This guy may be a greater danger to society, but we don't just lock him up again because of that.
Why not?
Lets take this to the logical conclusion.
Suppose this persons level of 'greater danger to society' amounted to a 100% chance of a future murder. Surely you would not want to let them out of prison in that case, right?
Ok, how about a 90% chance? Probably still wouldn't.. right?
80%?
70%?
60%?
Where the hell are you drawing this line of yours?
The fact is that MOST people in former soviet republic countries dont even *use* Internet Explorer, which makes the poster I replied to demonstrably myopic. Sure, maybe he isnt from America after all, but that doesnt change the fact that most of the world isnt stuck on Internet Explorer. IE usage is below 50% worldwide. Thats a fact. He begins with incorrect bullshit.
I'm not so sure about the "unless you have gone quite a few miles" thinking.
These cars also have wired braking systems and I simply have no idea how much pressure these wire systems can put on the brakes in a situation where the sensors never witness the tires slowing down, as I suspect the response to such a condition is to keep increasing the pressure.
As far as putting it in neutral.. maybe he did (are you sure that neutral does what is traditionally expected in these new cars?).. or maybe hes stupid..
That is not the issue on the table. The question is, out of all those plenty of things that they can do, has Apple done any of them?
There is no signed contract here. There is only a shrink-wrapped EULA, and those have only met limited success when tested in court. Still further, does this EULA truly address this issue?
Isnt the point of a good modern language that it offers what you want?
Most of C#'s features extend from things people use in other languages. Just one example is that while C++ was doing template masturbation, C# simply implemented the most common reasons for using templates in the first place, as first-class features.
And now C# is driving its way directly towards the most common reasons for using functional languages. C# may very well end up being THE gold standard language that all others wish they were. Yes I can mix references and pointers. Yes I can mix OO with procedural. And so on.. Thats a fucking good thing.
On top of it all is the most awesome IDE ever to be created. In this case it really IS 'developers, developers, developers!'
They key point is not fixed vs floating point, but rather base-10 vs base-2.
One common standardized datatype is known as Currency in older Microsoft languages (such as VBA), which is a 64-bit fixed point representation using a base of 1/10000th. Another older type is the floating point BCD you mentioned (there are several different common implementations, but they are not interchangeable due to encoding specifics, such as how sign and decimal point are represented)
Fixed-point in-and-of itself doesnt solve the problem. Its not a magic bullet. Typically fixed point is a power-of-two in every other area of usage, for obvious efficiency reasons. Many CPU's (especially DSP's) even have instructions for working with these directly.
the fact that 2.0/2.0 != 1 on almost all FPU's today.
Facts are supposed to be true.
You have taken a little bit of knowledge (that you shouldnt normally compare floats for equality) but have raced to the wrong conclusion.
2.0/2.0 is equal to 1.0 and no modern IEEE-compliant FPU will tell you any different, either.
If we changed your argument to, say, (0.2*10/2.0) != 1.0 then you would be right, if the programming language specified that these calculations were to be performed in IEEE floats or some other base-2 type (because 0.2 cannot be accurately represented)
Just about every financial institution has software designers that will patently disagree with you, and they'd be right.
You use the right tool for the job. A base-2 encoding (such as all those IEEE floating point formats) of 0.1 doesnt work, but a base-10 encoding (such as the datatypes used by financial institutions) does.
Most financial institutions are 100% accurate to at least 5 decimal digits beyond the decimal point, and in most transactions, spare accuracy is rounded to this limit.
These people wanted fat pipes, but private enterprise wasn't going to give it to them. So like any free group of people willing to pay the costs necessary to get what they wanted, they started gathering the money necessary to do it themselves.
Also keep in mind the 'bait and switch' nature of this all.
The State of Washington has had motive to allow Microsoft to do this for years and year, because it isn't like Microsoft generates zero tax revenue in Washington. Far from it, their payroll taxes must be a significant figure on the States revenue books.
The last thing they want is for Microsoft to move out. They have been satisfied with their share of this pie for many years, not wanting to disrupt the status quo and risk a real restructuring of Microsoft elsewhere.
But now they want to seize 1 billion dollars as a retroactive penalty, precisely when the state is in the most dire financial trouble its been in for decades? Please.
They didnt see a problem with it last year, so fuck them. Its OK for them to change the law and nail Microsoft on future money, but its not OK for them to have been complicit all these years essentially changing the rules of the game retroactively.
Which means that there may be 4 levels of charge inside the gate and you have to measure it precisely to tell which value it has.
Precisely?
They just have to distinguish between the charges with high reliability, which is way different than precise measurement.
This same concept has already been applied to just about every digital to analog communications strategy currently employed over the past 50 or so years.
This is in effect a digital to analog problem until they start using single electron increments for charge differentiations, which is never going to happen.
The fact is that MLC have two routes for improvements.. more levels per cell OR more cells per volume of space, and if you know a bit about information theory you know that these two seemingly different things are really the same thing. Cells can get a LOT smaller, therefore they can also handle a lot more levels at their current size.
Slashdot Mirror
Google Chrome, now with 30% faster version number increments.
This guy may be a greater danger to society, but we don't just lock him up again because of that.
Why not?
Lets take this to the logical conclusion.
Suppose this persons level of 'greater danger to society' amounted to a 100% chance of a future murder. Surely you would not want to let them out of prison in that case, right?
Ok, how about a 90% chance? Probably still wouldn't.. right?
80%?
70%?
60%?
Where the hell are you drawing this line of yours?
....or hang him at high noon.
You call it a flame because you are an American?
The fact is that MOST people in former soviet republic countries dont even *use* Internet Explorer, which makes the poster I replied to demonstrably myopic. Sure, maybe he isnt from America after all, but that doesnt change the fact that most of the world isnt stuck on Internet Explorer. IE usage is below 50% worldwide. Thats a fact. He begins with incorrect bullshit.
The thing is, most Americans see Internet Explorer as 'The Internet'
Fixed that for you.
I'm not so sure about the "unless you have gone quite a few miles" thinking.
.. or maybe hes stupid..
These cars also have wired braking systems and I simply have no idea how much pressure these wire systems can put on the brakes in a situation where the sensors never witness the tires slowing down, as I suspect the response to such a condition is to keep increasing the pressure.
As far as putting it in neutral.. maybe he did (are you sure that neutral does what is traditionally expected in these new cars?)
Is there a particular reason that you begin with misinformation?
More likely is that they arent telling you everything that they do online, such as visiting granny midget porn sites.
There is plenty Apple can do about it.
That is not the issue on the table. The question is, out of all those plenty of things that they can do, has Apple done any of them?
There is no signed contract here. There is only a shrink-wrapped EULA, and those have only met limited success when tested in court. Still further, does this EULA truly address this issue?
They didn't write the basic compiler, it was copied and badly copied at that.
What BASIC compiler were they copying from circa 1975 for the 8080?
You *DID* know that Microsoft was producing a BASIC compiler in 1975, right?
So you didnt actually have a retort to his point, right?
Isnt the point of a good modern language that it offers what you want?
Most of C#'s features extend from things people use in other languages. Just one example is that while C++ was doing template masturbation, C# simply implemented the most common reasons for using templates in the first place, as first-class features.
And now C# is driving its way directly towards the most common reasons for using functional languages. C# may very well end up being THE gold standard language that all others wish they were. Yes I can mix references and pointers. Yes I can mix OO with procedural. And so on.. Thats a fucking good thing.
On top of it all is the most awesome IDE ever to be created. In this case it really IS 'developers, developers, developers!'
They key point is not fixed vs floating point, but rather base-10 vs base-2.
One common standardized datatype is known as Currency in older Microsoft languages (such as VBA), which is a 64-bit fixed point representation using a base of 1/10000th. Another older type is the floating point BCD you mentioned (there are several different common implementations, but they are not interchangeable due to encoding specifics, such as how sign and decimal point are represented)
Fixed-point in-and-of itself doesnt solve the problem. Its not a magic bullet. Typically fixed point is a power-of-two in every other area of usage, for obvious efficiency reasons. Many CPU's (especially DSP's) even have instructions for working with these directly.
32-bit floats have 23 bit mantissa's + the implied bit (always 1) = 24 bits of accuracy
the fact that 2.0/2.0 != 1 on almost all FPU's today.
Facts are supposed to be true.
You have taken a little bit of knowledge (that you shouldnt normally compare floats for equality) but have raced to the wrong conclusion.
2.0/2.0 is equal to 1.0 and no modern IEEE-compliant FPU will tell you any different, either.
If we changed your argument to, say, (0.2*10/2.0) != 1.0 then you would be right, if the programming language specified that these calculations were to be performed in IEEE floats or some other base-2 type (because 0.2 cannot be accurately represented)
Just about every financial institution has software designers that will patently disagree with you, and they'd be right.
You use the right tool for the job. A base-2 encoding (such as all those IEEE floating point formats) of 0.1 doesnt work, but a base-10 encoding (such as the datatypes used by financial institutions) does.
Most financial institutions are 100% accurate to at least 5 decimal digits beyond the decimal point, and in most transactions, spare accuracy is rounded to this limit.
Your definition of profit is too limited, and I believe this stems from your over-stressing 'capitalism' instead of 'free market'
These people are willing to pay for the service they demand. Period.
umm hello?
The article is about a city, Monticello, that did exactly what you claim that there is "no way on Earth" a city would do.
These people wanted fat pipes, but private enterprise wasn't going to give it to them. So like any free group of people willing to pay the costs necessary to get what they wanted, they started gathering the money necessary to do it themselves.
This is capitalism at its finest.
Perhaps you didn't wrap it around your peni^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H^H drape it correctly.
Also keep in mind the 'bait and switch' nature of this all.
The State of Washington has had motive to allow Microsoft to do this for years and year, because it isn't like Microsoft generates zero tax revenue in Washington. Far from it, their payroll taxes must be a significant figure on the States revenue books.
The last thing they want is for Microsoft to move out. They have been satisfied with their share of this pie for many years, not wanting to disrupt the status quo and risk a real restructuring of Microsoft elsewhere.
But now they want to seize 1 billion dollars as a retroactive penalty, precisely when the state is in the most dire financial trouble its been in for decades? Please.
They didnt see a problem with it last year, so fuck them. Its OK for them to change the law and nail Microsoft on future money, but its not OK for them to have been complicit all these years essentially changing the rules of the game retroactively.
If it was an attempt at 128 cores, some of them would come off the fab with no defects and would be sold as 128's...
They arent going to intentionally roast up to 28 cores on every unit just to hit their advertised number.
I was indeed trying to be modded funny.
Mission accomplished.
In other words, what did they switch from.
They switched from capitalism to communism, silly.
Which means that there may be 4 levels of charge inside the gate and you have to measure it precisely to tell which value it has.
Precisely?
They just have to distinguish between the charges with high reliability, which is way different than precise measurement.
This same concept has already been applied to just about every digital to analog communications strategy currently employed over the past 50 or so years.
This is in effect a digital to analog problem until they start using single electron increments for charge differentiations, which is never going to happen.
The fact is that MLC have two routes for improvements.. more levels per cell OR more cells per volume of space, and if you know a bit about information theory you know that these two seemingly different things are really the same thing. Cells can get a LOT smaller, therefore they can also handle a lot more levels at their current size.