Schneier is well respected, and his opinion generally means a lot to people who actually care about security. And yes, I do prefer news, but then, I did not know about the real events.
It's not hard for them to make it mandatory. I've only recently figured out how to tweak the registry to allow me to disable automatic updates again. So all they have to do is change that registry setting and make it a critical update...
After calling up a guru friend of mine, I did eventually figure out how to turn DHCP on,
Huh. Mine Just Worked. But maybe it was a bit later than 98, anyway. I know I was using 98, and if 2k was out, it was actually less stable at the time.
Because they thought it would help more people than it actually helped?
I don't know, I just thought working software with a decent interface helps people more than "It looks like you're writing a letter."
Apple got rid of their nice help system in Mac OS 7.0+ that could actually circle the relevant parts of the screen and do other nifty stuff in favor of... just HTML files.
Speaking of circling parts of the screen, the Preferences screen is searchable. I can type the name of the setting I'm trying to change, and it'll dim the panel and hilight relevant things, with brightness depending on how relevant. For instance, typing "keys" hilights Displays, Keyboard & Mouse, Sound, QuickTime, and (slightly brighter) Universal Access, because in the autocomplete menu, the first suggestion is "Sticky Keys".
But HTML files are actually quite useful, and I find them a lot more helpful than help that comes after me.
And, hey, at least Office *has* help. That's more than you can say for a good proportion of Linux applications.
That's not only below the belt, it's irrelevant. I don't remember random apps like WinAmp having help, but if you want to compare apples to apples, OpenOffice has help. It even has an assistant, but the assistant is a non-animated light bulb that sits in the bottom right of the screen, whereas Clippy jumps up on top of whatever you're doing and blinks at you.
The people working on Aero are not on the same team as the people working on IE.
IE obviously needs more people working on it -- why not pull them from the Aero team?
And if it really is in such bad shape, why not give up and license Safari, or borrow Konqueror or Firefox? I really don't see a reason they need IE, other than to drag down web standards so Google Office doesn't become a reality and replace Windows.
Maybe a tinfoil hat theory, but do you have a better one?
Do you think it might be possible that Aero improves usability? Obviously, you're part of the Linux elite, so I'm sure you don't give a crap if mere mortals can use your software or not, but I happen to think improving usability is a noble cause.
Usability is like optimization. The usability of a nonworking program is irrelevant.
Make it work, first. Security, stability, standards compliance. Then we'll talk about pretty translucency and bouncy effects -- which Linux and OS X have already, by the way, but our software works.
Your metaphor doesn't work, because the Wii is simply more fun. You don't buy a car for fun, you buy it to get to work. You buy a game console for fun, not for eye candy and assorted BS.
The 20 somethings already bought a 360, just because Halo 3 will be on it. They don't want to have to buy a PS3, too, but they might still buy a Wii, because it's going to be so much cheaper.
PCs and MS operating systems do get bought by slashdotters -- at least 50% of us. But there isn't any sort of monopoly forcing us to buy a PS3. I know if I do, it'll be because Sony has made significant good-faith efforts (the onboard Linux should include ALL source code), the system is every bit as awesome as it's hyped to be (yay Cell), and the price has gone down significantly to where I won't just be upgrading my PC.
I don't know, I doubt they would have gotten it out the door any sooner. Look how much the Wii hurt them. Basically, they can't bear to not have every single feature everyone else has, and then some.
That, and Blu Ray does make sense here. I don't want to have a scenario like the old Final Fantasy games, where you needed four discs for one game. But then, I guess if the four discs combined cost less than a fourth of one Blu-Ray disc, not to mention savings on the price of the whole console, you may have a point...
Alpha means: "We're still working on it, but it kind of works, so go play with it."
Beta means: "Nothing major's going to change, but we want you to test it and help us shake out the bugs."
Release candidate means: "None of our Beta testers or developers can break it anymore."
If bugs are found in rc1, you fix them and put out rc2. You keep doing this until an rc -- no matter how late, could be rc15 -- survives for a fixed amount of time (usually measured in months) without any bugs reported at all. At that point, that particular rc is released, exactly as it was.
There is some fuzziness about what's pre-alpha, alpha, or beta. It's my opinion that MS betas are alpha quality, compared to the rest of the industry. But putting out a "release candidate" with known bugs is pure marketing bullshit, to keep them from getting crucified for further delays. When they "release" software, that's more marketing bullshit -- XP was certainly a release candidate before SP1, and arguably before SP2. Would you please stop defending their marketing bullshit?
CmdrTaco's last Windows experience might be with Windows 3.11, or maybe Windows 95, and yeah, those crashed. So did Mac OS at the same period of time. And while Linux may have been more stable, you couldn't DO jack with it (at least compared to Windows 95 and Mac OS 7.)
Around Win95 or 98, things were actually looking pretty good for Linux. Many developers were still using OpenGL, so Wine could play games. Windows crashed horribly, and was incredibly slow compared to anything else, even a Mac. Linux just seemed to be getting everything right, so that's when we were pushing the hardest for desktop users, because it was so much better, and we wanted to have native software -- sure, Wine works now, but it never worked for everything, certainly not drivers to really obscure hardware.
Constant mentions of "Clippy", which has been turned off by default for ages.
I seem to remember seeing it in Office 2000, but maybe it was my imagination. I just don't install it.
I think the point is, why would they spend resources developing crap like that, instead of making the software actually work? No one complains when Apple develops eye candy, because OS X mostly works. I still have a few complaints, but I can live with them. But MS continues to develop cute things like Vista's Aero, while IE still doesn't even come close to being a decent piece of software. It's not even secure, let alone standards compliant. Apple gets a pass on making things like Dashboard and core animations, because Safari already passes Acid2, and they gave that code back to Konqueror.
The list goes on and on. Every other modern desktop OS has long, long since gotten the stability and security issues pretty well solved. They run fast, you never have to worry about spyware or other crap. Basically, all the engineering is done. We've done our homework, so we get to play with eye candy. Yet MS continues to create things like Clippy, which annoy most people, and yeah, some people like it.
But their priorities are, and always have been, severely fucked up, and Clippy is the most visible reminder.
Mentions of Microsoft Bob. If I posted about how much Red Hat sucked in 1994, you'd get turned into -1 Flamebait instantly here. If you post about how much Microsoft Bob sucked, you'll get a +5 Informative.
I don't think Red Hat tried to do cutesy eye candy in 1994. And posts about MS Bob get +5 Funny -- you must have no sense of humor.
And it does say something interesting that the best I ever hear from Windows users is that Windows is really good enough, and you should just give up and use it, because then you won't have to worry about whether you can run your favorite software. When a Windows user wants to put down OS X, you see exactly the same kind of crap -- "I want a right mouse button!" And usually not even an actual complaint -- usually it's "If OS X was mainstream, it would have just as many security issues as Windows." Besides being flat-out wrong, it's also a sort of a pathetic insult -- "You aren't any better than we are!" Yeah, great... So why's that so bad?
In fact, even more often than that, we have MS apologists -- basically, you can't really find anything bad to say about other OSes, or anything good to say about Windows, all you can say is "Gee, people sure like to bash MS!"
And how, exactly, would a bug in their code force a reboot? On Linux, I just restart the service. Some of them (Apache) can even be restarted gracefully (no interruption in service), some (anything under xinetd) are effectively restarted automatically every connection.
But regardless, why would you have to reboot? I could believe it's Norton or McAffee, but a server app shouldn't have to touch anything deep enough to screw up XP (any more than it already is).
But seriously -- we're glad to have another contributor, especially as he lends credibility to Mozilla (and takes it from Microsoft). We honestly don't know whether he's the reason IE sucks, or whether it's something in corporate culture, or what.
And of course, Slashdot isn't always groupthink. We do sometimes diasgree with each other!
Of course it was the design. Question is whether it's an evil design. If the MS.NET starts really acting badly, we can always run Mono on Windows -- we do that already anyway.
Psyco doesn't work on 64-bit. Pypy doesn't work for me at all. But IronPython does work. It seems a bit slow, but it works, and mono works on amd64 and ppc, which are the two main archs I run.
I noticed this awhile ago, I didn't know it had been given a name.
Software goes to hardware when a significant amount of CPU is being spent on a well-defined task that can be done much more efficiently in specialized hardware. I'm not convinced AI is that well defined right now.
Hardware goes to software when the procedure assigned to hardware becomes too limiting, and we need flexibility again. An example of this is shading -- adding little, programmable software routines between the huge graphical hardware crunching. Hardware also goes to software when the procedure in question can be done in software using almost no CPU, and the specialized hardware is much more expensive than an extra few mhz of CPU.
So, I wouldn't be surprised by either outcome, at this point -- it's all about the economics. And it's not just about whether it's cheaper, it's about whether it's cheaper enough to compel enough game developers to use it, and whether it makes a noticeable improvement when you actually play the game.
Depending on how much cheaper it is, or what the difference in quality between revisions is, I'd probably rather buy them individually if they'll be used at all.
nVidia does actually want to make a "game card" -- they hate x86, and I don't blame them. But why are we running computers in the first place, and not consoles?
Customizability and upgradability. Take just the parts you need, only upgrade the bottlenecks.
That means that I would much rather have many, many little cards than one huge one. It may require a fundamentally different architecture -- even if there were enough slots to go around, there won't be enough cooling, and it'll be cumbersome as hell -- but that doesn't seem like a huge deal. I wouln't even object to putting the whole assembly in a separate box.
On the other hand, if I understand it, the point of the Cell processor is to do exactly that one one chip, and call that the CPU -- just throw in all kinds of different cores that are good at various things (on a more mathematical than practical level -- not physics vs AI, more like floating-point vs integer vs boolean logic), distribute the program between these cores. No need even for a graphics card. But I may be completely talking out my ass here.
Please stop assuming you can just keep adding latency to zone changes or level loads. Games are much more fun with dynamic loading. Also, I strongly suspect that as you've pointed out, AI is simply not enough of a concern to do much to a CPU -- at this point, your money is likely far better spent simply buying a slightly faster CPU and forgetting about it.
Except now they have to fool the human that's in the loop as well.
Fool the human into thinking what, exactly? What should the requirement be for gaining access?
I suppose you could always go the route of the age verification systems for pornography -- require a credit card, even if you claim you'll never charge it. But that still doesn't prevent multiple registrations by the same person.
This also makes it incredibly cumbersome for legitimate users to get in. Fill out a captcha, send an application to webmaster, wait for a reply, send another reply back...
And of course, the whole system seems vulnerable to the same attack. Pay someone money to break captchas and then break the subsequent authorization request.
There are people who still buy the HP 200LX handhelds. They run DOS on a 3 mhz processor, but because it's DOS, it runs faster than any of the "modern" Windows Mobile crap.
The real problem with this scheme is it means that instead of putting spam in your comments, they'll just put spam in their authorization requests. Before long, you'll have to either take down authorizations (and open your blog to the world) or take down comments at all.
I think what you're missing is that the value in real things comes from the real needs that they meet, not the fact they can be converted to money.
No, that's how it starts. Valuable item -> Food. Now we have valuable item -> money -> food. Since money can buy food, anything that can become money has real value.
And we can certainly agree that there are some quite unnecessary things that people are willing to spend money on -- more money, in fact, than they could possibly need for food or shelter in their entire life. So value != necessity.
Toy money in games has value because a small number of people like to play games with toy money. But it only has value to them, and maybe to speculators who have noticed this. But in general, it doesn't have value because in general no-one is interested in it...
Well, in general, no one is interested in the work that I do, and I doubt anyone outside the company I work for would care to pay me. But they do pay me.
Maybe it all comes back to real needs, but in reality, I can buy food, medical care, etc, with money. Thus, money has real value, because we all agree it does. My work has real value, because someone agrees it does and is willing to pay me money for it.
Money is just the universally agreed representation of the value humans, as a whole, place on real things. It's a symbol of value, not actual value.
Wrong. It's a container of value, and a universally agreed upon benchmark of value.
Anyway, the reason I say game money has value, and stealing game money should be a crime, is you can do this: Steal game money -> Sell game money for real money -> stop (real money == value). You can also do this: Work in-game -> generate in-game money -> sell in-game money for real money -> feed your family. This does happen in reality -- there are gold farmers in China who feed their families exclusively by making and selling World of Warcraft gold.
Now, the difficult question is, how do we define what things are legally allowed to have value, and how to punish them in the real world? I imagine most games make it illegal to buy and sell in game items for real money to avoid questions like these, because then -- well, what if they ban you? They're suddenly responsible for more than just your subscription fee -- you could sue them for all of your in-game possessions. And virtual worlds change and evolve so quickly, and they can die so easily, that it becomes a real challenge figuring out how to define, clearly, when a virtual item has real value and when it doesn't.
We should start by looking at our banking system, and figuring out why it is that money in a bank is real, but money in, say, Dope Wars, is not real, and never will be.
First of all, your step 2 is unnecessary if the CD is layed out at all sanely. At least with mkisofs, you have to use a file as the boot floppy image anyway, and this file is visible within the ISO filesystem. Thus, all you have to do is open up the CD and look around.
But also, my typical solution here was, download a FreeDOS boot CD, and put whatever tools I need either on a floppy or on a temporary FAT partition.
As for the real solution, I'd argue we just need standard, documented ways to flash the BIOS from inside an OS. But there are more intelligent ways. From what I'm told, OS X includes BIOS updates in Software Update, and flashes the BIOS when you reboot -- I've never seen it, but maybe I missed it. That'd be the second best way -- force a reboot, but at least this way you don't have to repartition. Third way would be loading from a partition on the hard disk -- just shrink your FS by a bit, make a tiny FAT partition, then type in the pathname, or have it check in standard places. Your solution is a distant fourth, and in fact, I believe my last computer did this, from a floppy. Some of the newer laptops do it from a USB keychain.
I would still much rather have the system be open enough that we didn't have to beg for things like these. Just use LinuxBIOS and have a standard way of flashing it -- then you can flash it from anything Linux can read, including network, NTFS partitions, the works.
Slashdot Mirror
Schneier is well respected, and his opinion generally means a lot to people who actually care about security. And yes, I do prefer news, but then, I did not know about the real events.
It's not hard for them to make it mandatory. I've only recently figured out how to tweak the registry to allow me to disable automatic updates again. So all they have to do is change that registry setting and make it a critical update...
Huh. Mine Just Worked. But maybe it was a bit later than 98, anyway. I know I was using 98, and if 2k was out, it was actually less stable at the time.
I don't know, I just thought working software with a decent interface helps people more than "It looks like you're writing a letter."
Speaking of circling parts of the screen, the Preferences screen is searchable. I can type the name of the setting I'm trying to change, and it'll dim the panel and hilight relevant things, with brightness depending on how relevant. For instance, typing "keys" hilights Displays, Keyboard & Mouse, Sound, QuickTime, and (slightly brighter) Universal Access, because in the autocomplete menu, the first suggestion is "Sticky Keys".
But HTML files are actually quite useful, and I find them a lot more helpful than help that comes after me.
That's not only below the belt, it's irrelevant. I don't remember random apps like WinAmp having help, but if you want to compare apples to apples, OpenOffice has help. It even has an assistant, but the assistant is a non-animated light bulb that sits in the bottom right of the screen, whereas Clippy jumps up on top of whatever you're doing and blinks at you.
IE obviously needs more people working on it -- why not pull them from the Aero team?
And if it really is in such bad shape, why not give up and license Safari, or borrow Konqueror or Firefox? I really don't see a reason they need IE, other than to drag down web standards so Google Office doesn't become a reality and replace Windows.
Maybe a tinfoil hat theory, but do you have a better one?
Usability is like optimization. The usability of a nonworking program is irrelevant.
Make it work, first. Security, stability, standards compliance. Then we'll talk about pretty translucency and bouncy effects -- which Linux and OS X have already, by the way, but our software works.
Your metaphor doesn't work, because the Wii is simply more fun. You don't buy a car for fun, you buy it to get to work. You buy a game console for fun, not for eye candy and assorted BS.
The 20 somethings already bought a 360, just because Halo 3 will be on it. They don't want to have to buy a PS3, too, but they might still buy a Wii, because it's going to be so much cheaper.
PCs and MS operating systems do get bought by slashdotters -- at least 50% of us. But there isn't any sort of monopoly forcing us to buy a PS3. I know if I do, it'll be because Sony has made significant good-faith efforts (the onboard Linux should include ALL source code), the system is every bit as awesome as it's hyped to be (yay Cell), and the price has gone down significantly to where I won't just be upgrading my PC.
See, gamers actually have choices.
I don't know, I doubt they would have gotten it out the door any sooner. Look how much the Wii hurt them. Basically, they can't bear to not have every single feature everyone else has, and then some.
That, and Blu Ray does make sense here. I don't want to have a scenario like the old Final Fantasy games, where you needed four discs for one game. But then, I guess if the four discs combined cost less than a fourth of one Blu-Ray disc, not to mention savings on the price of the whole console, you may have a point...
Get it through your head!
Alpha means: "We're still working on it, but it kind of works, so go play with it."
Beta means: "Nothing major's going to change, but we want you to test it and help us shake out the bugs."
Release candidate means: "None of our Beta testers or developers can break it anymore."
If bugs are found in rc1, you fix them and put out rc2. You keep doing this until an rc -- no matter how late, could be rc15 -- survives for a fixed amount of time (usually measured in months) without any bugs reported at all. At that point, that particular rc is released, exactly as it was.
There is some fuzziness about what's pre-alpha, alpha, or beta. It's my opinion that MS betas are alpha quality, compared to the rest of the industry. But putting out a "release candidate" with known bugs is pure marketing bullshit, to keep them from getting crucified for further delays. When they "release" software, that's more marketing bullshit -- XP was certainly a release candidate before SP1, and arguably before SP2. Would you please stop defending their marketing bullshit?
Around Win95 or 98, things were actually looking pretty good for Linux. Many developers were still using OpenGL, so Wine could play games. Windows crashed horribly, and was incredibly slow compared to anything else, even a Mac. Linux just seemed to be getting everything right, so that's when we were pushing the hardest for desktop users, because it was so much better, and we wanted to have native software -- sure, Wine works now, but it never worked for everything, certainly not drivers to really obscure hardware.
I seem to remember seeing it in Office 2000, but maybe it was my imagination. I just don't install it.
I think the point is, why would they spend resources developing crap like that, instead of making the software actually work? No one complains when Apple develops eye candy, because OS X mostly works. I still have a few complaints, but I can live with them. But MS continues to develop cute things like Vista's Aero, while IE still doesn't even come close to being a decent piece of software. It's not even secure, let alone standards compliant. Apple gets a pass on making things like Dashboard and core animations, because Safari already passes Acid2, and they gave that code back to Konqueror.
The list goes on and on. Every other modern desktop OS has long, long since gotten the stability and security issues pretty well solved. They run fast, you never have to worry about spyware or other crap. Basically, all the engineering is done. We've done our homework, so we get to play with eye candy. Yet MS continues to create things like Clippy, which annoy most people, and yeah, some people like it.
But their priorities are, and always have been, severely fucked up, and Clippy is the most visible reminder.
I don't think Red Hat tried to do cutesy eye candy in 1994. And posts about MS Bob get +5 Funny -- you must have no sense of humor.
And it does say something interesting that the best I ever hear from Windows users is that Windows is really good enough, and you should just give up and use it, because then you won't have to worry about whether you can run your favorite software. When a Windows user wants to put down OS X, you see exactly the same kind of crap -- "I want a right mouse button!" And usually not even an actual complaint -- usually it's "If OS X was mainstream, it would have just as many security issues as Windows." Besides being flat-out wrong, it's also a sort of a pathetic insult -- "You aren't any better than we are!" Yeah, great... So why's that so bad?
In fact, even more often than that, we have MS apologists -- basically, you can't really find anything bad to say about other OSes, or anything good to say about Windows, all you can say is "Gee, people sure like to bash MS!"
And how, exactly, would a bug in their code force a reboot? On Linux, I just restart the service. Some of them (Apache) can even be restarted gracefully (no interruption in service), some (anything under xinetd) are effectively restarted automatically every connection.
But regardless, why would you have to reboot? I could believe it's Norton or McAffee, but a server app shouldn't have to touch anything deep enough to screw up XP (any more than it already is).
I am lying.
Wikipedia knows what defenestrate means, even wrt MS.
It's called "Open Source" for a reason. I've never contributed a line of code, but I have all of the source on my box.
Unless there's an anti-MS clause in the Mozilla license, but I know there isn't in the GPL...
[family_guy]Can't it be both?[/family_guy]
But seriously -- we're glad to have another contributor, especially as he lends credibility to Mozilla (and takes it from Microsoft). We honestly don't know whether he's the reason IE sucks, or whether it's something in corporate culture, or what.
And of course, Slashdot isn't always groupthink. We do sometimes diasgree with each other!
And by the way, Python has been around for awhile, this is just a different, potentially better implementation.
Of course it was the design. Question is whether it's an evil design. If the MS .NET starts really acting badly, we can always run Mono on Windows -- we do that already anyway.
Psyco doesn't work on 64-bit. Pypy doesn't work for me at all. But IronPython does work. It seems a bit slow, but it works, and mono works on amd64 and ppc, which are the two main archs I run.
I noticed this awhile ago, I didn't know it had been given a name.
Software goes to hardware when a significant amount of CPU is being spent on a well-defined task that can be done much more efficiently in specialized hardware. I'm not convinced AI is that well defined right now.
Hardware goes to software when the procedure assigned to hardware becomes too limiting, and we need flexibility again. An example of this is shading -- adding little, programmable software routines between the huge graphical hardware crunching. Hardware also goes to software when the procedure in question can be done in software using almost no CPU, and the specialized hardware is much more expensive than an extra few mhz of CPU.
So, I wouldn't be surprised by either outcome, at this point -- it's all about the economics. And it's not just about whether it's cheaper, it's about whether it's cheaper enough to compel enough game developers to use it, and whether it makes a noticeable improvement when you actually play the game.
Depending on how much cheaper it is, or what the difference in quality between revisions is, I'd probably rather buy them individually if they'll be used at all.
nVidia does actually want to make a "game card" -- they hate x86, and I don't blame them. But why are we running computers in the first place, and not consoles?
Customizability and upgradability. Take just the parts you need, only upgrade the bottlenecks.
That means that I would much rather have many, many little cards than one huge one. It may require a fundamentally different architecture -- even if there were enough slots to go around, there won't be enough cooling, and it'll be cumbersome as hell -- but that doesn't seem like a huge deal. I wouln't even object to putting the whole assembly in a separate box.
On the other hand, if I understand it, the point of the Cell processor is to do exactly that one one chip, and call that the CPU -- just throw in all kinds of different cores that are good at various things (on a more mathematical than practical level -- not physics vs AI, more like floating-point vs integer vs boolean logic), distribute the program between these cores. No need even for a graphics card. But I may be completely talking out my ass here.
Please stop assuming you can just keep adding latency to zone changes or level loads. Games are much more fun with dynamic loading. Also, I strongly suspect that as you've pointed out, AI is simply not enough of a concern to do much to a CPU -- at this point, your money is likely far better spent simply buying a slightly faster CPU and forgetting about it.
Fool the human into thinking what, exactly? What should the requirement be for gaining access?
I suppose you could always go the route of the age verification systems for pornography -- require a credit card, even if you claim you'll never charge it. But that still doesn't prevent multiple registrations by the same person.
This also makes it incredibly cumbersome for legitimate users to get in. Fill out a captcha, send an application to webmaster, wait for a reply, send another reply back...
And of course, the whole system seems vulnerable to the same attack. Pay someone money to break captchas and then break the subsequent authorization request.
There are people who still buy the HP 200LX handhelds. They run DOS on a 3 mhz processor, but because it's DOS, it runs faster than any of the "modern" Windows Mobile crap.
The real problem with this scheme is it means that instead of putting spam in your comments, they'll just put spam in their authorization requests. Before long, you'll have to either take down authorizations (and open your blog to the world) or take down comments at all.
No, that's how it starts. Valuable item -> Food. Now we have valuable item -> money -> food. Since money can buy food, anything that can become money has real value.
And we can certainly agree that there are some quite unnecessary things that people are willing to spend money on -- more money, in fact, than they could possibly need for food or shelter in their entire life. So value != necessity.
Well, in general, no one is interested in the work that I do, and I doubt anyone outside the company I work for would care to pay me. But they do pay me.
Maybe it all comes back to real needs, but in reality, I can buy food, medical care, etc, with money. Thus, money has real value, because we all agree it does. My work has real value, because someone agrees it does and is willing to pay me money for it.
Wrong. It's a container of value, and a universally agreed upon benchmark of value.
Anyway, the reason I say game money has value, and stealing game money should be a crime, is you can do this: Steal game money -> Sell game money for real money -> stop (real money == value). You can also do this: Work in-game -> generate in-game money -> sell in-game money for real money -> feed your family. This does happen in reality -- there are gold farmers in China who feed their families exclusively by making and selling World of Warcraft gold.
Now, the difficult question is, how do we define what things are legally allowed to have value, and how to punish them in the real world? I imagine most games make it illegal to buy and sell in game items for real money to avoid questions like these, because then -- well, what if they ban you? They're suddenly responsible for more than just your subscription fee -- you could sue them for all of your in-game possessions. And virtual worlds change and evolve so quickly, and they can die so easily, that it becomes a real challenge figuring out how to define, clearly, when a virtual item has real value and when it doesn't.
We should start by looking at our banking system, and figuring out why it is that money in a bank is real, but money in, say, Dope Wars, is not real, and never will be.
First of all, your step 2 is unnecessary if the CD is layed out at all sanely. At least with mkisofs, you have to use a file as the boot floppy image anyway, and this file is visible within the ISO filesystem. Thus, all you have to do is open up the CD and look around.
But also, my typical solution here was, download a FreeDOS boot CD, and put whatever tools I need either on a floppy or on a temporary FAT partition.
As for the real solution, I'd argue we just need standard, documented ways to flash the BIOS from inside an OS. But there are more intelligent ways. From what I'm told, OS X includes BIOS updates in Software Update, and flashes the BIOS when you reboot -- I've never seen it, but maybe I missed it. That'd be the second best way -- force a reboot, but at least this way you don't have to repartition. Third way would be loading from a partition on the hard disk -- just shrink your FS by a bit, make a tiny FAT partition, then type in the pathname, or have it check in standard places. Your solution is a distant fourth, and in fact, I believe my last computer did this, from a floppy. Some of the newer laptops do it from a USB keychain.
I would still much rather have the system be open enough that we didn't have to beg for things like these. Just use LinuxBIOS and have a standard way of flashing it -- then you can flash it from anything Linux can read, including network, NTFS partitions, the works.
FreeDOS+qemu. Then the only thing that costs money is the app, if you care.