"that won't be enough to stop the ensuing shit storm. Last time something like this happened (iTunes 4.0 on the mac) it was discovered that you could enable music sharing over the entire internet. The RIAA jumped down Apple's throat."
The problem was the sharing across the internet, which was addressed by iTunes sharing being limited (in the first upgrade to iTunes) to LAN's, after which the record companies were fine with it -- people SHOULD be able to move music freely around their home, but not copy it to strangers across the internet. Since this program is functionally identical to programs that have been out for the Mac for many months, it doesn't introduce anything dramatically new, just evens up the PC users with the Mac users. Which is to say that music can be copied across a LAN but not the internet, and music sold by iTunes can be copied but not played without authorization. It's about the same as if someone turned on file sharing, and shared their Music folder to the LAN. That being said, I'd expect Apple to do what it can to discourage the availability of MyTunes, just as they did before with the comparable Mac app's.
It's not like Apple could prevent people from using file sharing to simply share their Music folders onto the LAN. It's not as slick a UI as iTunes, but certainly adequate.
iTunes doesn't really stream the music in the sense of a bit-rate limited version trickle delivered. It's more like QuickTime's auto-start download, where the full quality file is downloaded and played, only it's never written to disk. But if you capture and save the file, it's identical to the source file, in the same format, with the same DRM, etc.
This is different from operating a streaming server, where any sound to be broadcast is squished into a single continuous audio stream, at the desired bitrate and stream format. So, for example, I can run Nicecast (great app!) and it'll take whatever sounds play on my Mac, convert it to an icecast stream at 56K bps (or whatever I tell it), and stream it out. So if I play Protected AAC's, or WAV's, or movies, etc., it all ends up in one long stream, no files, no metadata.
Actually, I thought that "Windows Program Enables MP3 Downloading From iTunes" was unusually precise. It's a program for Windows that allows people to download files from iTunes shares that they can stream. Since iTunes only shares within a LAN, you can't see shares across the internet, so you can't download from them. It doesn't re-encode the files, so if they're "protected" they're still protected. So, all around, the headline was pretty accurate. Sorry you mis-read it, but you can't beat Slashdot up over that one. If it makes you feel better, there are plenty of candidates for that honor...
"Most people could care less about IE or WMP9 being bundled with Windows, and if they don't like using them....Then they stop using those specific programs. They don't dump the whole OS."
This ignores the catch 22 that it creates. If IE is bundled on 90% of computers, then web sites and application vendors all assume IE, which drives everyone else out of the browser business. Not because people can't choose alternative browsers, but because things won't work right unless you use what everyone else uses.
"if Al Gore had been president any of the following may have happened:"
Troll, you left out:
- The US wouldn't have alienated the rest of the planet by launching an unprovoked war on Iraq. - The US would have signed the Kyoto Protocol, and not alientated the rest of the planet on that front as well. - Hundreds of americans, and thousands of Iraquis, would still be alive. - The US economy would be about $130B better off (only counting money wasted on the war in Iraq -- more if you could secondary effects, like the business cost of companies refusing to do business with us, etc.). - Enron, etc., would have been better managed, because the white house wouldn't have been so busy covering things up instead of addressing them. - The US energy policy would be balanced, rather than written by the oil industry in order to maximize dependence on them, and thus on middle eastern oil supplies.
"Sure, who wants to sabotage PHPMyAdmin? But the voting software used in the US? That's a different story..."
I would hope that since voting is such a sensitive issue, it would be managed as tightly as the Linux and BSD kernel projects, where only well known, trusted people can check code into the project, and all code must be thoroughly reviewed by a whole team of people before it can be considered for checkin.
"open sourcing means it's easier to introduce a back door."
Open source doesn't mean that everyone can freely contribute code. Look at the Linux or BSD projects, for example. Anyone can read the code, and can make their own branch, but only a fairly rigorous process of code review and debate by well known and trusted people can get code into the project.
"Maybe the best thing is to have the federal government pay for the development of a paperless voting system. That brings up a new set of concerns though..."
What concerns? If the government issues a grant to, say, a university to develop software that achieves the spec (securely collect votes, auditable, etc.), and the result is open sourced, I'm not too worried about whether the government might perfer a particular election result.. they'd have to affect the NSF (or whoever) to affect the university to affect the developers, and then nobody inspecting the source would notice...
"Even then it's not enough. As my example of a word macro. Even if you know it's an allegedly passive type of document (like a word.doc) that STILL doesn't mean it's safe."
Good point! All that 'integration' in Windows has caused the barrier between data and code to blur. This wouldn't be so bad, except that Windows has no security model, so you can only open a "document" that does stuff (e.g. open a Word document, it runs a macro), there's no real limit on what it can do (e.g. open a word document that runs a macro that emails the document containing the macro to everyone in my address book, then adds the macro to every word document on my computer). So while I'm fan of the potential of API's to allow intelligent agents to do cool things for me, it has to be done in a manner that gives me control over what's going on.
"Who is going to install the full voting suite (voting software is much more than a voting terminal) and then hold mock elections in their home?"
Anyone with an interest in making sure that the voting process is secure -- universities, public interest groups, etc. Security analysts would scour the system in order to compete to issue their reports for their clients. Academics would scour the system for their research/publication. The same organizations that are so critical of the closed source voting systems organizations would audit the system. Heck -- look at the level of inspection that the leaked voting system code and documentation have received so far--no closed source project receives that kinda of effort outside of NASA. And, of course, any organization that holds formal votes (schools, churches, unions, etc.) could also use the system, which would both be an additional benefit (free system for them to use) and more users to flush out problems.
Of course, this all misses the most fundamental point -- voting is a critical function of the government, and must be performed in a transparent and auditable manner in order for people to trust it. Given the long and varied history of people committing voter fraud, there's no reason to think that people would _not_ attempt to manipulate the system. And with a system that can't be audited, doesn't support manual recounts, etc., there's no reason for anyone to trust it.
"The fault was not in the development model but in the failure of the project leadership."
The fault is in election boards allowing themselves to be hoodwinked by a bunch of sleazy opportunists chasing the $4B that the government has committed to updating the voting infrastructure. The closed source development model simply allows those bozos to conceal their shoddy workmanship.
"How the product is marketed is IRRELEVENT. Its how its utilized."
Apple's claim is that the dual G5 is the fastest "personal computer" -- a market segment. If you're saying that you have the fastest computer in a market segment ("personal computers"), then how the product is marketed matters quite a bit, because that defines what's in the market segment. There are rather obviously high-end workstations that can outperform the dual G5, and they can fit on a desk, but that's not terribly relevant. That's like arguing that because there are astoundingly fast specialized racecars that a mass market sports car couldn't claim to be the "fastest car in its class".
BTW, people have done some benchmarks of dial Xeon's against the Dual G5, and overall the G5 was faster. Of course, you could find some tests where the Xeon's were a little faster, but there were also tests where the G5's were a lot faster... and (though it doesn't affect performance) the fastest Xeon's cost $1-2K more than the G5. That would be (wait for it) because the dual Xeon's are sold as workstations, while the dual G5 is sold as a personal computer, thus larger volumes and lower margins.;-)
"I've speced out machines that run circles around the G5 for $2100... this is with a freakin 3Ghz P4".
Cool, for $2,100 you've outperformed the $1,600 G5. I think that perhaps you could try to configure a PC to outperform the model that Apple claimed was the fastest personal computer; when people tried to pull together dual Xeon's, they were around $4-5K (depending on vendor) so $1-2K more than Apple's dual G5. You might be able to get the fastest PC price down to Apple's price by "white boxing" it...
but even if the PC is more expensive than the dual G5, it's not clear that it could outperform it. The independent benchmarks that I've seen had either machine perform "best" on some tests, but overall the dual G5 outpeformed the fastest PC's, because when the PC was faster it was 10-20% faster, and when the Mac was faster, it was up to 50% faster. I suspect it had to do with which code could take advantage of the AltiVec instruction set, which is much more powerful/flexible than Intel's comparable SIMD instruction set...
Exactly. The design error is that Outlook doesn't contain its own viewer code to render data (i.e. a safe operation) it simply tells the OS to "open" the file, so the user uses the same operation to view a photo as to execute an application. Adding to the danger, Outlook by default hides file extensions, so users have to know how to recognize which icons are "safe" to open and which are not. For example, it's "safe" to open a text file, and "dangerous" to open a screen saver file, but they act and look quite similar in Outlook.
"I would argue it has, given their recent decision to prioritize security and stability. 2000 and XP and miles ahead of 95 as well."
It may be true that 2000 and XP are "miles ahead of 95" but that's not due to their recent emphasis on security and stability -- both operating systems were written before MS made that conversion. The NT-based versions of Windows are more stable and secure than the 9x versions because it's based on VMS, a fundamentally more secure OS model than the DOS-based 9x.
"It will accelerate the arms race between marketroids and Internet users"
This is certainly true. That being said, in the mass market right now the "marketroids" are the only ones fighting the war right now -- IE users are all on the receiving end of so much advertising that the general internet is all but useless. Sure, the advertisers will find new ways to abuse internet users, but that was going to happen anyway.
One side effect, though, will be that those of us using web browsers that have blocked pop-ups forever may be vulnerable to the next bit of advertising. But I'm confident that our browsers can move faster than IE...
The kind of "pop up" that these tools block is a pretty specific thing. They don't block every JavaScript that opens a window. They allow "pop ups" that are triggered by user actions, and block "pop ups" that are triggered by system events. So if you click on a "more info" button that pops up a window to tell you about a product while shopping, for example, you'll see the info. But if you browse a pr0n site (who would do that?!) that triggers a script whenever the page is closed that opens a new pop-up, that'll be blocked.
This sounds simple, but the impact is immense. With a browser that blocks pop-ups (i.e. anything by IE), the intenet is a pleasant place. With IE, the internet is a horrible maze of X10 camera ads and pr0n spam.
If IE finally catches up to the other browsers, and implements a good pop-up blocker that is on by default, everyone will benefit.
"Temperature recorders made before some point in the 20th century had a design defect that makes them accurate only to about 3 degrees, which is well outside the claimed variation."
Even if any individual measurement has a varience of 3 degrees, an analysis of many thousands of measurements will virtually eliminate that imprecision. Now assuming that the "global temperature" was measured in a reasonable number of locations (hundreds of cities?) fairly often (weekly? daily?) the number of samples would be large (10K's?).
Also, the study that started the whole global warming thing covered a much larger time period than when humans were measuring the temperature with thermometers -- they measured ice rings, etc., in order to estimate the temperature over many, many thousands of years. So even if the measurements for one recent decade were off, that wouldn't change a long term trend.
"The parent went a bit far by saying they could 'shut down' a company"
Well, Microsoft could (before the monopoly ruling) shut down any PC vendor by pulling their license to Windows. Of course, they don't need to -- more subtle threats, such as inflated pricing or witholding beta releases so that the vendor can't test for compatability, have so far been sufficient to cause every PC vendor to agree to anything MS has needed from them so far. I believe that one of the providions of the settlement is that MS has to sell Windows to any vendor off of the same pricing schedule, which at least limits their power to manipulate the market a little bit.
Of course, I agree with the rest of your post completely...
"I don't doubt that in the least, but considering they are probably being paid pretty well by the RIAA to do this survey and report the results, they don't really have an obligation to announce how inaccurate the data is if it reflects the "success" of the RIAA's litigious tactics."
NPD is in the business of providing accurate market survey data to its customers, who use that data to make critical business decisions. If it ever biased a survey in order to manufacture an inaccurate outcome, it would cost them far more in lost credibility (and thus clients) than they could every make performing that survey.
"Good answers? Usually I like correct answers more..."
OK, I'll elaborate on their answers to my questions, and why I found them satisfactory. For example, when I asked them about whether people tended to read the first questions on a survey more closely, and skim the later ones, they said that they knew that, and permute their question sequence across the sample set, so that the bias would average out. This also compensates for any sequencing artifact, where one question will affect the answer to the next question, because the questions will appear in the opposite sequence 1/2 the time. They've also mentioned compensating for people's desire to please surveyors by saying "yes" more often than "no" to questions, by phrasing test questions both positively and negatively (i.e. "yes" to one means "no" to the other) so that they can model the difference compensate across other questions. And they make sure that their panel mirrors the general public in terms of age, gender, income and geographical distribution. Heck, I saw one survey where they made sure that the panelists had a representative distribution of computer configurations and modem speeds so that they didn't bias the sample by having too many panelists on broadband and fast PC's...
So I can't say that I've talked with them about this study, but I'm pretty sure that they thought of obvious factors like "people deny doing illegal things".
"Uh did you not read the part that said attractive and reasonable? iTunes is the closest so far but Apple, not the RIAA, is the driving force there."
If you're waiting for RIAA to run a music store, you'll be waiting forever. RIAA doesn't sell music, and never will -- they're a lobbying group representing the record labels, not a retailer.
If you're saying that iTMS isn't "attractive and reasonable" -- what are you suggesting should be changed?
"My current job is supporting a website that uses pop-ups to display menus, etc. I didn't design it. I have no say in how to improve it. I just have to sit on the phone and tell people to disable their pop-up blockers."
You should do some quick cost estimation encourage someone in a decision making capacity at the company to consider whether the costs to the company in customer support and in lost customers possibly outweighs whatever advantage there is to the current design.
Aside from that, using pop-ups for navigation seems like an amazingly bad idea. But saying that probably would alienate the designers without changing anyone's mind. "costing the company money" is a much better business argument than "sucks".
"The biggest security problem isn't vulnerability, it's the creativity of the people trying to be obnoxious"
While it's certainly true that if there were nobody writing viruses or cracking systems we'd all be better off, we appear to be in a world where such things happen regularly. Given that, it makes a difference whether an OS or application designer makes decisions that promote security (e.g. BSD) or insecurity (e.g. Microsoft). I'm not saying that MS has a goal of making things insecure (they're not crazy), but that they routinely decide that transparent integration between applications, or between applications and the OS, is more important than security. This is how viruses can control Outlook to automaticaly spam everyone you know, or why IE and Outlook so often are the delivery vectors for trojan horses, etc. If there were a clear barrier between data and applications, as there is in every other OS and email client, then we'd have far fewer security problems. And beyond that, Microsoft cares more about adding lots of new features, and shipping rapidly, than in security. That may be a reasonable business strategy (they're certainly making money, and have avoided any liability for the flaws in their software) but it doesn't demonstate a commitment to security.
Let's compare Exchange with Lotus Notes. They have about the same market share (in corporate email systems), but Outlook is riddled with frequent security holes, while Notes hasn't had any worth mentioning for the past few years. Since they have the same market share, that isn't the difference. And while many crackers may hate Microsoft, it's hard to argue that they love IBM, so motivation isn't the answer. Perthaps design decisions have implications?
Or let's compare Apache and IIS. Apache has a much larger market share than IIS (about 2:1), but IIS is the target of more successful intrusions. Plenty of large, rich companies run either platform, so that isn't the difference. Perhaps the Apache developers put a higher priority on security?
Slashdot Mirror
"that won't be enough to stop the ensuing shit storm. Last time something like this happened (iTunes 4.0 on the mac) it was discovered that you could enable music sharing over the entire internet. The RIAA jumped down Apple's throat."
The problem was the sharing across the internet, which was addressed by iTunes sharing being limited (in the first upgrade to iTunes) to LAN's, after which the record companies were fine with it -- people SHOULD be able to move music freely around their home, but not copy it to strangers across the internet. Since this program is functionally identical to programs that have been out for the Mac for many months, it doesn't introduce anything dramatically new, just evens up the PC users with the Mac users. Which is to say that music can be copied across a LAN but not the internet, and music sold by iTunes can be copied but not played without authorization. It's about the same as if someone turned on file sharing, and shared their Music folder to the LAN. That being said, I'd expect Apple to do what it can to discourage the availability of MyTunes, just as they did before with the comparable Mac app's.
It's not like Apple could prevent people from using file sharing to simply share their Music folders onto the LAN. It's not as slick a UI as iTunes, but certainly adequate.
iTunes doesn't really stream the music in the sense of a bit-rate limited version trickle delivered. It's more like QuickTime's auto-start download, where the full quality file is downloaded and played, only it's never written to disk. But if you capture and save the file, it's identical to the source file, in the same format, with the same DRM, etc.
This is different from operating a streaming server, where any sound to be broadcast is squished into a single continuous audio stream, at the desired bitrate and stream format. So, for example, I can run Nicecast (great app!) and it'll take whatever sounds play on my Mac, convert it to an icecast stream at 56K bps (or whatever I tell it), and stream it out. So if I play Protected AAC's, or WAV's, or movies, etc., it all ends up in one long stream, no files, no metadata.
Actually, I thought that "Windows Program Enables MP3 Downloading From iTunes" was unusually precise. It's a program for Windows that allows people to download files from iTunes shares that they can stream. Since iTunes only shares within a LAN, you can't see shares across the internet, so you can't download from them. It doesn't re-encode the files, so if they're "protected" they're still protected. So, all around, the headline was pretty accurate. Sorry you mis-read it, but you can't beat Slashdot up over that one. If it makes you feel better, there are plenty of candidates for that honor...
"Most people could care less about IE or WMP9 being bundled with Windows, and if they don't like using them....Then they stop using those specific programs. They don't dump the whole OS."
This ignores the catch 22 that it creates. If IE is bundled on 90% of computers, then web sites and application vendors all assume IE, which drives everyone else out of the browser business. Not because people can't choose alternative browsers, but because things won't work right unless you use what everyone else uses.
"if Al Gore had been president any of the following may have happened:"
Troll, you left out:
- The US wouldn't have alienated the rest of the planet by launching an unprovoked war on Iraq.
- The US would have signed the Kyoto Protocol, and not alientated the rest of the planet on that front as well.
- Hundreds of americans, and thousands of Iraquis, would still be alive.
- The US economy would be about $130B better off (only counting money wasted on the war in Iraq -- more if you could secondary effects, like the business cost of companies refusing to do business with us, etc.).
- Enron, etc., would have been better managed, because the white house wouldn't have been so busy covering things up instead of addressing them.
- The US energy policy would be balanced, rather than written by the oil industry in order to maximize dependence on them, and thus on middle eastern oil supplies.
"Sure, who wants to sabotage PHPMyAdmin? But the voting software used in the US? That's a different story..."
I would hope that since voting is such a sensitive issue, it would be managed as tightly as the Linux and BSD kernel projects, where only well known, trusted people can check code into the project, and all code must be thoroughly reviewed by a whole team of people before it can be considered for checkin.
"open sourcing means it's easier to introduce a back door."
Open source doesn't mean that everyone can freely contribute code. Look at the Linux or BSD projects, for example. Anyone can read the code, and can make their own branch, but only a fairly rigorous process of code review and debate by well known and trusted people can get code into the project.
"Maybe the best thing is to have the federal government pay for the development of a paperless voting system. That brings up a new set of concerns though..."
What concerns? If the government issues a grant to, say, a university to develop software that achieves the spec (securely collect votes, auditable, etc.), and the result is open sourced, I'm not too worried about whether the government might perfer a particular election result.. they'd have to affect the NSF (or whoever) to affect the university to affect the developers, and then nobody inspecting the source would notice...
"Even then it's not enough. As my example of a word macro. Even if you know it's an allegedly passive type of document (like a word .doc) that STILL doesn't mean it's safe."
Good point! All that 'integration' in Windows has caused the barrier between data and code to blur. This wouldn't be so bad, except that Windows has no security model, so you can only open a "document" that does stuff (e.g. open a Word document, it runs a macro), there's no real limit on what it can do (e.g. open a word document that runs a macro that emails the document containing the macro to everyone in my address book, then adds the macro to every word document on my computer). So while I'm fan of the potential of API's to allow intelligent agents to do cool things for me, it has to be done in a manner that gives me control over what's going on.
"Who is going to install the full voting suite (voting software is much more than a voting terminal) and then hold mock elections in their home?"
Anyone with an interest in making sure that the voting process is secure -- universities, public interest groups, etc. Security analysts would scour the system in order to compete to issue their reports for their clients. Academics would scour the system for their research/publication. The same organizations that are so critical of the closed source voting systems organizations would audit the system. Heck -- look at the level of inspection that the leaked voting system code and documentation have received so far--no closed source project receives that kinda of effort outside of NASA. And, of course, any organization that holds formal votes (schools, churches, unions, etc.) could also use the system, which would both be an additional benefit (free system for them to use) and more users to flush out problems.
Of course, this all misses the most fundamental point -- voting is a critical function of the government, and must be performed in a transparent and auditable manner in order for people to trust it. Given the long and varied history of people committing voter fraud, there's no reason to think that people would _not_ attempt to manipulate the system. And with a system that can't be audited, doesn't support manual recounts, etc., there's no reason for anyone to trust it.
"The fault was not in the development model but in the failure of the project leadership."
The fault is in election boards allowing themselves to be hoodwinked by a bunch of sleazy opportunists chasing the $4B that the government has committed to updating the voting infrastructure. The closed source development model simply allows those bozos to conceal their shoddy workmanship.
"How the product is marketed is IRRELEVENT. Its how its utilized."
;-)
Apple's claim is that the dual G5 is the fastest "personal computer" -- a market segment. If you're saying that you have the fastest computer in a market segment ("personal computers"), then how the product is marketed matters quite a bit, because that defines what's in the market segment. There are rather obviously high-end workstations that can outperform the dual G5, and they can fit on a desk, but that's not terribly relevant. That's like arguing that because there are astoundingly fast specialized racecars that a mass market sports car couldn't claim to be the "fastest car in its class".
BTW, people have done some benchmarks of dial Xeon's against the Dual G5, and overall the G5 was faster. Of course, you could find some tests where the Xeon's were a little faster, but there were also tests where the G5's were a lot faster... and (though it doesn't affect performance) the fastest Xeon's cost $1-2K more than the G5. That would be (wait for it) because the dual Xeon's are sold as workstations, while the dual G5 is sold as a personal computer, thus larger volumes and lower margins.
"I've speced out machines that run circles around the G5 for $2100 ... this is with a freakin 3Ghz P4".
Cool, for $2,100 you've outperformed the $1,600 G5. I think that perhaps you could try to configure a PC to outperform the model that Apple claimed was the fastest personal computer; when people tried to pull together dual Xeon's, they were around $4-5K (depending on vendor) so $1-2K more than Apple's dual G5. You might be able to get the fastest PC price down to Apple's price by "white boxing" it...
but even if the PC is more expensive than the dual G5, it's not clear that it could outperform it. The independent benchmarks that I've seen had either machine perform "best" on some tests, but overall the dual G5 outpeformed the fastest PC's, because when the PC was faster it was 10-20% faster, and when the Mac was faster, it was up to 50% faster. I suspect it had to do with which code could take advantage of the AltiVec instruction set, which is much more powerful/flexible than Intel's comparable SIMD instruction set...
Exactly. The design error is that Outlook doesn't contain its own viewer code to render data (i.e. a safe operation) it simply tells the OS to "open" the file, so the user uses the same operation to view a photo as to execute an application. Adding to the danger, Outlook by default hides file extensions, so users have to know how to recognize which icons are "safe" to open and which are not. For example, it's "safe" to open a text file, and "dangerous" to open a screen saver file, but they act and look quite similar in Outlook.
"I would argue it has, given their recent decision to prioritize security and stability. 2000 and XP and miles ahead of 95 as well."
It may be true that 2000 and XP are "miles ahead of 95" but that's not due to their recent emphasis on security and stability -- both operating systems were written before MS made that conversion. The NT-based versions of Windows are more stable and secure than the 9x versions because it's based on VMS, a fundamentally more secure OS model than the DOS-based 9x.
"It will accelerate the arms race between marketroids and Internet users"
This is certainly true. That being said, in the mass market right now the "marketroids" are the only ones fighting the war right now -- IE users are all on the receiving end of so much advertising that the general internet is all but useless. Sure, the advertisers will find new ways to abuse internet users, but that was going to happen anyway.
One side effect, though, will be that those of us using web browsers that have blocked pop-ups forever may be vulnerable to the next bit of advertising. But I'm confident that our browsers can move faster than IE...
The kind of "pop up" that these tools block is a pretty specific thing. They don't block every JavaScript that opens a window. They allow "pop ups" that are triggered by user actions, and block "pop ups" that are triggered by system events. So if you click on a "more info" button that pops up a window to tell you about a product while shopping, for example, you'll see the info. But if you browse a pr0n site (who would do that?!) that triggers a script whenever the page is closed that opens a new pop-up, that'll be blocked.
This sounds simple, but the impact is immense. With a browser that blocks pop-ups (i.e. anything by IE), the intenet is a pleasant place. With IE, the internet is a horrible maze of X10 camera ads and pr0n spam.
If IE finally catches up to the other browsers, and implements a good pop-up blocker that is on by default, everyone will benefit.
Sorry to hear it. Perhaps you can save the product by convincing them to use a design that doesn't suck?
"Temperature recorders made before some point in the 20th century had a design defect that makes them accurate only to about 3 degrees, which is well outside the claimed variation."
Even if any individual measurement has a varience of 3 degrees, an analysis of many thousands of measurements will virtually eliminate that imprecision. Now assuming that the "global temperature" was measured in a reasonable number of locations (hundreds of cities?) fairly often (weekly? daily?) the number of samples would be large (10K's?).
Also, the study that started the whole global warming thing covered a much larger time period than when humans were measuring the temperature with thermometers -- they measured ice rings, etc., in order to estimate the temperature over many, many thousands of years. So even if the measurements for one recent decade were off, that wouldn't change a long term trend.
"The parent went a bit far by saying they could 'shut down' a company"
Well, Microsoft could (before the monopoly ruling) shut down any PC vendor by pulling their license to Windows. Of course, they don't need to -- more subtle threats, such as inflated pricing or witholding beta releases so that the vendor can't test for compatability, have so far been sufficient to cause every PC vendor to agree to anything MS has needed from them so far. I believe that one of the providions of the settlement is that MS has to sell Windows to any vendor off of the same pricing schedule, which at least limits their power to manipulate the market a little bit.
Of course, I agree with the rest of your post completely...
"I don't doubt that in the least, but considering they are probably being paid pretty well by the RIAA to do this survey and report the results, they don't really have an obligation to announce how inaccurate the data is if it reflects the "success" of the RIAA's litigious tactics."
NPD is in the business of providing accurate market survey data to its customers, who use that data to make critical business decisions. If it ever biased a survey in order to manufacture an inaccurate outcome, it would cost them far more in lost credibility (and thus clients) than they could every make performing that survey.
"Good answers? Usually I like correct answers more..."
OK, I'll elaborate on their answers to my questions, and why I found them satisfactory. For example, when I asked them about whether people tended to read the first questions on a survey more closely, and skim the later ones, they said that they knew that, and permute their question sequence across the sample set, so that the bias would average out. This also compensates for any sequencing artifact, where one question will affect the answer to the next question, because the questions will appear in the opposite sequence 1/2 the time. They've also mentioned compensating for people's desire to please surveyors by saying "yes" more often than "no" to questions, by phrasing test questions both positively and negatively (i.e. "yes" to one means "no" to the other) so that they can model the difference compensate across other questions. And they make sure that their panel mirrors the general public in terms of age, gender, income and geographical distribution. Heck, I saw one survey where they made sure that the panelists had a representative distribution of computer configurations and modem speeds so that they didn't bias the sample by having too many panelists on broadband and fast PC's...
So I can't say that I've talked with them about this study, but I'm pretty sure that they thought of obvious factors like "people deny doing illegal things".
"Uh did you not read the part that said attractive and reasonable? iTunes is the closest so far but Apple, not the RIAA, is the driving force there."
If you're waiting for RIAA to run a music store, you'll be waiting forever. RIAA doesn't sell music, and never will -- they're a lobbying group representing the record labels, not a retailer.
If you're saying that iTMS isn't "attractive and reasonable" -- what are you suggesting should be changed?
"My current job is supporting a website that uses pop-ups to display menus, etc. I didn't design it. I have no say in how to improve it. I just have to sit on the phone and tell people to disable their pop-up blockers."
You should do some quick cost estimation encourage someone in a decision making capacity at the company to consider whether the costs to the company in customer support and in lost customers possibly outweighs whatever advantage there is to the current design.
Aside from that, using pop-ups for navigation seems like an amazingly bad idea. But saying that probably would alienate the designers without changing anyone's mind. "costing the company money" is a much better business argument than "sucks".
"The biggest security problem isn't vulnerability, it's the creativity of the people trying to be obnoxious"
While it's certainly true that if there were nobody writing viruses or cracking systems we'd all be better off, we appear to be in a world where such things happen regularly. Given that, it makes a difference whether an OS or application designer makes decisions that promote security (e.g. BSD) or insecurity (e.g. Microsoft). I'm not saying that MS has a goal of making things insecure (they're not crazy), but that they routinely decide that transparent integration between applications, or between applications and the OS, is more important than security. This is how viruses can control Outlook to automaticaly spam everyone you know, or why IE and Outlook so often are the delivery vectors for trojan horses, etc. If there were a clear barrier between data and applications, as there is in every other OS and email client, then we'd have far fewer security problems. And beyond that, Microsoft cares more about adding lots of new features, and shipping rapidly, than in security. That may be a reasonable business strategy (they're certainly making money, and have avoided any liability for the flaws in their software) but it doesn't demonstate a commitment to security.
Let's compare Exchange with Lotus Notes. They have about the same market share (in corporate email systems), but Outlook is riddled with frequent security holes, while Notes hasn't had any worth mentioning for the past few years. Since they have the same market share, that isn't the difference. And while many crackers may hate Microsoft, it's hard to argue that they love IBM, so motivation isn't the answer. Perthaps design decisions have implications?
Or let's compare Apache and IIS. Apache has a much larger market share than IIS (about 2:1), but IIS is the target of more successful intrusions. Plenty of large, rich companies run either platform, so that isn't the difference. Perhaps the Apache developers put a higher priority on security?
"If you tell me though that this month Kazaa usage was down 10% I'll assume eDonkey's went up by at least that."
According to slyck.com, KaZaA is down a lot over the last few months, and the smaller p2p networks are up a little, so overall p2p usage is down.
"The RIAA needs to get over it and offer an attractive, reasonable alternative."
You mean like iTunes, Napster, MusicMatch, etc.?