OK have I missed something - cause one thing nobody has commented on is this para from TFA:
One of the agents of the RIAA, such as Media Sentry, downloads a file from an unsuspecting file-sharer. A screenshot is made of the individual's shared directory, or several files are downloaded to ensure a viable case. The individual's IP address is then obtained. The RIAA then subpoenas the file-sharer's ISP requesting the personal information associated with that IP address when the alleged upload occurred.
OK so, basically the RIAA sets up a file for you to download just so they can catch you (Entrapment), which then installs itself on your computer against your wishes or knowledge (distribution of Spyware), takes screenshots of your computer and sends them to RIAA (clear invasion of privacy).
And this is to enforce the law? Aren't there laws against doing such things?
Furthermore, what shitty evidence is an IP address. IP addresses do not equal individuals, for several reasons.
Um OK, if movie execs (and music execs, as the RIAA are in this case) are going to stoop as low as they have been for the past decade or so, they deserve everything that's coming to them.
Let them learn a musical instrument and create some art to earn a living instead of just exploiting those with talent.
The way I see it, there is no "lock-in monopoly" of the MP3 format. Any MP3 player company could easily switch their entire player to play the OGG format, and have their software convert music to OGG instead of MP3 (of course, you've still got the problem of the software having to read MP3, but at least that's a lot less problematic than having the actual device do it).
I foresee one huge problem though for anyone switching their player to OGG: This industry has the fatal problem of being called "MP3 players", not "digital music players" or anything of the sort. Therefore any "OGG players" would likely be seen as a "cheap knockoff". Why did we create this industry and call it after a patented format?
(In other words, we're not technologically constrained by the format, but market-wise constrained).
The only player who is really in the clear of this is Apple - it's not an "MP3 player", it's an "iPod". They could jump out of MP3.
It doesn't sound like this would be particularly exploitable because you'd need them to type the letters in the right order
This is true, but say you put it on a website where people type a lot into a text box - such as a forum, or web mail, or... AAAH! SLASHDOT!
It turns out that some percentage of users will probably type the correct string just by chance (hence the reason why the sample page used that silly expression about cheese).
Another way they could trick you into typing the letters in the correct order is to make a page saying they found a vulnerability in your browser, and to type the following into the box to "see this exploit in action", secretly stealing your file;)
Getting someone to type something might prove easier though now due to the prevalence of Capchas.
True, although I'd be a little suspicious if I saw a page which said:
"In order to prove you're human, please enter the following text into a box: C:\BOOT.INI"
In all seriousness, note that if you comply with what the author had in mind, you'd actually be typing in arbitrary upper/lowercasing. This is less of a security risk on POSIX systems than Windows, because filenames are case sensitive.
The irony of this comment is that the binary logarithm of 10k is still much smaller than 32. Radix sort is efficient for other reasons.
Oh of course! You have to be larger than 2^32 ~~ 4,000,000,000! (Duh). So unless you're sorting google's database of web pages, it's not going to give better performance than nlogn.
(Wow, now that you mention it, nlogn is pretty close to n really, isn't it).
First it doesn't require the whole dataset to exist at a time. This is nice for things like bzip2 where the data itself is partly generated.
That doesn't seem right to me (since you need to do k iterations over the whole dataset, you must require it to exist in memory concurrently). But I'll have to think about it.
Note that he lists all of his "skills" as languages and programs. Languages != skills.
And was it intentional that his hobbies is an infinite list of the string "Computers"?
Hobbies : Loop ......................AddHobbie ("Computers"); ................Do Until 1=2
(Is that VB, with a semicolon?)
So.. anyway, is it just me or has everything been taken off the site now? I thought it was a browser or slashdotting problem, but now there's a link to "Radix sort" on Wikipedia. Do you think they read all the slashdot posts, realised it was radix sort, then deleted everything and replaced it with a wikilink? Well I guess our job is done...
Problem: What if one significant reason a Linux machine is better than a Windows machine is the lack of malware? Thus, a fresh (non-OEM) Windows install will get loaded with malware infinitely faster than a fresh Linux install, as the Linux install will never get hit with malware.
OK this is true. It does decrease one big advantage (well, it wouldn't be an issue if the Linux system was set up correctly, but then if you trust that to people who are going to preinstall malware, chances are it won't be). However a) it's easy enough to wipe+reinstall Linux, at least for power users and above, and b) it shouldn't turn Windows people off Linux, because it's not worse off than Windows.
Also I wasn't aware we were talking about malware - these "crapplets", I thought, were just ads. I don't know, I've never bought a computer from a major retailer like this.
It's hardly a free/open platform if it comes with BonziBuddy Unix Edition.
Lol... my brother installed BonziBuddy on his laptop years and years ago. I was quite cautious that he was an evil little ape so I never let it touch my machines. Turns out I was right. But he was just so cute!
No, I don't really see a problem here. If Windows machines are cheaper than Linux machines because they're crapplet-subsidised, I say bring on the crapplets.
After all, a Linux machine with crapplets is better than a Windows machine with crapplets by the same amount that a Linux machine is better than a Windows machine.
I haven't seen the article - it seems to be vanished or something. Anyway since everyone's saying it's just RadSort I'll go on that.
The difference between Radix Sort and one of those other sorts is in the complexity. Most sort algorithm's complexity is based on the number of elements solely. That is, Qsort and MergeSort are O(nlogn), Selection sort is O(n^2), etc. (Forgive me if I got any of these wrong, it's been awhile).
The point is that all of these sort algorithms have a maximum complexity of nlogn. Nobody has ever developed (nor is it thought possible) an algorithm which sorts faster than nlogn.
Enter Radix Sort: Which sorts with complexity O(n*m). Where m is the number of "bits" or the number of "radixes" you're going to select. So in general, this is nowhere near as good as Qsort or Mergesort, since m could be anything. A good example is for sorting strings - you can use radix sort (or I believe it's called "bucket sort" when you're not using bits), but the complexity will be the number of strings times the number of characters.
However, Radix sort shines when there is a fixed number of bits. Obviously that's true of integers. If you're sorting 32-bit integers, your complexity is suddenly O(n*32), which of course reduces to O(n). So, wow, we have a linear-time sorting algorithm! This is a good thing, but with three important caveats:
1. The constant multiplier factor is much larger than for most algorithms (ie. it's actually going to take 32 times longer than you might think just looking at O(n)). However, the reason why we eliminate constant multipliers from O notation is that they're meaningless with large data sets. So if you're sorting 10000 integers, then n*32 is far better than nlogn. In other words radix sort keeps getting better as the data size increases, but the bit size remains the same. 2. This doesn't work, or doesn't work well, in general. It only works on specific data types, such as integers. Other data types, such as strings, will perform poorly. Furthermore, the entire implementation is data-type dependent (that is, it has to be written totally differently for sorting ints, floats, strings, etc). This is as opposed to almost any other sorting algorithm where you just provide a comparison function for two elements of a given type, and it works. 3. I just thought of this but it's a biggie: This whole algorithm is very low-level. That is, it's language-depentend, platform-dependent and compiler-dependent as to how things are stored at the bit level. You need to be aware of if numbers are in two's complement. You need to know the exact floating-point representation if you're sorting FP numbers (and not all FP representations can even be bit sorted). And your language needs to support bit accesses - this is not possible in high level languages like Python or Haskell or maybe even Java. So if you write radix sort, even in C, it won't be properly portable code.
I don't pirate TV, I just wait and watch it when we catch up (I'm in Aus btw). But it's getting to the point that it's hard to avoid all these spoilers.
Firstly, you have the Internet. Obviously most fiction-related websites have a "grace period" where they don't discuss spoilers of recent episodes. But when you're 5 months to a year behind, you simply can't visit any websites related to these shows (including Wikipedia articles) or you'll see months and months of spoilers.
Secondly, you have an increasing number of friends who give up, cross the barrier, and simply start downloading episodes at the same time as the US. And when your friends reach a critical mass (as they have done with shows such as Heroes and Lost), it's "did you see the latest episode of..." referring to the latest episode from the US, and we who choose to watch the televised version have to go hide in a corner for 5 months while we wait to catch up.
All in all it adds up to a slippery slope of more and more viewers turning to BitTorrent because the TV networks can't keep up - whoever's fault that is. If you're going to take the stance of "It isn't the TV networks' fault because it's really expensive to get shows early," then I say the US TV producers need to realise that they'll lose viewership (or contracts) with foreign networks unless they start making it feasible to get shows at the same time, or near enough.
Network Ten in Australia is making headway here, showing Jericho at the same time as the US, the final episode of the OC within one day of the US, and Smallville (which began 2 seasons behind the US when Ch10 bought it) rocketing through three seasons without a break to catch up - currently half a season behind.
It's not that there's too much "stuff to do" per se. It's that there's a difference between cool stuff (which for me is pretty much the exploration, being able to see places and meet characters - but also just includes fun imaginitive quests), and boring grinding (and I consider endless "kill x monsters" quests to be part of the grinding).
WoW is the first game I've found where the quests themselves can be considered grinding, since they are just so tedious and repetitive.
So to be clear: I'm not quitting because there's too much stuff, I'm quitting because it's too hard to wade through the crap to get to the good stuff.
Er, great solution. Until ebooks become feasible and portable enough that it sees mass adoption, books begin to be distributed more and more in this way until it reaches critical mass, and sure enough, the books are protected by insane DRM (to stop people from "stealing passages" of the work).
Due to the demand to read ebooks on general purpose computers (and the incredible ease of copying text - far moreso than audio or video), general purpose computers will be locked down more and more, copy and paste will become a "protected channel" and ebook programs will be given high-security status.
And the cycle will continue... the Apple eBooks will be incompatible with the Microsoft eBooks, the open source eBooks will be an indie niche, and you'll have to buy and carry around multiple eBook readers to handle the competing standards. We're talking about plain text becoming a format war.
And I am not kidding or exaggerating. This little-known technology is already present in Office and Internet Explorer (including "protected HTML") since 2003.
If you want to avoid DRM, I recommend you get behind the anti-DRM movement now, not stick to formats yet-to-have-been conquered.
- WinDVD did its best to protect its device key - It's impossible to protect a device key in a program that people can reverse-engineer [true] - We'd better not allow any software to read AACS-protected content
That's a good outcome. With hardware-only DRM, well we lose rights to our movies. But at least we get to keep our computers.
The way it went with Vista, however, is that we lost the rights to our movies, and now we've lost the rights to have access to our computers too. It's all locked up, to the detriment of all future computer users.
So at the very least, DRM can get the hell off my computer.
You raise some good points but two points I'd like to make are:
Patents are not equivalent to copied code. Patents are for methods. It's still possible to detect patent use in closed software, because the methods are still known (though it's harder to figure out how it works).
Open source is not a wiki, it's not like anybody can contribute code. Any open source project manager will carefully check the code. (Though of course that's no guarantee it doesn't violate patents, again).
So Novell either doesn't know or won't disclose what it paid 40 million dollars for?
Nobody outside of Microsoft knows which patents, if any, MS thinks Linux has violated. They wouldn't have told Novell - Novell is Microsoft's enemy and pawn.
Microsoft got in on this deal for the ability to sling FUD, saying "Novell has basically admitted that Linux violates patents or they wouldn't have gone in on this deal."
Novell got in on this deal for the ability to say, "See, if you use Linux you may get sued by MS, so buy our Linux, the only patent-safe way to get it."
Both will benefit from the FUD. Nobody will ever admit anything or sue anyone. Novell will eventually bite the bullet when this all cools off. But in the meantime, they don't care which patents are in question - it's all about the FUD.
I think the point is that is MS actually does sue Linux users, there will be a huge retaliation. (And there always has been this threat). So they won't.
Before you say "won't that nullify the point of this site, since it won't prove anything about whether MS has patents or not", I'll point out that if "we can't sue or you'll retaliate by suing us" is MS's response (or excuse for not doing anything at all), then it proves that they can't do anything about it and we can get on without all this FUD.
It would also serve to show that they are equally, if not more, in violation of the open source community's patents than we are of them.
I was also interested to read that the original study (the "283 patents" which Ballmer refers to) was actually a finding that Linux contained 283 potential patents in total. It estimated that 1/3 of the patents were owned by the OS community companies themselves (eg. IBM). And Microsoft owned just 10% of them - which if my calculations are correct is just over 28 patents, not over 200 as Ballmer claims.
The patent FUD really has nothing to do with the general public. Even if MS did sue, it wouldn't be individuals, it'd be big business who are using Linux. So it's for the business customers that this needs to be cleared up.
I think it's a very dangerous thing to do (and if it is a hoax, suicidal), because when you use software in general, you typically place a great degree of trust that each individual program is not going to be malicious. It only takes one malicious program to mess up everything (particularly in Windows). I simply couldn't have that level of user/developer trust with any program which had code in it to "delete my home directory", even if it were guaranteed it wouldn't be used if I cooperated (which there is no guarantee of).
In fact I'd be wary in general of any program with the word "Eater" in the title.
Slashdot Mirror
And this is to enforce the law? Aren't there laws against doing such things?
Furthermore, what shitty evidence is an IP address. IP addresses do not equal individuals, for several reasons.
Um OK, if movie execs (and music execs, as the RIAA are in this case) are going to stoop as low as they have been for the past decade or so, they deserve everything that's coming to them.
Let them learn a musical instrument and create some art to earn a living instead of just exploiting those with talent.
The way I see it, there is no "lock-in monopoly" of the MP3 format. Any MP3 player company could easily switch their entire player to play the OGG format, and have their software convert music to OGG instead of MP3 (of course, you've still got the problem of the software having to read MP3, but at least that's a lot less problematic than having the actual device do it).
I foresee one huge problem though for anyone switching their player to OGG: This industry has the fatal problem of being called "MP3 players", not "digital music players" or anything of the sort. Therefore any "OGG players" would likely be seen as a "cheap knockoff". Why did we create this industry and call it after a patented format?
(In other words, we're not technologically constrained by the format, but market-wise constrained).
The only player who is really in the clear of this is Apple - it's not an "MP3 player", it's an "iPod". They could jump out of MP3.
It turns out that some percentage of users will probably type the correct string just by chance (hence the reason why the sample page used that silly expression about cheese).
Another way they could trick you into typing the letters in the correct order is to make a page saying they found a vulnerability in your browser, and to type the following into the box to "see this exploit in action", secretly stealing your file
"In order to prove you're human, please enter the following text into a box:
C:\BOOT.INI"
In all seriousness, note that if you comply with what the author had in mind, you'd actually be typing in arbitrary upper/lowercasing. This is less of a security risk on POSIX systems than Windows, because filenames are case sensitive.
You can't. DRM relies solely on "security through obscurity," long considered a "false" security by security experts.
Surprise: No DRM has ever survived being cracked.
(Wow, now that you mention it, nlogn is pretty close to n really, isn't it).That doesn't seem right to me (since you need to do k iterations over the whole dataset, you must require it to exist in memory concurrently). But I'll have to think about it.
Uh there have been lots of systems available to the general public for the twenty years Windows has been around. People didn't get over it.
And was it intentional that his hobbies is an infinite list of the string "Computers"?(Is that VB, with a semicolon?)
So.. anyway, is it just me or has everything been taken off the site now? I thought it was a browser or slashdotting problem, but now there's a link to "Radix sort" on Wikipedia. Do you think they read all the slashdot posts, realised it was radix sort, then deleted everything and replaced it with a wikilink? Well I guess our job is done...
Also I wasn't aware we were talking about malware - these "crapplets", I thought, were just ads. I don't know, I've never bought a computer from a major retailer like this.Lol... my brother installed BonziBuddy on his laptop years and years ago. I was quite cautious that he was an evil little ape so I never let it touch my machines. Turns out I was right. But he was just so cute!
No, I don't really see a problem here. If Windows machines are cheaper than Linux machines because they're crapplet-subsidised, I say bring on the crapplets.
After all, a Linux machine with crapplets is better than a Windows machine with crapplets by the same amount that a Linux machine is better than a Windows machine.
(In other words, ((L+c) - (W+c) == L - W))
(L >> W)
I haven't seen the article - it seems to be vanished or something. Anyway since everyone's saying it's just RadSort I'll go on that.
The difference between Radix Sort and one of those other sorts is in the complexity. Most sort algorithm's complexity is based on the number of elements solely. That is, Qsort and MergeSort are O(nlogn), Selection sort is O(n^2), etc. (Forgive me if I got any of these wrong, it's been awhile).
The point is that all of these sort algorithms have a maximum complexity of nlogn. Nobody has ever developed (nor is it thought possible) an algorithm which sorts faster than nlogn.
Enter Radix Sort: Which sorts with complexity O(n*m). Where m is the number of "bits" or the number of "radixes" you're going to select. So in general, this is nowhere near as good as Qsort or Mergesort, since m could be anything. A good example is for sorting strings - you can use radix sort (or I believe it's called "bucket sort" when you're not using bits), but the complexity will be the number of strings times the number of characters.
However, Radix sort shines when there is a fixed number of bits. Obviously that's true of integers. If you're sorting 32-bit integers, your complexity is suddenly O(n*32), which of course reduces to O(n). So, wow, we have a linear-time sorting algorithm! This is a good thing, but with three important caveats:
1. The constant multiplier factor is much larger than for most algorithms (ie. it's actually going to take 32 times longer than you might think just looking at O(n)). However, the reason why we eliminate constant multipliers from O notation is that they're meaningless with large data sets. So if you're sorting 10000 integers, then n*32 is far better than nlogn. In other words radix sort keeps getting better as the data size increases, but the bit size remains the same.
2. This doesn't work, or doesn't work well, in general. It only works on specific data types, such as integers. Other data types, such as strings, will perform poorly. Furthermore, the entire implementation is data-type dependent (that is, it has to be written totally differently for sorting ints, floats, strings, etc). This is as opposed to almost any other sorting algorithm where you just provide a comparison function for two elements of a given type, and it works.
3. I just thought of this but it's a biggie: This whole algorithm is very low-level. That is, it's language-depentend, platform-dependent and compiler-dependent as to how things are stored at the bit level. You need to be aware of if numbers are in two's complement. You need to know the exact floating-point representation if you're sorting FP numbers (and not all FP representations can even be bit sorted). And your language needs to support bit accesses - this is not possible in high level languages like Python or Haskell or maybe even Java. So if you write radix sort, even in C, it won't be properly portable code.
Wow that's genius!
:)
I was a big fan of dangermouse's esoteric languages some years ago:
http://www.dangermouse.net/esoteric/
I didn't realise he'd been developing these esoteric algorithms
I don't pirate TV, I just wait and watch it when we catch up (I'm in Aus btw). But it's getting to the point that it's hard to avoid all these spoilers.
..." referring to the latest episode from the US, and we who choose to watch the televised version have to go hide in a corner for 5 months while we wait to catch up.
Firstly, you have the Internet. Obviously most fiction-related websites have a "grace period" where they don't discuss spoilers of recent episodes. But when you're 5 months to a year behind, you simply can't visit any websites related to these shows (including Wikipedia articles) or you'll see months and months of spoilers.
Secondly, you have an increasing number of friends who give up, cross the barrier, and simply start downloading episodes at the same time as the US. And when your friends reach a critical mass (as they have done with shows such as Heroes and Lost), it's "did you see the latest episode of
All in all it adds up to a slippery slope of more and more viewers turning to BitTorrent because the TV networks can't keep up - whoever's fault that is. If you're going to take the stance of "It isn't the TV networks' fault because it's really expensive to get shows early," then I say the US TV producers need to realise that they'll lose viewership (or contracts) with foreign networks unless they start making it feasible to get shows at the same time, or near enough.
Network Ten in Australia is making headway here, showing Jericho at the same time as the US, the final episode of the OC within one day of the US, and Smallville (which began 2 seasons behind the US when Ch10 bought it) rocketing through three seasons without a break to catch up - currently half a season behind.
It's not that there's too much "stuff to do" per se. It's that there's a difference between cool stuff (which for me is pretty much the exploration, being able to see places and meet characters - but also just includes fun imaginitive quests), and boring grinding (and I consider endless "kill x monsters" quests to be part of the grinding).
WoW is the first game I've found where the quests themselves can be considered grinding, since they are just so tedious and repetitive.
So to be clear: I'm not quitting because there's too much stuff, I'm quitting because it's too hard to wade through the crap to get to the good stuff.
It was necessary.
Er, great solution. Until ebooks become feasible and portable enough that it sees mass adoption, books begin to be distributed more and more in this way until it reaches critical mass, and sure enough, the books are protected by insane DRM (to stop people from "stealing passages" of the work).
... the Apple eBooks will be incompatible with the Microsoft eBooks, the open source eBooks will be an indie niche, and you'll have to buy and carry around multiple eBook readers to handle the competing standards. We're talking about plain text becoming a format war.
Due to the demand to read ebooks on general purpose computers (and the incredible ease of copying text - far moreso than audio or video), general purpose computers will be locked down more and more, copy and paste will become a "protected channel" and ebook programs will be given high-security status.
And the cycle will continue
And I am not kidding or exaggerating. This little-known technology is already present in Office and Internet Explorer (including "protected HTML") since 2003.
If you want to avoid DRM, I recommend you get behind the anti-DRM movement now, not stick to formats yet-to-have-been conquered.
The way it went with Vista, however, is that we lost the rights to our movies, and now we've lost the rights to have access to our computers too. It's all locked up, to the detriment of all future computer users.
So at the very least, DRM can get the hell off my computer.
Microsoft got in on this deal for the ability to sling FUD, saying "Novell has basically admitted that Linux violates patents or they wouldn't have gone in on this deal."
Novell got in on this deal for the ability to say, "See, if you use Linux you may get sued by MS, so buy our Linux, the only patent-safe way to get it."
Both will benefit from the FUD. Nobody will ever admit anything or sue anyone. Novell will eventually bite the bullet when this all cools off. But in the meantime, they don't care which patents are in question - it's all about the FUD.
I think the point is that is MS actually does sue Linux users, there will be a huge retaliation. (And there always has been this threat). So they won't.
Before you say "won't that nullify the point of this site, since it won't prove anything about whether MS has patents or not", I'll point out that if "we can't sue or you'll retaliate by suing us" is MS's response (or excuse for not doing anything at all), then it proves that they can't do anything about it and we can get on without all this FUD.
It would also serve to show that they are equally, if not more, in violation of the open source community's patents than we are of them.
I was also interested to read that the original study (the "283 patents" which Ballmer refers to) was actually a finding that Linux contained 283 potential patents in total. It estimated that 1/3 of the patents were owned by the OS community companies themselves (eg. IBM). And Microsoft owned just 10% of them - which if my calculations are correct is just over 28 patents, not over 200 as Ballmer claims.
Ooh :/ that was a mistake.
Best. link. ever.
I think it's a very dangerous thing to do (and if it is a hoax, suicidal), because when you use software in general, you typically place a great degree of trust that each individual program is not going to be malicious. It only takes one malicious program to mess up everything (particularly in Windows). I simply couldn't have that level of user/developer trust with any program which had code in it to "delete my home directory", even if it were guaranteed it wouldn't be used if I cooperated (which there is no guarantee of).
In fact I'd be wary in general of any program with the word "Eater" in the title.