Baloney. The atomic unit (sector size) of every modern HD is 512 bytes. Thus all HD operations (read/write) occur on base-2 aligned sector boundaries.
And that's where it ends. Everything else is pretty much arbitrary. Okay, there's LBA as well, perhaps, but when you get to drive geometry there are no powers of 2.
It would have been nice if they'd have thought about this problem earlier. But I'm not going to knock Hard disk manufacturers for getting it right, even if it is just for marketting purposes. I'm more inclined to criticise RAM manufacturers for getting it wrong for marketting purposes all those years ago, and sticking with a flawed system.
Because everyone damn well knew what "kilobyte", "megabyte" and so on were supposed to mean, and the only reason the meanings "changed" was because some fuck in marketing saw a good excuse to inflate quoted hard drive capacity.
Marketing corrected something that engineers got wrong (albeit for the wrong reasons). Precision should be more important to engineers rather than considering 1024 bytes to be close enough to call it a kilobyte.
Yeah, you can nit-pick that kilo et al should only apply to powers of 10 (or 1000)
No, by definition the kilo prefix means 1000. It's not a nitpick. It's vagueness on the part of computer manufacturers.
but the 2^10-based system makes a hell of a lot more sense in a computer-based context.
Only in RAM. And only because RAM is addressed using a binary system. Hard disks are not. They are an arbitrary size with an arbitrary number of platters, and an arbitrary numbr of cylinders. A base 2 system makes no sense for these.
If the computer manufacturers are so damn keen on the "correct" use of kilo, why the hell aren't they selling memory by the kibibyte, or quoting the capacity in "true" kilo-bytes?
I have no idea. Considering they always seem to be quoting MBs and GBs rather than Megabytes and GigaBytes, I doubt anyone would bat an eyelid if they quoted MiB and GiB instead. And I really don't know why they don't claim that machines come with 536MB rather than 512MB.
Only in addressable RAM. Hard disks are not indexed by a binary memory location, so the hard disk manufacturers see no reason stick to the power of 2 system that RAM manufacturers use. Thus making it consistent with the SI system of units, but inconsistent with the de-facto notation for memory sizes.
In the past, the 2.4% difference in KB didn't matter too much, but when we're talking about petabytes, the difference goes up to 12.6% which is quite significant.
Because of this, the IEC has proposed the term Kibibyte (KiB), Mebibyte (MiB) (and so on) to refer to the power of 2 system, but this doesn't appear to be catching on.
I have no problem with this. What I do see as a problems are:
People try to use analogies as proof (as opposed to illustrations)
When the analogy is flawed, you get half a dozen people attempting to correct it and making it more unweidly in the process rather than saying "No, the analogy sucks, because...".
I doubt its that hard to get hold of someone's file anyway. After all, anyone you apply to for a loan can get hold of it, as can their agents. Since they have to make it avaialable to a large number of people, it's difficult to offer that much protection.
I could be wrong, but it looks like the prison term would only apply if the spam is fraudulent. If this is the case, then I don't have a problem with it. Prison seems reasonable for attempted fraud.
The problem here is that people like to bitch because Valve has found a solution that works to enable easy distribution direct to gamers and gives warezers a hard time.
I'm sorry? Did you read the comment?
The problem here is that Valve have decided to inconvenience the people who buy the product, and refuse to let people play it unless they agree to terms and conditions that are totally skewed in the favour of the company.
As far as non-"all-rights resrved licences" go, a few companies have toyed with GPL and other similar licenses for applications (e.g. Sun with staroffice, and the MPL for Mozilla). But this is the exception rather than the rule. Plus the occasional public spirited company making software free when they retire it.
But these tend to be the exception rather than the rule.
Well, yes. The law has been written and enacted, challenged in the court, and dismissed. Of course it's legal. This is why I was asking about morality. I was questioning whether the law was correct.
"3) Congress carefully considers the meaning of "promotes the progress of arts and science" every time they extend this"
This is the worst one. Congress does no such thing.
Well perhaps, perhaps not. But it's not up to the court to speculate on whether the law does "promotes the progress of arts and science". They could only overrule it on this grounds, if it was quite clear that the obvious result of this extension was detrimental to the progress of arts and science. While there is some evidence to support this position, it could be argued that longer copyright protection has encouraged some publishers to continue publishing their work for longer.
Personally, I think the arguments a load of hokum, but unless I can disprove this and any similar argument, then it has to accepted as the wisdom of congress.
Indeed. We have a cultural notion that the owner of an item remains the owner and has exclusive rights to it whether he needs it or not. You have the moral right to hoard it, and it does have a tangible value, albeit not realised until you either use it or sell it. Someone taking this would be depriving you of this property.
Should this cultuiral notion also be applied to "intellectual property"? A concept that exists in order to encourage publication and distribution of works? Someone making a copy is not depriving you of any tangible property. If you were distributing it, then they would be depriving you of potential sales, but why do you have a moral right to restrict people from making copies when no harm is done to you?
The court rules on whether the act is legal. Not whether it is a good idea. And there's simply no reason for it not to be.
It is a limited monopoly. It is for a limited time.
I guess what we need is an actual abanodware act. Something that makes it legal under certain conditions (e.g. not for commercial use), to make copies of works that are no longer being published, but a change in the law is what is needed. Challenges to existing laws will not succeed.
It's probably best to ask them to clarify this clause.
For example, if you wrote a virus, would they want to take full responsibility for the damage it caused.
What about if you inadvertantly use code that has an expensive licence for commercial use in one of your own creations.
Your HR people may be a little more amenable to rewriting that clause to clarify that it doesn't include work that's not related to the company's core business, and that you do in your own time.
Or how about non-resizable dialogs with a set number of items in a list which displays all of the items minus one.
Watcom does this. The add file requester has a type selector (*.c; *.cpp etc.) that only displays some of the items. Fortunately it comes with a dialog editor that can fix this.
Sure, until ATSC gets replaced by something else using different compression. Then everyone can just buy an upgrade disk to flash the new codec. (Nahh, they'd rather sell a new TV)
NTSC lasted for 50 years. ATSC will probably last close to as long before being replaced. The extra cost of general purpose CPUs seems a little extravagant considering the life expectancy of a TV is only about 10 years.
DSP and Custom designs are great for high-power applications, but they're far more expensive to design and deploy than a programmable design with "enough" horsepower.
It makes more sense to buy smaller cheaper custom components. Economies of scale offer diminishing returns once you get to the millions of units. It may have cost an extra million to design, but if you can produce 1 million chips for each at $1.01 less, then you're up on the deal. If you can force them to upgrade in 10 years time, you win again.
Even if you do want something more general, you'll want to design a component with a specialised instruction set optimised for video decoding.
The write up suggests it is, and lots of comments have concluded this, but it looks like they're actually trying to make something that will work with modern lightweight materials and construction techniques.
Sounds like that's more of a job for a DSP or a custom component. But even if you use a CPU, you don't need more power thana modern CPU that's considered quite slow. Even for High definition.
I remember telling someone caling to offer me cheaper calls that I don't have a telephone.
For some reason, the idea just didn't seem to register. Might have been because I clearly had a phone...
I was also tempted to give my address to someone selling building extensions as "3rd floor, Merideth court...". I thought an extra room jutting out of the side of the building would look quite nice.
Nope. Although you get a pretty good piece of kit for $189. a 500MHz machine with a gig of ram is pretty functional. Hell, you can do quite a bit with a 50MHz machine and 4 megs of RAM. More than enough to play doom at a decent frame rate (except you'd possibly want a monitor).
On the plus side, something that size is very easy to stop, and very easy to steer away from danger. Not getting into an accident in the first place is something that is greatly undervalued as a safety feature.
Some of those, fair enough, but Aston Martins are hand built, and the build quality often suffers as a result (lovely cars though), and Rolls Royces are only assembled in England. Most of it is made in Germany.
Slashdot Mirror
Baloney. The atomic unit (sector size) of every modern HD is 512 bytes. Thus all HD operations (read/write) occur on base-2 aligned sector boundaries.
And that's where it ends. Everything else is pretty much arbitrary. Okay, there's LBA as well, perhaps, but when you get to drive geometry there are no powers of 2.
It would have been nice if they'd have thought about this problem earlier. But I'm not going to knock Hard disk manufacturers for getting it right, even if it is just for marketting purposes. I'm more inclined to criticise RAM manufacturers for getting it wrong for marketting purposes all those years ago, and sticking with a flawed system.
Because everyone damn well knew what "kilobyte", "megabyte" and so on were supposed to mean, and the only reason the meanings "changed" was because some fuck in marketing saw a good excuse to inflate quoted hard drive capacity.
Marketing corrected something that engineers got wrong (albeit for the wrong reasons). Precision should be more important to engineers rather than considering 1024 bytes to be close enough to call it a kilobyte.
Yeah, you can nit-pick that kilo et al should only apply to powers of 10 (or 1000)
No, by definition the kilo prefix means 1000. It's not a nitpick. It's vagueness on the part of computer manufacturers.
but the 2^10-based system makes a hell of a lot more sense in a computer-based context.
Only in RAM. And only because RAM is addressed using a binary system. Hard disks are not. They are an arbitrary size with an arbitrary number of platters, and an arbitrary numbr of cylinders. A base 2 system makes no sense for these.
If the computer manufacturers are so damn keen on the "correct" use of kilo, why the hell aren't they selling memory by the kibibyte, or quoting the capacity in "true" kilo-bytes?
I have no idea. Considering they always seem to be quoting MBs and GBs rather than Megabytes and GigaBytes, I doubt anyone would bat an eyelid if they quoted MiB and GiB instead. And I really don't know why they don't claim that machines come with 536MB rather than 512MB.
Only in addressable RAM. Hard disks are not indexed by a binary memory location, so the hard disk manufacturers see no reason stick to the power of 2 system that RAM manufacturers use. Thus making it consistent with the SI system of units, but inconsistent with the de-facto notation for memory sizes.
In the past, the 2.4% difference in KB didn't matter too much, but when we're talking about petabytes, the difference goes up to 12.6% which is quite significant.
Because of this, the IEC has proposed the term Kibibyte (KiB), Mebibyte (MiB) (and so on) to refer to the power of 2 system, but this doesn't appear to be catching on.
Why does every /. analogy involve a car?
I have no problem with this. What I do see as a problems are:
People try to use analogies as proof (as opposed to illustrations)
When the analogy is flawed, you get half a dozen people attempting to correct it and making it more unweidly in the process rather than saying "No, the analogy sucks, because...".
I doubt its that hard to get hold of someone's file anyway. After all, anyone you apply to for a loan can get hold of it, as can their agents. Since they have to make it avaialable to a large number of people, it's difficult to offer that much protection.
In Soviet Russia, the dead horse flogs YOU.
But if I play without paying, how is this doing more harm than if I don't buy it and don't play it?
I could be wrong, but it looks like the prison term would only apply if the spam is fraudulent. If this is the case, then I don't have a problem with it. Prison seems reasonable for attempted fraud.
The problem here is that people like to bitch because Valve has found a solution that works to enable easy distribution direct to gamers and gives warezers a hard time.
I'm sorry? Did you read the comment?
The problem here is that Valve have decided to inconvenience the people who buy the product, and refuse to let people play it unless they agree to terms and conditions that are totally skewed in the favour of the company.
As far as non-"all-rights resrved licences" go, a few companies have toyed with GPL and other similar licenses for applications (e.g. Sun with staroffice, and the MPL for Mozilla). But this is the exception rather than the rule. Plus the occasional public spirited company making software free when they retire it.
But these tend to be the exception rather than the rule.
Well, yes. The law has been written and enacted, challenged in the court, and dismissed. Of course it's legal. This is why I was asking about morality. I was questioning whether the law was correct.
"3) Congress carefully considers the meaning of "promotes the progress of arts and science" every time they extend this"
This is the worst one. Congress does no such thing.
Well perhaps, perhaps not. But it's not up to the court to speculate on whether the law does "promotes the progress of arts and science". They could only overrule it on this grounds, if it was quite clear that the obvious result of this extension was detrimental to the progress of arts and science. While there is some evidence to support this position, it could be argued that longer copyright protection has encouraged some publishers to continue publishing their work for longer.
Personally, I think the arguments a load of hokum, but unless I can disprove this and any similar argument, then it has to accepted as the wisdom of congress.
Indeed. We have a cultural notion that the owner of an item remains the owner and has exclusive rights to it whether he needs it or not. You have the moral right to hoard it, and it does have a tangible value, albeit not realised until you either use it or sell it. Someone taking this would be depriving you of this property.
Should this cultuiral notion also be applied to "intellectual property"? A concept that exists in order to encourage publication and distribution of works? Someone making a copy is not depriving you of any tangible property. If you were distributing it, then they would be depriving you of potential sales, but why do you have a moral right to restrict people from making copies when no harm is done to you?
The court rules on whether the act is legal. Not whether it is a good idea. And there's simply no reason for it not to be.
It is a limited monopoly. It is for a limited time.
I guess what we need is an actual abanodware act. Something that makes it legal under certain conditions (e.g. not for commercial use), to make copies of works that are no longer being published, but a change in the law is what is needed. Challenges to existing laws will not succeed.
It's probably best to ask them to clarify this clause.
For example, if you wrote a virus, would they want to take full responsibility for the damage it caused.
What about if you inadvertantly use code that has an expensive licence for commercial use in one of your own creations.
Your HR people may be a little more amenable to rewriting that clause to clarify that it doesn't include work that's not related to the company's core business, and that you do in your own time.
No. The correct way of writing the date is number of seconds since midnight, January 1st, 1970.
So, happy 1103932800-1104019200 to all!
Or how about non-resizable dialogs with a set number of items in a list which displays all of the items minus one.
Watcom does this. The add file requester has a type selector (*.c; *.cpp etc.) that only displays some of the items. Fortunately it comes with a dialog editor that can fix this.
Sure, until ATSC gets replaced by something else using different compression. Then everyone can just buy an upgrade disk to flash the new codec. (Nahh, they'd rather sell a new TV)
NTSC lasted for 50 years. ATSC will probably last close to as long before being replaced. The extra cost of general purpose CPUs seems a little extravagant considering the life expectancy of a TV is only about 10 years.
DSP and Custom designs are great for high-power applications, but they're far more expensive to design and deploy than a programmable design with "enough" horsepower.
It makes more sense to buy smaller cheaper custom components. Economies of scale offer diminishing returns once you get to the millions of units. It may have cost an extra million to design, but if you can produce 1 million chips for each at $1.01 less, then you're up on the deal. If you can force them to upgrade in 10 years time, you win again.
Even if you do want something more general, you'll want to design a component with a specialised instruction set optimised for video decoding.
The write up suggests it is, and lots of comments have concluded this, but it looks like they're actually trying to make something that will work with modern lightweight materials and construction techniques.
Sounds like that's more of a job for a DSP or a custom component. But even if you use a CPU, you don't need more power thana modern CPU that's considered quite slow. Even for High definition.
Why not? Having your very own piece of military equipment definitely qualifies for geek coolness!
I remember telling someone caling to offer me cheaper calls that I don't have a telephone.
For some reason, the idea just didn't seem to register. Might have been because I clearly had a phone...
I was also tempted to give my address to someone selling building extensions as "3rd floor, Merideth court...". I thought an extra room jutting out of the side of the building would look quite nice.
Nope. Although you get a pretty good piece of kit for $189. a 500MHz machine with a gig of ram is pretty functional. Hell, you can do quite a bit with a 50MHz machine and 4 megs of RAM. More than enough to play doom at a decent frame rate (except you'd possibly want a monitor).
On the plus side, something that size is very easy to stop, and very easy to steer away from danger. Not getting into an accident in the first place is something that is greatly undervalued as a safety feature.
Some of those, fair enough, but Aston Martins are hand built, and the build quality often suffers as a result (lovely cars though), and Rolls Royces are only assembled in England. Most of it is made in Germany.