Additionally, it's trivially easy to read files off of a passworded hard drive. The password is stored in an EEPROM on the board, so all you have to do is buy an nearly identical drive and swap the circuit board to read all the documents.
Okay, I agree that this may not be a great approach, but I don't think that a procedure that requires skilled technicians and a cleanroom to do reliably qualifies as "trivial".
Okay, I *have* to know. Did you try this with the intention of dropping the internal voltage enough to zero the CMOS or were you freezing them for some other reason? I would never have thought of trying this.
That said, I hope and pray that the author was smart enough to encode it with CSS, so we can actually have an example of using Fair Use policy to circumvent CSS encryption.
The MPAA has made for some bizarre cases. First, there was the instance where they legally threatened Prof. Felton for releasing an academic paper on CSS. The EFF *really* wanted to fight a case where the MPAA was trying to squash academic discussion, so they jumped on this. The MPAA dropped the case like a hot potato and for a while, the EFF was trying to argue to keep the judge from letting the MPAA drop the case. Now, we have the case where the EFF would be defending the MPAA to support fair use rights...
The way it reads to me the filmmaker was itching to stir something up to begin with, I mean why else would you specify that the MPAA not make any copies of the film...
"The way it reads to me the MPAA was itching to stir something up to begin with, I mean why else would you specify that the consumer not make any copies of the film..."
It'll only get fun if the film maker decides to try to take it somewhere. Right now it looks like the film wasn't harmed financially, and I don't think the film maker has a major studio backing him with financial and legal muscle.
It might be a worthwhile PR case for EFF...but honestly, the publicity from an ongoing lawsuit might drive his movie sales up enough to pay for any legal costs.
It's not like **AA-folk are some kind of ideologically-driven crusaders. They're a buncha people working for a business that happens to be nastier than most, but I'd hardly expect them to be anything other than the Average Joe. And Average Joe likes downloading music.
Sounds plausible to me. After all, the MPAA is an entirely altruistic organization devoted to facilitating advances in art. They couldn't care less about money.
I doubt that the *code* per se is that insecure -- if someone was going to manipulate it, presumably it would be Diebold, and while I don't know much about the machines, I doubt they have many vectors to manipulate them.
My concern is simply that e-voting inherently is broken -- there's no guarantee of what you're sending to the vote counter. The only reason for Diebold to be making these is for Diebold to get state money. Everyone else loses out.
I strongly suspect that anyone involved in intentional vote manipulation would not be this foolhardy. If it really is incorrect, I'd sooner blame incompetence.
Open source wouldn't help e-voting. The problem with e-voting is fundamental to e-voting -- it's that there is no guarantee that the vote transmitted is the vote you see. There will *always* be a vulnerability in e-voting, with a tremendous prize for anyone who can manipulate it. This is an inherent guarantee of the traditional paper system.
For that matter, what does *Flash* do that people need?
I've seen:
* Indie animations, most of which are pretty bad.
* Custom interfaces for webpages. These are, in my experience, much slower and more annoying to navigate than regular ol' HTML interfaces, are fixed at a size too small for my father to easily read, often (irritatingly) play sounds, usually have awful color schemes -- there's a *reason* that I have my foreground and background colors set to black and white, have sluggish reimplementations of scrollbars that don't look like scrollbars, and don't really IMHO do much for anyone other than the designer, who gets to play with a fun toy for a while.
* Ads. Animated, computer-bogging-down ads. Ads with sounds. Horrible, awful things which make computers without Flashblock miserable to use. Probably the primary use of Flash today.
* Small web games. While I have played these occasionally, the best of them don't come close to the best full-blown native games.
* Splash screens, which many companies inexplicably stuff in front of their website's main page. I would assume that this is to drive the less-than-dedicated away.
I mean, seriously, how does Flash make life better for the browser *user*? Okay, granted, perhaps in some very indirect way (advertisers will maybe pay more for Flash ads, that money goes to fund the production of websites that users want), but in general, Flash doesn't seem to be a net positive for my web browsing at all.
He's cranky that everyone in the world doesn't share his opinion that LISP is the Second Coming.
Actually, in reading PG's articles, I have to say that the thing that I think I disagree most with him on is his opinions about LISP. He comes off as if he's spent *way* too many hours having to be defensive about his favorite language and now it's wormed its way into his personality.
LISP may have some nice features, but it just isn't *that* big a deal, and he gives it way too much credit for stunning returns in productivity or what-have-you.
Also, I'm not sure that if I got a successful company going that I'd want to keep growing it and eventually sell out. It might be nice to just produce a nice product and not have nonsense coming down from On High. PG talks a lot about how great it is to run out and build a business and then sell it -- but every time he describes how he felt post-takeover at Yahoo, he was obviously not happy with his situation.
I read an interesting article that pointed out that data secrecy isn't really worth all that much. It might have been Paul Graham or Joel Spolsky -- not sure. Basically, startups are always running around having people sign NDAs left and right -- but, honestly, their idea is probably nothing that crucial -- that'll buy them some market lead, but it's not going to make them stunningly rich. What matters is whether or not they can manage to compete after their competitors introduce competing products.
I'm sure that we can all thing of worst-case scenarios that are pretty scary, but honestly, does your company pay you to engage in data espionage against other companies? How likely do you think it is that your counterparts at another company are being paid to engage in data espionage?
Apple trumpeted around UI mockups of what was going to be their latest-and-greatest operating system for *years* Pink, Taligent, Copeland, Rhapsody...before finally actually sitting down and doing OS X.
OSS strengths and weaknesses
on
IE7 Leaked
·
· Score: 1
I agree that there is a place for opensource software, but I would also say that there is a place for proprietary software.
Open source software does well at:
* Production of development tools. The open source world has first rate, free development tools. Emacs+gdb+valgrind+gcc+splint+gprof+oprofile+GNU make+doxygen is one powerful toolset, and it doesn't cost a dime. Every software developer out there develops software, so the "my development tool can't do X" itch has been thoroughly scratched.
* Production of small components. (This fits nicely with the Unix methodology.) Commercial software packages often suffer from feature creep -- the developers have to keep earning their keep, and if the company has a product, it keeps getting features. On the other hand, if I'm a GIMP developer and some feature would be better suited for inclusion in ImageMagick, I can just pop over the the ImageMagick project and add my feature there.
* Security tools. Lots of paranoid crypto-types play around with open source.
* Things with a security and/or performance aspect. It's generally a lot easier to sell features to a customer than security or some (difficult to measure) performance difference. Open source developers aren't necessarily trying to sell features, so they put a higher percentage of time into security and performance work (IMHO, of course).
* Honesty. I've found that open source projects tend to be more straighforward about the failings in their code than traditional software projects are.
* Writing software with sane resource requirements. I'm boggled by the size of commercial software packages. A large Linux distribution is maybe five CDs and includes a *huge* range of software packages. A single typical commercial software package is probably at least half a CD.
The open source world does badly (most of the time) in the following areas:
* Games. Most developers don't want to work on something that *they* don't get to play themselves. If you make an adventure game yourself, it's no fun to play it. The few games that have done really well in the open source world have almost perfect replayability factor (i.e. you could play them for five years and they wouldn't lose their value). However, cinematics, flashy graphics, scripted sequences, and high-quality music are almost unheard of in the open source world -- these things are fun for a short period of time, but don't help the replayability of the game that much, so there's little benefit to the developers. There are whole genres of games that just don't really go anywhere in the OSS world because they have limited replayability -- adventure games (except for the notable exception of text-based adventure games), plot-oriented RPGs (not dungeon crawlers, which the open source world does do well), and single-player FPSes.
* Software that developers don't use. Developers are pretty advanced users. Software designed exclusively for newbies, like effective basic help systems, doesn't see much work in the open source world (unless there's some research value and some grad student's decided to do a prototype, or unless some company decides to donate some labor).
* Usability. Software companies can afford to hire some guy to devise studies to see exactly what people are doing wrong when they try to use the software, and then come up with fixes. The problem is that this sort of stuff isn't (IMHO) nearly as much fun as developing software -- and since people are generally expected to implement their own ideas, a UI guy has to also be able to code with the best of them to contribute ideas. Sun needed to fund some folks in this area for GNOME because the open source world wasn't handling these problems themselves.
* Big software packages. Okay, there are some -- xorg is big, the Linux kernel is big, and so forth. But looking at some of the big OSS packages (notably Open Office and Mozilla), you discover that these had development heavily funded by traditional software comp
I've found that Microsoft does provide pretty solid user documentation, contrary to what you are saying.
I agree that they do not provide much information about bugs compared to the Linux world (and they don't have "bugs", they have "issues" -- sigh).
Where they really fall down is on specs. POSIX has a spec. It was designed by people who knew what they were doing and were thinking hard about the future. Win32 has MSDN documentation, which is not a spec. It means that you can't easily "write to the spec" under Windows.
I have a Greasemonkey script that adds a second, Coral Cache-ized link to all the off-site links on Slashdot.
It's awfully handy, and I'd recommend that any FireFox users out there give it a shot.
The only drawbacks I've found: every now and then, you'll get many "[CC]" coral cache links in comment headers, and if you want to copy text to paste in a response, you're liable to have [CC] after each link in the original text in your pasted text.
Re:Leaks? I'll show you LEAKS!
on
IE7 Leaked
·
· Score: 2, Informative
Heap defragmentation in languages with pointers kind of sucks, though. The classic Mac OS memory manager did this was *really* aimed at not wasting any memory -- it could move chunks of memory around and even free chunks of memory that you had marked as unimportant. It turned out to be really difficult to program bug-free software for something like this.
Re:Leaks? I'll show you LEAKS!
on
IE7 Leaked
·
· Score: 1
To leak it would have to not return some RAM after the app is closed.
That would be a page leak in the kernel, and while this is not impossible (I remember some rather infamous Linux releases), it is certainly not what people are talking about when they say "leak memory". They're talking about not releasing used memory so that the application can use that memory for other things. (Theoretically, the application could release that memory for *other* applications to use, but since we don't do heap defragmentation these days, the application just reuses some and the rest gets shuffled off to virtual memory where it waits quietly until the application restarts.
Re:So where's the .torrent?
on
IE7 Leaked
·
· Score: 1
You'd install a web browser that has a less-than-stellar reputation, which is tightly tied into your OS, which is currently arriving to you in the form of a leaked beta from a random pirate site?
I mean, are you just bored and looking for more excitement and danger on your computer?
but without obnoxious "i spread my article over infinite pages in order to get more clicks" practice.
If you use Firefox as your browser, you have the antipagination plugin available, and no reason to be bothered by people that spread their articles over multiple pages.
(Also, try reading webcomics using antipagination. Woo!)
Re:Please use Kelvin to compare % temperature chan
on
Undervolting a Laptop
·
· Score: 1
I would think that you'd care about percentage deviation from ambient temperature, not the deviation from absolute zero.
You have absolutely no idea what you're talking about! How on earth did this get modded up!?
He may be a little bit more in the Windows-style commercial software approach, but what he's saying is not unreasonable.
Basically, if I'm a company, I can do what a lot of companies have done for Linux for a while now -- throw a driver out, and then not provide any specific guarantees. If it works, great, if it doesn't work, too bad. In reality, this works pretty well.
If a company is trying to seriously make a product and market to Linux users, though, they generally have to provide official support. Official support conventionally means testing under all the platforms you support. If you're supporting a piece of software under WinME and WinXP, you need to actually test it on each. You can't ship it with a "WinME support" sticker and then find out from your customers that it doesn't actually work under WinME.
Now, under Linux there are an awful lot of distros. What a lot of companies do is to almost certainly support Red Hat, then SuSE and Debian, and maybe Mandrake. They don't want to commit to supporting a platform that they haven't gone through full-blown testing on.
By the same token, the Linux API isn't as unstable as "keeping the API open" suggests. There are many drivers available in the kernel that have been there for... a LONG time. Most of them were ported to 2.6 with no trouble at all.
I would still say that there are more changes than under Windows. Devfs gets introduced, code needs to be introduced to register a driver. Linus decides (probably correctly) that the IP filtering system could be improved, and tosses out the existing stuff. The sound API changes, and vendors that arranged to have an OSS driver have to worry about an ALSA driver.
As a person who has written device drivers I can tell you that writing and maintaining a Linux driver is significantly easier. The docs and community support is all there, and everything makes sense. It's pretty much the opposite when it comes to Windows driver development.
I have done a very minimal amount of driver work on Linux and Windows, but that taste left me in full agreement with you. A given Linux driver is much shorter, simpler, and easier to write. I remember being astounded that the Linux ramdisk driver (an admittedly trivial example) was only 180 lines of code, and unimpressed with the complexity of handling pageable memory pools and power management under Windows.
Simply not true. And the beautiful part about Linux is that even if a driver does need updating, there's a significant chance that if the driver is used by enough people, some person will just fix it on their own. But let me just reiterate that this is completely untrue in most cases. At least not any more than it's true for Windows.
That's true, but that's a scary thing for a company to do -- to rely on someone else to be producing the code that affects their reputation.
But then again look at nVidia. Most Linux users are quite loyal to nV for their awesome Linux driver support. Some complain to them that they don't open source their drivers but in my opinion this is their perogative... they support it, and that's great.
I bought the one of the last ATI cards with decent open source drivers -- the Radeon 9250. I'm going to hang onto that until someone comes out with a new card with decent Linux open source drivers. It might be quite a while, but then, there aren't *that* many 3d games for Linux.
Slashdot Mirror
Additionally, it's trivially easy to read files off of a passworded hard drive. The password is stored in an EEPROM on the board, so all you have to do is buy an nearly identical drive and swap the circuit board to read all the documents.
Okay, I agree that this may not be a great approach, but I don't think that a procedure that requires skilled technicians and a cleanroom to do reliably qualifies as "trivial".
Okay, I *have* to know. Did you try this with the intention of dropping the internal voltage enough to zero the CMOS or were you freezing them for some other reason? I would never have thought of trying this.
That said, I hope and pray that the author was smart enough to encode it with CSS, so we can actually have an example of using Fair Use policy to circumvent CSS encryption.
The MPAA has made for some bizarre cases. First, there was the instance where they legally threatened Prof. Felton for releasing an academic paper on CSS. The EFF *really* wanted to fight a case where the MPAA was trying to squash academic discussion, so they jumped on this. The MPAA dropped the case like a hot potato and for a while, the EFF was trying to argue to keep the judge from letting the MPAA drop the case. Now, we have the case where the EFF would be defending the MPAA to support fair use rights...
I don't think that California's state government is very likely to go after Hollywood's powerful in any event.
The way it reads to me the filmmaker was itching to stir something up to begin with, I mean why else would you specify that the MPAA not make any copies of the film...
"The way it reads to me the MPAA was itching to stir something up to begin with, I mean why else would you specify that the consumer not make any copies of the film..."
It'll only get fun if the film maker decides to try to take it somewhere. Right now it looks like the film wasn't harmed financially, and I don't think the film maker has a major studio backing him with financial and legal muscle.
It might be a worthwhile PR case for EFF...but honestly, the publicity from an ongoing lawsuit might drive his movie sales up enough to pay for any legal costs.
It's not like **AA-folk are some kind of ideologically-driven crusaders. They're a buncha people working for a business that happens to be nastier than most, but I'd hardly expect them to be anything other than the Average Joe. And Average Joe likes downloading music.
Given that the MPAA's intent isn't financial gain
Sounds plausible to me. After all, the MPAA is an entirely altruistic organization devoted to facilitating advances in art. They couldn't care less about money.
Right or wrong, they would be able to outlast me in court, strictly from a financial standpoint.
Ah, yes. Copyright law. Defender of the creativity of the little guy.
I doubt that the *code* per se is that insecure -- if someone was going to manipulate it, presumably it would be Diebold, and while I don't know much about the machines, I doubt they have many vectors to manipulate them.
My concern is simply that e-voting inherently is broken -- there's no guarantee of what you're sending to the vote counter. The only reason for Diebold to be making these is for Diebold to get state money. Everyone else loses out.
I strongly suspect that anyone involved in intentional vote manipulation would not be this foolhardy. If it really is incorrect, I'd sooner blame incompetence.
Open source wouldn't help e-voting. The problem with e-voting is fundamental to e-voting -- it's that there is no guarantee that the vote transmitted is the vote you see. There will *always* be a vulnerability in e-voting, with a tremendous prize for anyone who can manipulate it. This is an inherent guarantee of the traditional paper system.
For that matter, what does *Flash* do that people need?
I've seen:
* Indie animations, most of which are pretty bad.
* Custom interfaces for webpages. These are, in my experience, much slower and more annoying to navigate than regular ol' HTML interfaces, are fixed at a size too small for my father to easily read, often (irritatingly) play sounds, usually have awful color schemes -- there's a *reason* that I have my foreground and background colors set to black and white, have sluggish reimplementations of scrollbars that don't look like scrollbars, and don't really IMHO do much for anyone other than the designer, who gets to play with a fun toy for a while.
* Ads. Animated, computer-bogging-down ads. Ads with sounds. Horrible, awful things which make computers without Flashblock miserable to use. Probably the primary use of Flash today.
* Small web games. While I have played these occasionally, the best of them don't come close to the best full-blown native games.
* Splash screens, which many companies inexplicably stuff in front of their website's main page. I would assume that this is to drive the less-than-dedicated away.
I mean, seriously, how does Flash make life better for the browser *user*? Okay, granted, perhaps in some very indirect way (advertisers will maybe pay more for Flash ads, that money goes to fund the production of websites that users want), but in general, Flash doesn't seem to be a net positive for my web browsing at all.
He's cranky that everyone in the world doesn't share his opinion that LISP is the Second Coming.
Actually, in reading PG's articles, I have to say that the thing that I think I disagree most with him on is his opinions about LISP. He comes off as if he's spent *way* too many hours having to be defensive about his favorite language and now it's wormed its way into his personality.
LISP may have some nice features, but it just isn't *that* big a deal, and he gives it way too much credit for stunning returns in productivity or what-have-you.
Also, I'm not sure that if I got a successful company going that I'd want to keep growing it and eventually sell out. It might be nice to just produce a nice product and not have nonsense coming down from On High. PG talks a lot about how great it is to run out and build a business and then sell it -- but every time he describes how he felt post-takeover at Yahoo, he was obviously not happy with his situation.
I read an interesting article that pointed out that data secrecy isn't really worth all that much. It might have been Paul Graham or Joel Spolsky -- not sure. Basically, startups are always running around having people sign NDAs left and right -- but, honestly, their idea is probably nothing that crucial -- that'll buy them some market lead, but it's not going to make them stunningly rich. What matters is whether or not they can manage to compete after their competitors introduce competing products.
I'm sure that we can all thing of worst-case scenarios that are pretty scary, but honestly, does your company pay you to engage in data espionage against other companies? How likely do you think it is that your counterparts at another company are being paid to engage in data espionage?
Apple trumpeted around UI mockups of what was going to be their latest-and-greatest operating system for *years* Pink, Taligent, Copeland, Rhapsody...before finally actually sitting down and doing OS X.
I agree that there is a place for opensource software, but I would also say that there is a place for proprietary software.
Open source software does well at:
* Production of development tools. The open source world has first rate, free development tools. Emacs+gdb+valgrind+gcc+splint+gprof+oprofile+GNU make+doxygen is one powerful toolset, and it doesn't cost a dime. Every software developer out there develops software, so the "my development tool can't do X" itch has been thoroughly scratched.
* Production of small components. (This fits nicely with the Unix methodology.) Commercial software packages often suffer from feature creep -- the developers have to keep earning their keep, and if the company has a product, it keeps getting features. On the other hand, if I'm a GIMP developer and some feature would be better suited for inclusion in ImageMagick, I can just pop over the the ImageMagick project and add my feature there.
* Security tools. Lots of paranoid crypto-types play around with open source.
* Things with a security and/or performance aspect. It's generally a lot easier to sell features to a customer than security or some (difficult to measure) performance difference. Open source developers aren't necessarily trying to sell features, so they put a higher percentage of time into security and performance work (IMHO, of course).
* Honesty. I've found that open source projects tend to be more straighforward about the failings in their code than traditional software projects are.
* Writing software with sane resource requirements. I'm boggled by the size of commercial software packages. A large Linux distribution is maybe five CDs and includes a *huge* range of software packages. A single typical commercial software package is probably at least half a CD.
The open source world does badly (most of the time) in the following areas:
* Games. Most developers don't want to work on something that *they* don't get to play themselves. If you make an adventure game yourself, it's no fun to play it. The few games that have done really well in the open source world have almost perfect replayability factor (i.e. you could play them for five years and they wouldn't lose their value). However, cinematics, flashy graphics, scripted sequences, and high-quality music are almost unheard of in the open source world -- these things are fun for a short period of time, but don't help the replayability of the game that much, so there's little benefit to the developers. There are whole genres of games that just don't really go anywhere in the OSS world because they have limited replayability -- adventure games (except for the notable exception of text-based adventure games), plot-oriented RPGs (not dungeon crawlers, which the open source world does do well), and single-player FPSes.
* Software that developers don't use. Developers are pretty advanced users. Software designed exclusively for newbies, like effective basic help systems, doesn't see much work in the open source world (unless there's some research value and some grad student's decided to do a prototype, or unless some company decides to donate some labor).
* Usability. Software companies can afford to hire some guy to devise studies to see exactly what people are doing wrong when they try to use the software, and then come up with fixes. The problem is that this sort of stuff isn't (IMHO) nearly as much fun as developing software -- and since people are generally expected to implement their own ideas, a UI guy has to also be able to code with the best of them to contribute ideas. Sun needed to fund some folks in this area for GNOME because the open source world wasn't handling these problems themselves.
* Big software packages. Okay, there are some -- xorg is big, the Linux kernel is big, and so forth. But looking at some of the big OSS packages (notably Open Office and Mozilla), you discover that these had development heavily funded by traditional software comp
I've found that Microsoft does provide pretty solid user documentation, contrary to what you are saying.
I agree that they do not provide much information about bugs compared to the Linux world (and they don't have "bugs", they have "issues" -- sigh).
Where they really fall down is on specs. POSIX has a spec. It was designed by people who knew what they were doing and were thinking hard about the future. Win32 has MSDN documentation, which is not a spec. It means that you can't easily "write to the spec" under Windows.
I have a Greasemonkey script that adds a second, Coral Cache-ized link to all the off-site links on Slashdot.
It's awfully handy, and I'd recommend that any FireFox users out there give it a shot.
The only drawbacks I've found: every now and then, you'll get many "[CC]" coral cache links in comment headers, and if you want to copy text to paste in a response, you're liable to have [CC] after each link in the original text in your pasted text.
Heap defragmentation in languages with pointers kind of sucks, though. The classic Mac OS memory manager did this was *really* aimed at not wasting any memory -- it could move chunks of memory around and even free chunks of memory that you had marked as unimportant. It turned out to be really difficult to program bug-free software for something like this.
To leak it would have to not return some RAM after the app is closed.
That would be a page leak in the kernel, and while this is not impossible (I remember some rather infamous Linux releases), it is certainly not what people are talking about when they say "leak memory". They're talking about not releasing used memory so that the application can use that memory for other things. (Theoretically, the application could release that memory for *other* applications to use, but since we don't do heap defragmentation these days, the application just reuses some and the rest gets shuffled off to virtual memory where it waits quietly until the application restarts.
You'd install a web browser that has a less-than-stellar reputation, which is tightly tied into your OS, which is currently arriving to you in the form of a leaked beta from a random pirate site?
I mean, are you just bored and looking for more excitement and danger on your computer?
but without obnoxious "i spread my article over infinite pages in order to get more clicks" practice.
If you use Firefox as your browser, you have the antipagination plugin available, and no reason to be bothered by people that spread their articles over multiple pages.
(Also, try reading webcomics using antipagination. Woo!)
I would think that you'd care about percentage deviation from ambient temperature, not the deviation from absolute zero.
You have absolutely no idea what you're talking about! How on earth did this get modded up!?
He may be a little bit more in the Windows-style commercial software approach, but what he's saying is not unreasonable.
Basically, if I'm a company, I can do what a lot of companies have done for Linux for a while now -- throw a driver out, and then not provide any specific guarantees. If it works, great, if it doesn't work, too bad. In reality, this works pretty well.
If a company is trying to seriously make a product and market to Linux users, though, they generally have to provide official support. Official support conventionally means testing under all the platforms you support. If you're supporting a piece of software under WinME and WinXP, you need to actually test it on each. You can't ship it with a "WinME support" sticker and then find out from your customers that it doesn't actually work under WinME.
Now, under Linux there are an awful lot of distros. What a lot of companies do is to almost certainly support Red Hat, then SuSE and Debian, and maybe Mandrake. They don't want to commit to supporting a platform that they haven't gone through full-blown testing on.
By the same token, the Linux API isn't as unstable as "keeping the API open" suggests. There are many drivers available in the kernel that have been there for... a LONG time. Most of them were ported to 2.6 with no trouble at all.
I would still say that there are more changes than under Windows. Devfs gets introduced, code needs to be introduced to register a driver. Linus decides (probably correctly) that the IP filtering system could be improved, and tosses out the existing stuff. The sound API changes, and vendors that arranged to have an OSS driver have to worry about an ALSA driver.
As a person who has written device drivers I can tell you that writing and maintaining a Linux driver is significantly easier. The docs and community support is all there, and everything makes sense. It's pretty much the opposite when it comes to Windows driver development.
I have done a very minimal amount of driver work on Linux and Windows, but that taste left me in full agreement with you. A given Linux driver is much shorter, simpler, and easier to write. I remember being astounded that the Linux ramdisk driver (an admittedly trivial example) was only 180 lines of code, and unimpressed with the complexity of handling pageable memory pools and power management under Windows.
Simply not true. And the beautiful part about Linux is that even if a driver does need updating, there's a significant chance that if the driver is used by enough people, some person will just fix it on their own. But let me just reiterate that this is completely untrue in most cases. At least not any more than it's true for Windows.
That's true, but that's a scary thing for a company to do -- to rely on someone else to be producing the code that affects their reputation.
But then again look at nVidia. Most Linux users are quite loyal to nV for their awesome Linux driver support. Some complain to them that they don't open source their drivers but in my opinion this is their perogative... they support it, and that's great.
I bought the one of the last ATI cards with decent open source drivers -- the Radeon 9250. I'm going to hang onto that until someone comes out with a new card with decent Linux open source drivers. It might be quite a while, but then, there aren't *that* many 3d games for Linux.