is black. You could do that by having a set of angled tubes with black exteriors pointing at the projector.
Five years is a long time. What is the quality of like like, and what am I achieving while I am there?
Five years sitting inside a small capsule just to prove it can be done - forget it.
Five years in a moderately cramped environment with good communications, building part of a real space station, participiting in the escape from Earth - you're on.
While danger is not irrelevant, the cause, the goal, is much more relevant. People have taken huge risks for a cause they believe in - and lost, not infrequently. I believe in trying to ensure that humanity is not limited by the finite resources of the Earth. I want humans to inherit the stars. I am prepared to risk quite a lot of danger, and quite a lot of discomfort, in that goal. But not infinite danger, and not infinite discomfot.
So - give me a worthwhile job to do, and I'll sign up.
Old news. Artillerymen refer to "direct fire" - shooting at things you can see - and "indirect fire" - shooting over the horizon. Been going on since the 19th century. Originally you needed spotters, who would signal back how you were doing and correct your aim. That is why the British Army had a Balloon Corps before any radar existed. The original function of the Royal Flying Corps was artillery observation - and the fighters only developed to attack/defend the artillery observation planes.
Now radar does it for you. You can have on-ship radar, land or carrier based radar aircraft. The arcraft, provided they are high enough, can be well away from the target - but need defence.
One of the reason naval vessels still have artillery as well as missiles is that missiles so damned expensive. It doesn't make sense to attack a coaster valued at $100,000 with a misslile costing $1,500,000. Same probably applies to shelling shore targets. Sure, a missile will take out anything you aim it at. But an awful lot of military fiting is just blasting at the enemy to make hime keep down while your kuys move in - covering fire. Dumb shells from a railgun are likely to be maybe 1/1000 of the price of a missile - andto be able to carry thousnads of them rather than the tens of mmissiles you can carry (80 in the ship mentioned). For general mayhem, possibly a lot more valuable.
Doesn't necessary follow. Railguns might have more precise control than exploding propellant, and a missile which depends upon sheer mass kinetic energy may be less easy to push off course than one which delivers explosive.
No weapon hits percisely where aimed. Rather, you can define the size of the circle in which the weapon has a 25%, 50%, 90%... chance of hitting. Referred to as Circular Error Probability, or CEP. Part of the traning of those who lainch heavy weapons, be it heavy artillery in the army, naval artillery, or bombs from the air force, is to know and allow for CEP. Do not fire at enemies with sensitive innicents inside the CEP - or accept that innocent blood will be spilt/
If these weapons are deployed, those CEPs will be evaluated, and the armed forces will know and allow for them. What probability of hitting innocent bystanders is acceptable is essentially a political problem, not a military one. The armed forces will tell the politicians "At this level of risk, I can achieve X. More risk, more unintended casualties, I can achieve more. Less risk, a cleaner result, but less achievement".
The politicians would like to believe that, if they give the armed forces enough equipment, training etc., they can achieve a war in which only the bad guys are killed. When innocents or our guys ae killed, they blame the military for not using all these expensive weapons right. This is rubbich. Whatever weapons you use, if you are using enough force to overcome a determined ememy, some of it is going to go astray and kill the wrong people. Politicians should face up to this: the military know it well enough.
Which means that a cop can stop and question me, at any time, on a whim, claiming 'suspicion' to justify his actions later. And there's no way I can prove otherwise, since it's just my word against his.
That interpretation is not in the judgement as I read it. It gave no right to stop you. All it said was that if he had already stopped you, for reasons under existing law, he is entitled to demand that you tell him your name.
This sort of extrapolation does not help a rational argument. The police already have the right to stop you if they are suspicious - and they have to have that right in order to stop the real bad guys. The provision of proper checks and balances on that power is a very important matter - but not the one under discussion here,
Seriously, this would seem to necessarily take us one step closer to requiring a national ID card in the most extreme of likely outcomes of the court ruling.
According to the article, the judgement said exactly the opposite. You may be required to tell the police officer your name, but not to produces a drivers licence or other identifying document.
Only Adams could have the genius and foresight to record the part before he died and when it wasn't planned to dramatise those later books.
Actually, DA was trying to get the BBC to dramatise the books. It didn't come off during his lifetime, but as part of the planning for it, he did a readthrough in which he "acted" all the voices - including Agrajag, and this was taped. They reckoned they had enough to voice the (relatively minor, but very Adams-ish) part. (From BBC radio today).
Many years ago we bought a development system which supported maximum of six developers. The manufacturer justified this by pointing to research saying that this was the larges sixe of team that could work together on a single project. As you increased the size of the team, the productivity increase associated with each new person fell. The sixth person on the project only increased productivity by 10% of the productivity of the first person on. The seventh perfson decreased productivity.
Their view was that if you want to deploy more people on a project, you have to divide it into sub-projects wiith relatuively much more formal and documented interfaces between the separate teams.
Slashdot Mirror
MySQL was not designed as a robust relational database, and its creators didn't seem to be intent on making it so, or else they'd have designed it differently. It was designed as a very quick and quite dirty SQL frontend/ISAM backend system to support small, informal databases (or so it seems)
On MySQL doing only "informal" databases, you have a point: it is not good on points like referential integrity. But on "small", I think you are unfair. It scales very well to very large numbers of records/gigabytes, and performs well when doing so. If your challenge is sheer numbers of records, MySQL is good. It is still OK if you have very few developers who understand the entire database and its constraints. If you have many tables with complex relationships and many developers, some of indifferent quality, you need a lot more enforcement than MySQL gives you.
Leaning over to take corners at speed is cool unless it throws everyones coffee into the isle.
On the contrary, tilting over stops the passengers coffeee spilling. Like leaning into the corner on a motorcycle or banking a plane - if you get it right, everything feels perfectly normal to those inside.
The problems come if you fail to tilt when you should - then everybody's coffee does get thrown around, and you have to slow down to non-tilt speed, making you very late. Or more dangerously, if you fail to untilt or tilt when you shouldn't, becasue then you could be tilting into the other guys track just as he comes towards you tilting into your space. It really has to be very, very reliable - and it wasn't.
The "life sized map" goes back at least to Lewis Carroll, in Sylvie and Bruno, back in the mid 19th century, and has been take up by other writers since. I think it enough of a cliche not to need attribution.
Broadcasters would use it. If they could afford it, they would like to record e very frame of every camera shot that they record, and every frame of every channel they transmit. That is too expensive, even on tape - but at the prices quoted, tehy would use it.
Mind you, I have to say that this sounds too good to be true. Million-fold increases in capacity don't come in one jump. They may well come in 20 doublings over 20 years, or perhaps a little faster with the occasional quadrupling.
And I don't think I would trust one single electron with my bit. The little buggers are to slippery. Current generation storage works on a kind of "electon mob rule" you line up a few hundred of the blighters, and if one of them gets out of line, the others soon pull him back. Even if the guy is right about nudging single electrons, You need a lot more thqan that to make some storage.
Right, now I've had may say, I'll go and RTFA.
No.
All the fuss about nanobots is overrated. Two classes of nanobots have existed since time immemorial - non-fueled types (viruses) and the much larger fuelled types (bacteria). If we manage to build nanobots, it will be reinventing something that nature has done before - sonar, cameras (eyes), levers, packaging (eggs) and so on.
Bacteria are Darwinianly designed nanobots.
someone needs to create a nano robot who enjoys the stuff found in bathrooms
It's been done - they are called bacteria.
If you're talking about project managers, I think they ought to know more than their underlings.
I disagree. I work for a company which produces hardware. The Project Manager for whom I am working is usually a hardware engineer. Often a good hardware engineer. And one who can cut a fair bit of software - in these days of programmable devices, hardware looks a lot like software. But he has not got, and knows he has not got, my software skills. He cannot do my job, and would be wrong to try to.
In multi-discipliniary teams, it is unreasonable to expect anyone to be able to master the all the skills used in the project. Managers have to have only enough skill to judge outcomes, not to do the work themselves.
I was hoping that this would provide some real evidence for Dark Matter. I have a problem with something so massive which, as far as I can see, is invented to explain a single fact: the anomalously fast rotation of galaxies.
But this article doesn't do that. It says, as I understand it, if the rotation of galaxies is caused by dark matter then dark matter has these properties. If the unexpected rotation is caused by something else, then this is just a curious kind of meta-measurement,
It is a bit like the phlogiston theory. If fire were caused by the release of phlogiston, you could measure the mass of phlogiston - and come out with a negative mass. Which is perfectly logical, but counter-intuitive. Further investigation then makes the phlogiston theory even less attractive - but in the short term, the theory can be patched to work.
Unfortunately, I think the blame lies at least in large part with the consumer. As long as people put up with programming errors in a $500 software suite that they would never accept in an $80 DVD player, we will continue to have these problems.
Another blame is featuritis. In a competitive marketplace, Marketing and hence Management want the now features now. Which means that you tnd to start implementing stuff without doing a full design, and implement enough to get a demo version running quickly, and then are not allowed to go back and rewrite before shipping.
In many cases, what marketing want is not defined enough to write a specification in Z. "It must be fast and user friendly" is too soft. So you knock something up and say "how about that" - and they ship it. Formal specifications alone would be a big gain.
Or better "Noun: a ten letter acrostic"
I have all my private domains forwarding immediately to GMail. Nobody sees my gmail address, but I use it to read most mail. GMail's spam filter is pretty good, so that, while I get 1-200 spams a day, only about 3-4 make it to the inbox.
For me, the switch to using GMail has been pretty positive. And if I have to change to GoogleMail, I only have to repoint my three private domains.
They did factor in the cost of manufacturing - $8. The others, being fixed rather than per-unit, depend upon the volume sold. If they sell hundreds of thousands, that would be high; if they sell tens of millions, that would be low. They are working only BOM*2 rather than the *4 you remember - which is only viable if you expect huge volumes.
Or, to put it another way, if it cost $8 million to develop (design, software etc. - just a guess) and they are making $80 per unit, they need to sell 100,000 units to break even. 200,000 units gives a 100% ROI, whcih is the sort of return Apple needs to make off *every* product if it is to live. A million units is the sort of figure they need to keep their stock looking cheerful. Given the buzz they have, this doesn't look unlikely.
All these are cases when phone calls are made to the company which need to get to a particular person who may be on another call, away from their desk etc, so I will have to be put on hold. In none of those cases could somebody else be of any help, except in taking a message if I get tired of waiting.
That presupposes that the people on the end are interchangeable. If I want to talk to *the* guy who knows what I want to akk, bu he is on another call, I would rather be put on hold with decent hold music than routed to another person who may be very nice to me but cannot answer the question/threat/complain/offer I want to deliver. You want the person with your file on their desk, not just any bod.
To answer the question: popular classics, or popular pop (e.g. Beatles). But turn of the damned voice that breaks in every 15 seconds to say how important your call is. If the music is bad, peopel will hang up enyway. If the music is good, don't break into it just as they are beginning to enjoy it.
Wrong, I think. They index the databases using hashes, because otherwise a linear search of the whole database would take forever. But if they get a hit, they pull up the raw data to let a human have an opinion of the quality of the match. It doesn'rt matter during the detection stage, but courts won't take a machine's word for it on the match - they insist on a human experts opinion on the match between suspect and scene-of-crime.
Thirty years ago, the parent of a friend wanted to solve the same problem before the days of small computers. He had a nice house by the river, with a big square lawn reaching down to the waters edge. He hammered a stake into the centre of the lawn, tied a rope to the it, and tied the other end to the mower. Jam the mower throttle open, and the mower goes round in ever-decreasing circles as the rope winds round the stake. It works, for a couple of orbits, so he goes into the house to get the camera to get evidence of his cleverness. Unfortunately, the rope provides a neat rocking pressure on the stake guaranteed to maximise its chances of pulling out of the ground. When he comes back with the camera, there is a nice neat mowed line leading to the water's edge, and a pair of mower handles sticking out of the water.
They don't go into the physics. If, as you say, it only reflects certain frequencise, it can approach black - it just depends upon what you define as the cutoff albedo for black: 10%? 2%? 0.1%?
On the other hand, it might be an angle thing. It absorbs light coming from all angles except a special one, and reflects spherically light coming from that angle. In which case, from all other angles it is black. You could do that by having a set of angled tubes with black exteriors pointing at the projector.
Five years is a long time. What is the quality of like like, and what am I achieving while I am there?
Five years sitting inside a small capsule just to prove it can be done - forget it.
Five years in a moderately cramped environment with good communications, building part of a real space station, participiting in the escape from Earth - you're on.
While danger is not irrelevant, the cause, the goal, is much more relevant. People have taken huge risks for a cause they believe in - and lost, not infrequently. I believe in trying to ensure that humanity is not limited by the finite resources of the Earth. I want humans to inherit the stars. I am prepared to risk quite a lot of danger, and quite a lot of discomfort, in that goal. But not infinite danger, and not infinite discomfot.
So - give me a worthwhile job to do, and I'll sign up.
Old news. Artillerymen refer to "direct fire" - shooting at things you can see - and "indirect fire" - shooting over the horizon. Been going on since the 19th century. Originally you needed spotters, who would signal back how you were doing and correct your aim. That is why the British Army had a Balloon Corps before any radar existed. The original function of the Royal Flying Corps was artillery observation - and the fighters only developed to attack/defend the artillery observation planes.
Now radar does it for you. You can have on-ship radar, land or carrier based radar aircraft. The arcraft, provided they are high enough, can be well away from the target - but need defence.
One of the reason naval vessels still have artillery as well as missiles is that missiles so damned expensive. It doesn't make sense to attack a coaster valued at $100,000 with a misslile costing $1,500,000. Same probably applies to shelling shore targets. Sure, a missile will take out anything you aim it at. But an awful lot of military fiting is just blasting at the enemy to make hime keep down while your kuys move in - covering fire. Dumb shells from a railgun are likely to be maybe 1/1000 of the price of a missile - andto be able to carry thousnads of them rather than the tens of mmissiles you can carry (80 in the ship mentioned). For general mayhem, possibly a lot more valuable.
Doesn't necessary follow. Railguns might have more precise control than exploding propellant, and a missile which depends upon sheer mass kinetic energy may be less easy to push off course than one which delivers explosive.
No weapon hits percisely where aimed. Rather, you can define the size of the circle in which the weapon has a 25%, 50%, 90%... chance of hitting. Referred to as Circular Error Probability, or CEP. Part of the traning of those who lainch heavy weapons, be it heavy artillery in the army, naval artillery, or bombs from the air force, is to know and allow for CEP. Do not fire at enemies with sensitive innicents inside the CEP - or accept that innocent blood will be spilt/
If these weapons are deployed, those CEPs will be evaluated, and the armed forces will know and allow for them. What probability of hitting innocent bystanders is acceptable is essentially a political problem, not a military one. The armed forces will tell the politicians "At this level of risk, I can achieve X. More risk, more unintended casualties, I can achieve more. Less risk, a cleaner result, but less achievement".
The politicians would like to believe that, if they give the armed forces enough equipment, training etc., they can achieve a war in which only the bad guys are killed. When innocents or our guys ae killed, they blame the military for not using all these expensive weapons right. This is rubbich. Whatever weapons you use, if you are using enough force to overcome a determined ememy, some of it is going to go astray and kill the wrong people. Politicians should face up to this: the military know it well enough.
Which means that a cop can stop and question me, at any time, on a whim, claiming 'suspicion' to justify his actions later. And there's no way I can prove otherwise, since it's just my word against his.
That interpretation is not in the judgement as I read it. It gave no right to stop you. All it said was that if he had already stopped you, for reasons under existing law, he is entitled to demand that you tell him your name.
This sort of extrapolation does not help a rational argument. The police already have the right to stop you if they are suspicious - and they have to have that right in order to stop the real bad guys. The provision of proper checks and balances on that power is a very important matter - but not the one under discussion here,
Seriously, this would seem to necessarily take us one step closer to requiring a national ID card in the most extreme of likely outcomes of the court ruling.
According to the article, the judgement said exactly the opposite. You may be required to tell the police officer your name, but not to produces a drivers licence or other identifying document.
Only Adams could have the genius and foresight to record the part before he died and when it wasn't planned to dramatise those later books.
Actually, DA was trying to get the BBC to dramatise the books. It didn't come off during his lifetime, but as part of the planning for it, he did a readthrough in which he "acted" all the voices - including Agrajag, and this was taped. They reckoned they had enough to voice the (relatively minor, but very Adams-ish) part. (From BBC radio today).
Many years ago we bought a development system which supported maximum of six developers. The manufacturer justified this by pointing to research saying that this was the larges sixe of team that could work together on a single project. As you increased the size of the team, the productivity increase associated with each new person fell. The sixth person on the project only increased productivity by 10% of the productivity of the first person on. The seventh perfson decreased productivity.
Their view was that if you want to deploy more people on a project, you have to divide it into sub-projects wiith relatuively much more formal and documented interfaces between the separate teams.
My experioence would not contradict this at all.