FAT is particularly prone to fragmentation. Finding any free block in FAT is O(n), since you need to scan the free blocks bitmap. Finding an optimally-sized free space is even more complex, since you have to scan the entire free space bitmap and then compare the sizes. Most other filesystems use some kind of tree for extents of free space, making this much easier to implement, so they are more resistant to fragmentation.
Can't you just read the bitmap into memory at mount-time and then deal with it there in whatever format you want?
But even full data journaling isn't going to keep you 100% safe from data loss...
Especially since the data=journal implementation is less used and therefore less debugged. Several years ago, I found that mounting with data=journal inside user-mode-linux inside another data=journal filesystem resulted in filesystem corruption on the inner filesystem. The problem went away when I mounted with data=ordered on the outer filesystem.
64-bit multiplication is 4 times faster on a 64-bit processor than on a 32-bit platform
Do a lot of that do ya?
Yes, I do. Audio and video encoding and decoding both use, among other things, FFTs, which use lots of multiplications. Encryption and decryption use lots of multi-word operations as well, and since Intel's instruction set architecture supports very few general-purpose registers, fitting twice as many bits in register eliminates not only half the instructions, but also the extra register-shuffling instructions that would otherwise be needed.
Most applications do not utilize the additional address space. In fact, the relative speed increase you would get is negligible. The bottle neck is often the speed of the disk controller and IO throughput. It's like getting a 1GBit network card in hopes your download speed from the internet will be faster.
"Address space"... I do not think it means what you think it means. I'm talking about the size of the data word, not the address space.
So you have a need for more than 4GB of RAM in your system?
You realize that 64-bit processors are a _lot_ faster than 32-bit processors, right? For example, a 64-bit multiplication is 4 times faster on a 64-bit processor than on a 32-bit platform. It's not just about address space.
I think we need to move even farther from the ISA. One way that you can do that is with VMs and JIT compiles like Java, Mono, and.Net does. What I would love to see is moving to a system like the old IBM model 38/AS400.
Every program is compiled to a "perfect" ISA. When you first run the program a JIT compiler compiles that ISA to your native ISA and then stores that compilation on your system.
This isn't a problem worth solving if you have the source code to your software.
I'm so tired of this crap
on
Trick or Treatment
·
· Score: 2, Insightful
I agree that there are a lot of quacks that are in it for the money, but when I was in China my friend with a slipped disk was having some serious back pains and went to a doctor of Chinese medicine. After a fire-cupping and drinking herbal teas for a week he helt much better.
Not again...
"I know about the placebo effect, but it worked for me!"
"Censorship is a dangerous tool for the powerful to have, but we need to filter the Internet so we can catch those spammers and traders of music and child porn!"
"Those Mormons and Scientologists are crazy, but MY religion deserves respect!"
Hypocrites. You're just as bad as those you decry.
"Yeah, all those OTHER forms of woo are bunk, but MY pet woo is for real!"
"Never rewrite anything" is great advice for those who have no hope of ever understanding what their code actually does, or---more importantly---what it should do. It's not surprising that Joel, an ex-employee of Microsoft, would give that advice, but that doesn't make it universally applicable.
The reality is that it depends on a number of factors, including on how much technical debt the project has accrued and how this impacts the organization's flexibility. Having simple feature additions blow up in your face because nobody really understands the current code base is an excellent reason to ditch that code base at the earliest opportunity.
There are many other factors to consider, but what I've written suffices to show that a blanket statement like "rewrites considered harmful" is simply wrong.
Or the company could offer to send it to the person via a WHOLE BUNCH of SMS text messages. A very common way to "exchange data" nowadays. LOL!
SMS text messages are arguably not a "medium", and even if they were, they're not customarily used for software interchange.
There's no real strength in a license itself, until there's real court cases to back it up. It's still quite "untested" after all these years, which I find interesting, considering lots of companies have stolen code and used it.
What, do you live under a rock? The GPL has been in court many times. Very few defendants have bothered to carry through to trial, but that's just because their lawyers advised them that it would be a good idea to settle. That shouldn't undermine one's confidence in the enforceability of the license.
"We demand that you immediately provide the complete corresponding machine-readable source code, which must be distributed under the terms of Sections 1 and 2 of GPLv2 on a medium customarily used for software interchange, or we will be forced to initiate legal proceedings."
RMS already thought of that. Even paper tape won't work for you in 2008.
It kinda sounds like the OP is willing to pay for project work, which is therefore not free. The OP appears to simply want work done with, or rather on, software that's OSI compliant.
But why does he want that? If he's willing to pay for implementation, he ought to be willing to pay the developers too. If he wants an open source product so he can add his own extentions, wait a second, he's going to have to pay for development there too. If he doesn't want to do anything to the software, then why does he need the code?
Controlling English writing in Quebec, from what I understand, is accomplished by Quebec's National Assembly (provincial legislative assembly) invoking the notwithstanding clause every 5 years.
And I would hazard a guess that, the whole reason Debian is so stable... is because of it's policies and processes. I would say that APT/DEB has very little to do with this stability.
I'd say it's the reverse: Debian Policy is the reason why APT/DEB are so good. Because Debian Policy could say "thou shalt place a file in/usr/share/menu", stuff like the "menu" package (which centralized the management of various window managers' application menus) actually happened before any other distro had it.
Do the same restrictive regulations apply in less densely-populated areas? Somebody living on an acreage in an unforested area really shouldn't have to deal with the same regulations as someone in, say, a 12-storey apartment building in the middle of a metropolis.
This is just a bunch of people who are religious zealots about the Creative Commons licensing suite who... hate the viral nature of the Free Software Foundation license suite.
Well this won't help those zealots, since the new GFDL only allows migrating to the ShareAlike license, which also has a "viral nature".
Like I said, there are problems with the GFDL, but throwing it away just because a small group (and it is a very small group of individuals pushing to make this change) doesn't like the Free Software Foundation. Unfortunately this small group making this change also have some prominent positions within the WMF.
So you think RMS would change the GFDL just because prominent people at the Wikimedia Foundation (allegely) don't like the FSF? I don't buy it.
You cannot "taste test" fruit in a grocery store, you cannot download copyrighted documents with impunity, and you cannot give these things away to the general public because it is stealing and it is wrong.
Don't be a jackass. If you aren't going to bother addressing the obvious differences between copyright infringement and petty theft, then you don't have any business expressing your ignorant opinion on the subject.
Singlecasting of 30 million copies of CSI via internet is not logical.
Yes, it is, and it's a worthless straw-man argument.
By "it is", I mean "it is illogical". Trying to single-cast digital TV would be insane.
You might as well argue that the DNS can't work because the root servers can't handle every DNS query in the world. The reality is that they don't have to.
Take a typical show like CSI which is watched by 30 million viewers each week. It's broadcast by 200 stations at approximately 15 megabit/s. If broadcast television was discontinued and CSI provided by ABC.com, it would need a bandwidth of 500,000 gigabit/s bandwidth.
Uh, 15 Mbps * 200 = 3000 Mbps, and that's assuming you wouldn't use a better codec, which you would. If you were going to deploy Internet access nationwide as a replacement for over-the-air TV, you would make sure multicast worked.
Singlecasting of 30 million copies of CSI via internet is not logical.
Yes, it is, and it's a worthless straw-man argument.
Here's an idea from left field: Why not treat Internet access as an essential service, and work toward getting it to everyone? You can broadcast TV over the Internet, and Wikipedia access alone would improve the average American's education by an order of magnitude.
Carmack isn't a particularly brilliant programmer. He's just pragmatic and dedicated to putting in the effort to create something.
The code in Q1-Q3 is your standard stuff, nothing special in there and in fact it's too limited in many areas that could use a little more flexibility (ie. it was a bad design).
The code in Q1-Q3 was "standard stuff" because "standard" in that era was defined as "whatever John Carmack does".
John Carmack had a strength most other programmers lack: reading academic papers and applying them. Hell, in that era, reading RFCs made one an above-average programmer.
If you think John Carmack wasn't brilliant, then you haven't examined the work of his contemporaries. Comparing the code he wrote 10 years ago against code others write today isn't fair.
Slashdot Mirror
Ironic that the user is the cause of just about all windows issues.
Sure. If nobody used Windows, there would be no issues.
Seriously, you've clearly never tried to write an innovative program on a Windows box. It's absolutely painful.
FAT is particularly prone to fragmentation. Finding any free block in FAT is O(n), since you need to scan the free blocks bitmap. Finding an optimally-sized free space is even more complex, since you have to scan the entire free space bitmap and then compare the sizes. Most other filesystems use some kind of tree for extents of free space, making this much easier to implement, so they are more resistant to fragmentation.
Can't you just read the bitmap into memory at mount-time and then deal with it there in whatever format you want?
But even full data journaling isn't going to keep you 100% safe from data loss...
Especially since the data=journal implementation is less used and therefore less debugged. Several years ago, I found that mounting with data=journal inside user-mode-linux inside another data=journal filesystem resulted in filesystem corruption on the inner filesystem. The problem went away when I mounted with data=ordered on the outer filesystem.
64-bit multiplication is 4 times faster on a 64-bit processor than on a 32-bit platform
Do a lot of that do ya?
Yes, I do. Audio and video encoding and decoding both use, among other things, FFTs, which use lots of multiplications. Encryption and decryption use lots of multi-word operations as well, and since Intel's instruction set architecture supports very few general-purpose registers, fitting twice as many bits in register eliminates not only half the instructions, but also the extra register-shuffling instructions that would otherwise be needed.
Most applications do not utilize the additional address space. In fact, the relative speed increase you would get is negligible. The bottle neck is often the speed of the disk controller and IO throughput. It's like getting a 1GBit network card in hopes your download speed from the internet will be faster.
"Address space"... I do not think it means what you think it means. I'm talking about the size of the data word, not the address space.
I run 64-bit everywhere.
So you have a need for more than 4GB of RAM in your system?
You realize that 64-bit processors are a _lot_ faster than 32-bit processors, right? For example, a 64-bit multiplication is 4 times faster on a 64-bit processor than on a 32-bit platform. It's not just about address space.
I think we need to move even farther from the ISA. One way that you can do that is with VMs and JIT compiles like Java, Mono, and .Net does. What I would love to see is moving to a system like the old IBM model 38/AS400.
Every program is compiled to a "perfect" ISA. When you first run the program a JIT compiler compiles that ISA to your native ISA and then stores that compilation on your system.
This isn't a problem worth solving if you have the source code to your software.
I agree that there are a lot of quacks that are in it for the money, but when I was in China my friend with a slipped disk was having some serious back pains and went to a doctor of Chinese medicine. After a fire-cupping and drinking herbal teas for a week he helt much better.
Not again...
"I know about the placebo effect, but it worked for me!"
"Censorship is a dangerous tool for the powerful to have, but we need to filter the Internet so we can catch those spammers and traders of music and child porn!"
"Those Mormons and Scientologists are crazy, but MY religion deserves respect!"
Hypocrites. You're just as bad as those you decry.
"Yeah, all those OTHER forms of woo are bunk, but MY pet woo is for real!"
A re-write from scratch is nearly always the wrong approach.
I'm tired of hearing Joel Spolsky's opinion touted as an irrefutable fact.
"Never rewrite anything" is great advice for those who have no hope of ever understanding what their code actually does, or---more importantly---what it should do. It's not surprising that Joel, an ex-employee of Microsoft, would give that advice, but that doesn't make it universally applicable.
The reality is that it depends on a number of factors, including on how much technical debt the project has accrued and how this impacts the organization's flexibility. Having simple feature additions blow up in your face because nobody really understands the current code base is an excellent reason to ditch that code base at the earliest opportunity.
There are many other factors to consider, but what I've written suffices to show that a blanket statement like "rewrites considered harmful" is simply wrong.
However, as the rootkit is armed with an algorithm that periodically generates new domain names where the malware then looks for new instructions . . .
Couldn't the registrars run that algorithm ahead of time and ban (or track down) new registrations for those domains?
"Free laptop upgrades for life"... sounds like "unlimited bandwidth" and "Plays4Sure".
No thanks.
Or the company could offer to send it to the person via a WHOLE BUNCH of SMS text messages. A very common way to "exchange data" nowadays. LOL!
SMS text messages are arguably not a "medium", and even if they were, they're not customarily used for software interchange.
There's no real strength in a license itself, until there's real court cases to back it up. It's still quite "untested" after all these years, which I find interesting, considering lots of companies have stolen code and used it.
What, do you live under a rock? The GPL has been in court many times. Very few defendants have bothered to carry through to trial, but that's just because their lawyers advised them that it would be a good idea to settle. That shouldn't undermine one's confidence in the enforceability of the license.
Win 95 and to some extent 98 were basically 16 bit systems with 32 bit support
No, they were 32-bit systems that allowed you to run 16-bit code in privileged mode.
"We demand that you immediately provide the complete corresponding machine-readable source code, which must be distributed under the terms of Sections 1 and 2 of GPLv2 on a medium customarily used for software interchange, or we will be forced to initiate legal proceedings."
RMS already thought of that. Even paper tape won't work for you in 2008.
Nice try, though.
It kinda sounds like the OP is willing to pay for project work, which is therefore not free. The OP appears to simply want work done with, or rather on, software that's OSI compliant.
But why does he want that? If he's willing to pay for implementation, he ought to be willing to pay the developers too. If he wants an open source product so he can add his own extentions, wait a second, he's going to have to pay for development there too. If he doesn't want to do anything to the software, then why does he need the code?
Uh, maybe to avoid vendor lock-in?
It's insurance.
Controlling English writing in Quebec, from what I understand, is accomplished by Quebec's National Assembly (provincial legislative assembly) invoking the notwithstanding clause every 5 years.
But seriously, I don't see that yum is inferior to apt.
Press Ctrl-C sometime.
And I would hazard a guess that, the whole reason Debian is so stable... is because of it's policies and processes. I would say that APT/DEB has very little to do with this stability.
I'd say it's the reverse: Debian Policy is the reason why APT/DEB are so good. Because Debian Policy could say "thou shalt place a file in /usr/share/menu", stuff like the "menu" package (which centralized the management of various window managers' application menus) actually happened before any other distro had it.
Do the same restrictive regulations apply in less densely-populated areas? Somebody living on an acreage in an unforested area really shouldn't have to deal with the same regulations as someone in, say, a 12-storey apartment building in the middle of a metropolis.
This is just a bunch of people who are religious zealots about the Creative Commons licensing suite who ... hate the viral nature of the Free Software Foundation license suite.
Well this won't help those zealots, since the new GFDL only allows migrating to the ShareAlike license, which also has a "viral nature".
Like I said, there are problems with the GFDL, but throwing it away just because a small group (and it is a very small group of individuals pushing to make this change) doesn't like the Free Software Foundation. Unfortunately this small group making this change also have some prominent positions within the WMF.
So you think RMS would change the GFDL just because prominent people at the Wikimedia Foundation (allegely) don't like the FSF? I don't buy it.
You cannot "taste test" fruit in a grocery store, you cannot download copyrighted documents with impunity, and you cannot give these things away to the general public because it is stealing and it is wrong.
Don't be a jackass. If you aren't going to bother addressing the obvious differences between copyright infringement and petty theft, then you don't have any business expressing your ignorant opinion on the subject.
Singlecasting of 30 million copies of CSI via internet is not logical.
Yes, it is, and it's a worthless straw-man argument.
By "it is", I mean "it is illogical". Trying to single-cast digital TV would be insane.
You might as well argue that the DNS can't work because the root servers can't handle every DNS query in the world. The reality is that they don't have to.
Take a typical show like CSI which is watched by 30 million viewers each week. It's broadcast by 200 stations at approximately 15 megabit/s. If broadcast television was discontinued and CSI provided by ABC.com, it would need a bandwidth of 500,000 gigabit/s bandwidth.
Uh, 15 Mbps * 200 = 3000 Mbps, and that's assuming you wouldn't use a better codec, which you would. If you were going to deploy Internet access nationwide as a replacement for over-the-air TV, you would make sure multicast worked.
Singlecasting of 30 million copies of CSI via internet is not logical.
Yes, it is, and it's a worthless straw-man argument.
Here's an idea from left field: Why not treat Internet access as an essential service, and work toward getting it to everyone? You can broadcast TV over the Internet, and Wikipedia access alone would improve the average American's education by an order of magnitude.
0x07? You support censorship?
Carmack isn't a particularly brilliant programmer. He's just pragmatic and dedicated to putting in the effort to create something.
The code in Q1-Q3 is your standard stuff, nothing special in there and in fact it's too limited in many areas that could use a little more flexibility (ie. it was a bad design).
The code in Q1-Q3 was "standard stuff" because "standard" in that era was defined as "whatever John Carmack does".
John Carmack had a strength most other programmers lack: reading academic papers and applying them. Hell, in that era, reading RFCs made one an above-average programmer.
If you think John Carmack wasn't brilliant, then you haven't examined the work of his contemporaries. Comparing the code he wrote 10 years ago against code others write today isn't fair.