Gosh this all sounds pretty sensible doesn't it? Well, yes it does. It's not often I support anything from Mr John (slightly to the right of Ghengis Khan) Howard, but this seems pretty reasonable. (I still won't vote for him though).
Can't say I'm a huge fan of Howard (although he does fit into my "least worst option" category), but he's not even close to being "far right". Indeed, most of our American friends would probably consider him centre-Left (although that's somewhat skewed by the whole US system being biased a long way Right in the first place).
Howard, like Rudd, is pretty much centrist (ie: populist). Which is why it would be quite difficult to pick between them, if it wasn't for that Communist wench standing just slightly behind ol' Kevin.
So, she was still referring to "a gigabyte of power" like she was on the 7:30 Report a few hours earlier, was she?
She actually said "a gigabit" - and while the terminology is grating to people with Clues, what she actually meant was perfectly clear in context (for those who didn't - or couldn't - watch, a gigabit of bandwidth ("power") [into the home]).
However, people with such a poor grasp of the technology shouldn't be in charge of it. While I can excuse Howard for clearly not having the foggiest clue what the bloke meant when he was talking about "spectrum", for the Minister of Telecommunications, etc, not to know the terminology (or to get so flustered as to bollocks it up) is ridiculous.
>i>The $600 price is with no discount for signing a contract, that is full retail price. Apple has stated there will be no discount for a contract - it's just that they won't sell the phone without a contract at any price.
Ie: "that's the discounted price for going on a contract, but we won't call it that because customers hate the idea of contracts".
(It's exactly the same as the "full price, full version" OS X that's actually an upgrade (because you can only run it on an existing Mac).)
It can be detected, but not trivially if the cheating is subtle. Coding the cheat to occur randomly and only if a certain period of time has elapsed since the previous vote was entered would make it very difficult to detect. In most elections in the US the difference between the leading candidates is usually just a couple of percentage points so randomly changing one vote in twenty cast at least two minutes apart would make the testing of millions of simulated votes difficult.
The testing scenario is not difficult. Certainly, it requires planning, but in and of itself, it is not difficult, because the scope is so tiny and the variables so few. Indeed, the scope is so small that the use of a formally verifiable specification and implementation (to prove the code is bug free) is more than feasible, making detection of outside influences even easier. An absolute code freeze (for all components - hardware and software) a fixed time period prior to the election (say, 90 days - IIRC US elections run on a fixed schedule, making this part even easier). Scripted testing of the system, with suitable delays between individual votes, running on multiple parallel systems and simulating multiple locations and voter demographics. Always have the machine's date set to the election date. Only run the tests during the hours voting would actually be happening. Etc, etc.
There's only a (very) limited number of variables that a vote-rigging system can key off to determine whether or not the "election" is real. Account for them and you can be confident that if no discrepencies show up in the testing, then even if they (undetectably) exist they're also too subtle to matter in the actual election.
That, and, of course, verification via the paper trail should be a requirement.
To look at it another way: Yes, it's feasible that vote-rigging could be introduced via the OS. But no more feasible that the same vote-rigging being introduced in any of the myriad pieces of hardware involved (and there would be a *lot* of hardware components involved) or even in an open-source solution using that backdoor-in-the-compiler trick.
Again, as I said, with a very limited set if possible inputs and outputs (like you would have, say, in an election), it is trivial to determine if the environment is a factor.
Set the machine date appropriately (in fact, run the test for every day in the year), feed in a million votes, check the output for discrepencies.
Indeed, I would expect such basic functional verification to be part of the development process - to catch bugs - even _before_ the stupidly remote possibility of vote-rigging-via-OS was considered.
With that said, a paper trail is essential in any electronic voting system. There's simply too many ways inconsistent results can be introduced, by both fair means and foul (though mostly foul).
Personally, I believe that if Microsoft wants to continue to enjoy their current status as the leading provider of security-challenged applications and operating systems, they should be required to provide free updates. Either that, or Microsoft should have to reimburse everyone else for the bandwidth costs incurred by the hundreds of millions of infected Windows installations worldwide. For that matter, how about all the millions of man-hours lost because of instabilities in Microsoft's products. Heck, I'd say providing Windows updates is the least that company can do, given the grief they've caused so many others.
Why should Microsoft be paying for problems that are, by and large, caused by either their customers or third parties ?
I've often wondered what would have happened if Windows had never been, if some other OS (say a Unix variant of some kind, or perhaps one of the other OSes that Microsoft eclipsed) had become dominant, one that was fundamentally more secure than Windows.
Like what ? UNIX (until relatively recently - SELinux, et al) was fundamentally *less* secure than Windows NT. BeOS, OS/2, Classic MacOS, AmigaOS and the like were single-user (heck, AmigaOS didn't even have protected memory) and didn't even _have_ security. What "other OSes" were you thinking of ?
We might never have seen the billions of dollars being lost to spam and armies of rooted Windows boxes.
Indeed. Instead it would have been billions of dollars lost to spam and armies of root $OTHER_OS boxes.
The vast majority of "security problems" are the fault of the end user. Harsh, but true.
The issue is that the underlying OS has the ability to modify the results of any application it runs. Say, for example, Gates decides to run for President. Balmer could easily have the input coded to skew the results in Bill's favor in very subtle ways that only a thorough examination of the underlying system code would expose.
Utter tripe. A simple examination and verification of the (limited and known) possible inputs and outputs would provide confirmation of correct operation.
If you have a calculator that's only designed to do a dozen different sums, it's *trivial* to prove whether or not it is doing those sums correctly, or whether some sort of environment effect is affecting the results.
You should keep in mind two things. First, a manual transmission is more fuel-efficient and a pleasure to drive. The only rational reason not to get one is that you don't know how to drive a stick.
Anyone who buys a manual transmission vehicle for a car whose primary purpose is commuting or inner-city drive is crazy.
It's highly unlikely you'll be able to drive with the discipline to achieve better mileage (assuming you can at all - modern autos are extremely good and most people are terrible drivers), your gearbox - especially the second gear synchromesh - will suffer very harsh wear and tear and it's simply more work (and hence more stress) to do.
I say this as someone who grew up driving "stick" (as Americans call it) and derives a great deal of pleasure from a drive along some twisty roads - although I prefer a motorbike these days. I wouldn't buy a manual car for city driving with someone else's money.
You are just hoping. What evidence do you have that business computing requirements will explode? (No, 3D eye candy is not a "requirement".) We should be seeing some hints of any such demand today. If you are betting on some software breakthrough that is far beyond the horizon, then you are really dreaming.
What's this "explode" ? I'm simply observing that over time, the baseline PC specification has risen, and I am predicting it will continue to do so. A decade ago a 486 provided enough power for "business requirements".
Software expands to fill the hardware resources allotted to it.
And, in fact, "3D eye candy" may well become a "requirement" if it forms part of UI improvements that lead to productivity increases. Despite common belief on Slashdot, the vt100 terminal was not the pinnacle of user interface design.
Hah! Point is when you get something from Microsoft they wrote or purchased every scrap of software that goes with it.
So they didn't write all of it, is what you're saying ?
When you use Mac or Linux developers are competing to get their software in the system.
You mean like the various programmers and departments at Microsoft - not to mention the third party code they have bought - do ?
Remember Mac uses quite a bit of OSS too. Essentially, Mac and Linux operate in a free market while everything installed on Windows is affected by some board of directors in Redmond and their decisions are not based on what is best for the user, it is based on what is best for the bottom line, even if it means shipping IE6 for years after it's a dead horse.
Well, that's one of the more... interesting takes on the "free market" I've heard for a while.
If you don't think Apple dictate what goes into OS X in exactly the same way Microsoft to, and Linus and co. do the same for Linux, I have some excellent real estate to sell you.
Corruption is the form of cheating I expect Microsoft to try.
What, exactly, do you expect them to "corrupt" ?
You certainly haven't shown it.
Why ? Because I have come to a different conclusion ?
What you consistently refuse to understand is that technology does not stand still.
On the contrary, it appears I understand it vastly better than you do, since you seem convinced that the computing requirements of today are going to remain static while the "per-$100" specifications of computers rise to meet (or exceed) them.
Not to mention the difficulty you seem to have understanding that Microsoft, also, do not stand still.
Sneer at the OLPC all you want. But $100 computers will not stay flimsy and underpowered forever; soon, their hardware will be everything that 99% of businesses will ever need, and their software will be free.
Firstly, I'm not "sneering" at the OLPC.
Secondly, your conclusion is doubtful to say the least. There's a limit to how cheap you can make a computer, especially one where the buyer does not consider price to be the most important feature.
So, if Windows is only designed for two or four processors, why even consider eight?
Best not to listen to marketing dweebs for technical information. Windows NT ("Vista") is - and always has been - designed from the ground up to work very well with multiple CPUs. It's heavily multithreaded, fully re-entrant, kernel locking is very fine-grained, etc, etc.
I have no idea what this person thinks they're saying, but Windows NT4 was available for machines with 8 CPUs a decade ago and Windows 2000 has been running on 64-CPU machines for years. It's possibly some sort of incredibly poorly communicated misunderstanding about how modern machines are more likely to find multiple cores on a single package, rather than discrete CPUs, but even that would only require scheduler tweaking and certainly nothing "fundamentally different". It may also be a reference to Singularity.
What is clear, is that "Microsoft executive Ty Wilson" has NFI what he's talking about and needs to be whacked with a clue-by-four (and probably was). There's nothing at all wrong with Windows' SMP support, especially in the context of the hardware it typically runs on.
Of course, that's Microsoft... How does OSX and Linux handle eight processors?
OSX, not very well. They've only moved away from a single big kernel lock relatively recently - although Leopard is supposed to have some significant improvements in this area - and there's lots of work that needs to be done. Linux's SMP support is excellent (almost certainly better than Windows') and it's been running on machines with quite large CPU counts for years.
Yes: Microsoft could cheat, and I have little doubt they will try.
Indeed, they could lower prices. The scoundrels !
What you refuse to understand is this: when a normal computer with all the software trimmings costs a hundred dollars, how much do you think Microsoft will make from it? And will that be enough to continue feeding a $50 billion/year monster? Of course not. Microsoft is due for a massive shrink.
I understand it perfectly well.
What *you* don't seem to understand is:
* The $100 PC isn't likely to take over the world. Expand it ? Yes. Replace the significant proportion of machines that sit into the same price brackets as the Mac Mini and the iMac ? No chance.
* Most of Microsoft's income comes from sources other than the entry-level, OEM-home-desktop Windows and Office licenses that might be displaced on a $100 PC.
* Lower per-unit revenue will be at least somewhat mitigated by higher volumes.
* Other revenue sources will likely increase (eg: Xbox, more legitimate purchases in growing economies)
Microsoft are a long, long way from the peril people like you would like to pretend they're in.
In a world where ultra cheap computers are normal, Microsoft will not be getting anywhere near $100 per customer per year, not when excellent, featureful, reliable software is available for free. Conclusion: while Microsoft may never completely disappear, the company is due for a massive shrink.
There's a mighty big elephant in your living room that you're ignoring.
The laws of economics are inexorable: reasonable quality + low price = winner.
What's your point ? As I keep trying to point out, and you keep ignoring, to the vast majority of Microsoft's customers, *their software isn't expensive*.
i>For most of these, I've not found an equivalent in Windows or Linux which is available by default. I know the Command-W has the Ctrl-W equivalent in many Windows apps now, but it would be nice if everything supported it - Visual Studio, I'm looking at you.
There are certainly equivalents to most in Windows (except for things which don't apply to the Windows UI, like Expose). However, ultimately it is up to the application to actually implement them.
I'm not saying OS X complete lacks keyboard shortcuts, I'm saying Windows is vastly more accessible via the keyboard because it has been designed from day 1 to be completely usable without any mouse attached. This lets you access arbitrary functions by what are, eseentially, keyboard shortcuts (although they might require multiple keypresses).
That, I'll grant. I ignored it since it's an idealistic, not technical, issue.
The journaling support was introduced in version 5, which shipped with windows 2000...
NTFS journals metadata and always has. It's important to realise, however, that in the architecture of NTFS the file "data" is itself considered metadata, so the end result is very similar to full block journaling.
Can't speak for posix compatibility tho
NTFS has been fully POSIX compliant since its first relase.
The anti-competitive cases are usually about getting Microsoft to focus on their core functionality, like the security of the operating system, rather than write up stupid little weather bug clones for the desktop. Get M$ out of browser space, out of desktop search, get them to quit trying to own everything the user touches and quit using their monopoly status to ship this crap that snuffs out any market emerging on the desktop.
In the market Microsoft are selling to, browsers, desktop search, email, file managers, etc *are* core functionality.
/car analogy/ If Microsoft was a car company, [...]
Here's a better one. Years ago, Microsoft used to sell basic cars. They had an engine, wheels, brakes, seats, etc, but not much in the way of steroes, ABS brakes, anti-theft devices, performance modifications, etc. But then their customers started asking them to include things like airbags, GPSes, 18" rims, spoilers, leather seats, air conditioning, cruise control and the like. Partly this was because their competitors were already offering these features, but the other reason was because some of those features had become so commonly added as aftermarket modifications, that most people had come to consider them standard.
So now when you buy a Microsoft car, you get all the basics, plus a GPS, fancy stereo, ABS, airbags, 18" rims, spoiler, fart pipe, air conditioning, leather seats and dozens of other little optional extras that you used to have to go out and buy yourself after you'd picked the car up - just like when you buy an Apple car or an Ubuntu car. Why ? Because that's what y'all have told them you want.
No Microsoft does not deserve any kind of forgiveness for shipping crap, no business does.
They don't ship "crap", by objective, rational and sane measurements (unless you genuinely think most software is "crap").
Slashdot Mirror
Gosh this all sounds pretty sensible doesn't it? Well, yes it does. It's not often I support anything from Mr John (slightly to the right of Ghengis Khan) Howard, but this seems pretty reasonable. (I still won't vote for him though).
Can't say I'm a huge fan of Howard (although he does fit into my "least worst option" category), but he's not even close to being "far right". Indeed, most of our American friends would probably consider him centre-Left (although that's somewhat skewed by the whole US system being biased a long way Right in the first place).
Howard, like Rudd, is pretty much centrist (ie: populist). Which is why it would be quite difficult to pick between them, if it wasn't for that Communist wench standing just slightly behind ol' Kevin.
So, she was still referring to "a gigabyte of power" like she was on the 7:30 Report a few hours earlier, was she?
She actually said "a gigabit" - and while the terminology is grating to people with Clues, what she actually meant was perfectly clear in context (for those who didn't - or couldn't - watch, a gigabit of bandwidth ("power") [into the home]).
However, people with such a poor grasp of the technology shouldn't be in charge of it. While I can excuse Howard for clearly not having the foggiest clue what the bloke meant when he was talking about "spectrum", for the Minister of Telecommunications, etc, not to know the terminology (or to get so flustered as to bollocks it up) is ridiculous.
The scary thing is, compared to the biggest luddite in history she's a towering intellect regarding the Intarwebs.
>i>The $600 price is with no discount for signing a contract, that is full retail price. Apple has stated there will be no discount for a contract - it's just that they won't sell the phone without a contract at any price.
Ie: "that's the discounted price for going on a contract, but we won't call it that because customers hate the idea of contracts".
(It's exactly the same as the "full price, full version" OS X that's actually an upgrade (because you can only run it on an existing Mac).)
I'd just like to say, that is an excellent "in a nutshell" explanation of SELinux.
It can be detected, but not trivially if the cheating is subtle. Coding the cheat to occur randomly and only if a certain period of time has elapsed since the previous vote was entered would make it very difficult to detect. In most elections in the US the difference between the leading candidates is usually just a couple of percentage points so randomly changing one vote in twenty cast at least two minutes apart would make the testing of millions of simulated votes difficult.
The testing scenario is not difficult. Certainly, it requires planning, but in and of itself, it is not difficult, because the scope is so tiny and the variables so few. Indeed, the scope is so small that the use of a formally verifiable specification and implementation (to prove the code is bug free) is more than feasible, making detection of outside influences even easier. An absolute code freeze (for all components - hardware and software) a fixed time period prior to the election (say, 90 days - IIRC US elections run on a fixed schedule, making this part even easier). Scripted testing of the system, with suitable delays between individual votes, running on multiple parallel systems and simulating multiple locations and voter demographics. Always have the machine's date set to the election date. Only run the tests during the hours voting would actually be happening. Etc, etc.
There's only a (very) limited number of variables that a vote-rigging system can key off to determine whether or not the "election" is real. Account for them and you can be confident that if no discrepencies show up in the testing, then even if they (undetectably) exist they're also too subtle to matter in the actual election.
That, and, of course, verification via the paper trail should be a requirement.
To look at it another way: Yes, it's feasible that vote-rigging could be introduced via the OS. But no more feasible that the same vote-rigging being introduced in any of the myriad pieces of hardware involved (and there would be a *lot* of hardware components involved) or even in an open-source solution using that backdoor-in-the-compiler trick.
if(app == voting_app) ..if(date == second_tues_in_november) ....if(vote_time > too_soon) ......if(selected!=our_guy) ........if(random_cheat(cheat_factor)) ..........selected=our_guy
C:\>date %ELECTION_DATE%
C:\>voting_app.exe
Again, as I said, with a very limited set if possible inputs and outputs (like you would have, say, in an election), it is trivial to determine if the environment is a factor.
Set the machine date appropriately (in fact, run the test for every day in the year), feed in a million votes, check the output for discrepencies.
Indeed, I would expect such basic functional verification to be part of the development process - to catch bugs - even _before_ the stupidly remote possibility of vote-rigging-via-OS was considered.
With that said, a paper trail is essential in any electronic voting system. There's simply too many ways inconsistent results can be introduced, by both fair means and foul (though mostly foul).
Personally, I believe that if Microsoft wants to continue to enjoy their current status as the leading provider of security-challenged applications and operating systems, they should be required to provide free updates. Either that, or Microsoft should have to reimburse everyone else for the bandwidth costs incurred by the hundreds of millions of infected Windows installations worldwide. For that matter, how about all the millions of man-hours lost because of instabilities in Microsoft's products. Heck, I'd say providing Windows updates is the least that company can do, given the grief they've caused so many others.
Why should Microsoft be paying for problems that are, by and large, caused by either their customers or third parties ?
I've often wondered what would have happened if Windows had never been, if some other OS (say a Unix variant of some kind, or perhaps one of the other OSes that Microsoft eclipsed) had become dominant, one that was fundamentally more secure than Windows.
Like what ? UNIX (until relatively recently - SELinux, et al) was fundamentally *less* secure than Windows NT. BeOS, OS/2, Classic MacOS, AmigaOS and the like were single-user (heck, AmigaOS didn't even have protected memory) and didn't even _have_ security. What "other OSes" were you thinking of ?
We might never have seen the billions of dollars being lost to spam and armies of rooted Windows boxes.
Indeed. Instead it would have been billions of dollars lost to spam and armies of root $OTHER_OS boxes.
The vast majority of "security problems" are the fault of the end user. Harsh, but true.
So why do you need schwindoze (or schlinux) to do all those basic things????
Because it's a lot cheaper to buy (or download) an off-the-shelf OS than get someone to write the code you are talking about.
The issue is that the underlying OS has the ability to modify the results of any application it runs. Say, for example, Gates decides to run for President. Balmer could easily have the input coded to skew the results in Bill's favor in very subtle ways that only a thorough examination of the underlying system code would expose.
Utter tripe. A simple examination and verification of the (limited and known) possible inputs and outputs would provide confirmation of correct operation.
If you have a calculator that's only designed to do a dozen different sums, it's *trivial* to prove whether or not it is doing those sums correctly, or whether some sort of environment effect is affecting the results.
This blogger did nothing wrong according to OUR CONSTITUTION.
"Freedom of speech" is not synonymous with "freedom from consequence".
You should keep in mind two things. First, a manual transmission is more fuel-efficient and a pleasure to drive. The only rational reason not to get one is that you don't know how to drive a stick.
Anyone who buys a manual transmission vehicle for a car whose primary purpose is commuting or inner-city drive is crazy.
It's highly unlikely you'll be able to drive with the discipline to achieve better mileage (assuming you can at all - modern autos are extremely good and most people are terrible drivers), your gearbox - especially the second gear synchromesh - will suffer very harsh wear and tear and it's simply more work (and hence more stress) to do.
I say this as someone who grew up driving "stick" (as Americans call it) and derives a great deal of pleasure from a drive along some twisty roads - although I prefer a motorbike these days. I wouldn't buy a manual car for city driving with someone else's money.
You are just hoping. What evidence do you have that business computing requirements will explode? (No, 3D eye candy is not a "requirement".) We should be seeing some hints of any such demand today. If you are betting on some software breakthrough that is far beyond the horizon, then you are really dreaming.
What's this "explode" ? I'm simply observing that over time, the baseline PC specification has risen, and I am predicting it will continue to do so. A decade ago a 486 provided enough power for "business requirements".
Software expands to fill the hardware resources allotted to it.
And, in fact, "3D eye candy" may well become a "requirement" if it forms part of UI improvements that lead to productivity increases. Despite common belief on Slashdot, the vt100 terminal was not the pinnacle of user interface design.
Hah! Point is when you get something from Microsoft they wrote or purchased every scrap of software that goes with it.
So they didn't write all of it, is what you're saying ?
When you use Mac or Linux developers are competing to get their software in the system.
You mean like the various programmers and departments at Microsoft - not to mention the third party code they have bought - do ?
Remember Mac uses quite a bit of OSS too. Essentially, Mac and Linux operate in a free market while everything installed on Windows is affected by some board of directors in Redmond and their decisions are not based on what is best for the user, it is based on what is best for the bottom line, even if it means shipping IE6 for years after it's a dead horse.
Well, that's one of the more... interesting takes on the "free market" I've heard for a while.
If you don't think Apple dictate what goes into OS X in exactly the same way Microsoft to, and Linus and co. do the same for Linux, I have some excellent real estate to sell you.
Corruption is the form of cheating I expect Microsoft to try.
What, exactly, do you expect them to "corrupt" ?
You certainly haven't shown it.
Why ? Because I have come to a different conclusion ?
What you consistently refuse to understand is that technology does not stand still.
On the contrary, it appears I understand it vastly better than you do, since you seem convinced that the computing requirements of today are going to remain static while the "per-$100" specifications of computers rise to meet (or exceed) them.
Not to mention the difficulty you seem to have understanding that Microsoft, also, do not stand still.
Sneer at the OLPC all you want. But $100 computers will not stay flimsy and underpowered forever; soon, their hardware will be everything that 99% of businesses will ever need, and their software will be free.
Firstly, I'm not "sneering" at the OLPC.
Secondly, your conclusion is doubtful to say the least. There's a limit to how cheap you can make a computer, especially one where the buyer does not consider price to be the most important feature.
So you're a moral relativist?
Everyone is a moral relativist. Some just aren't prepared to admit it for fear of breaking their neck falling off such a high horse.
So, if Windows is only designed for two or four processors, why even consider eight?
Best not to listen to marketing dweebs for technical information. Windows NT ("Vista") is - and always has been - designed from the ground up to work very well with multiple CPUs. It's heavily multithreaded, fully re-entrant, kernel locking is very fine-grained, etc, etc.
I have no idea what this person thinks they're saying, but Windows NT4 was available for machines with 8 CPUs a decade ago and Windows 2000 has been running on 64-CPU machines for years. It's possibly some sort of incredibly poorly communicated misunderstanding about how modern machines are more likely to find multiple cores on a single package, rather than discrete CPUs, but even that would only require scheduler tweaking and certainly nothing "fundamentally different". It may also be a reference to Singularity.
What is clear, is that "Microsoft executive Ty Wilson" has NFI what he's talking about and needs to be whacked with a clue-by-four (and probably was). There's nothing at all wrong with Windows' SMP support, especially in the context of the hardware it typically runs on.
Of course, that's Microsoft... How does OSX and Linux handle eight processors?
OSX, not very well. They've only moved away from a single big kernel lock relatively recently - although Leopard is supposed to have some significant improvements in this area - and there's lots of work that needs to be done. Linux's SMP support is excellent (almost certainly better than Windows') and it's been running on machines with quite large CPU counts for years.
All 8 cores and 16 Gigs of RAM fully accessible by the OS, unlike say Win XP.
Windows XP 64-bit will "access" your 8 cores and 16 gigs of RAM just fine, and to boot will do a better job of utilising them than OS X does.
Yes: Microsoft could cheat, and I have little doubt they will try.
Indeed, they could lower prices. The scoundrels !
What you refuse to understand is this: when a normal computer with all the software trimmings costs a hundred dollars, how much do you think Microsoft will make from it? And will that be enough to continue feeding a $50 billion/year monster? Of course not. Microsoft is due for a massive shrink.
I understand it perfectly well.
What *you* don't seem to understand is:
* The $100 PC isn't likely to take over the world. Expand it ? Yes. Replace the significant proportion of machines that sit into the same price brackets as the Mac Mini and the iMac ? No chance.
* Most of Microsoft's income comes from sources other than the entry-level, OEM-home-desktop Windows and Office licenses that might be displaced on a $100 PC.
* Lower per-unit revenue will be at least somewhat mitigated by higher volumes.
* Other revenue sources will likely increase (eg: Xbox, more legitimate purchases in growing economies)
Microsoft are a long, long way from the peril people like you would like to pretend they're in.
In a world where ultra cheap computers are normal, Microsoft will not be getting anywhere near $100 per customer per year, not when excellent, featureful, reliable software is available for free. Conclusion: while Microsoft may never completely disappear, the company is due for a massive shrink.
There's a mighty big elephant in your living room that you're ignoring.
The laws of economics are inexorable: reasonable quality + low price = winner.
What's your point ? As I keep trying to point out, and you keep ignoring, to the vast majority of Microsoft's customers, *their software isn't expensive*.
i>For most of these, I've not found an equivalent in Windows or Linux which is available by default. I know the Command-W has the Ctrl-W equivalent in many Windows apps now, but it would be nice if everything supported it - Visual Studio, I'm looking at you.
There are certainly equivalents to most in Windows (except for things which don't apply to the Windows UI, like Expose). However, ultimately it is up to the application to actually implement them.
I'm not saying OS X complete lacks keyboard shortcuts, I'm saying Windows is vastly more accessible via the keyboard because it has been designed from day 1 to be completely usable without any mouse attached. This lets you access arbitrary functions by what are, eseentially, keyboard shortcuts (although they might require multiple keypresses).
I've read that three times now, and I'm prepared to admit defeat. What the hell are you raving about ?
NTFS is not open...
That, I'll grant. I ignored it since it's an idealistic, not technical, issue.
The journaling support was introduced in version 5, which shipped with windows 2000...
NTFS journals metadata and always has. It's important to realise, however, that in the architecture of NTFS the file "data" is itself considered metadata, so the end result is very similar to full block journaling.
Can't speak for posix compatibility tho
NTFS has been fully POSIX compliant since its first relase.
NTFS has only journaling on the file system level, not on the file level.
Functionally, it ssentially does, since a file's "data" is, by the architecture, considered metadata, and NTFS journals metadata.
It's not a transactional system on each file as what is understood under journaling these days.
Most "journaling filesystems" - especially in their default configuration - only journal metadata.
The anti-competitive cases are usually about getting Microsoft to focus on their core functionality, like the security of the operating system, rather than write up stupid little weather bug clones for the desktop. Get M$ out of browser space, out of desktop search, get them to quit trying to own everything the user touches and quit using their monopoly status to ship this crap that snuffs out any market emerging on the desktop.
In the market Microsoft are selling to, browsers, desktop search, email, file managers, etc *are* core functionality.
Here's a better one. Years ago, Microsoft used to sell basic cars. They had an engine, wheels, brakes, seats, etc, but not much in the way of steroes, ABS brakes, anti-theft devices, performance modifications, etc. But then their customers started asking them to include things like airbags, GPSes, 18" rims, spoilers, leather seats, air conditioning, cruise control and the like. Partly this was because their competitors were already offering these features, but the other reason was because some of those features had become so commonly added as aftermarket modifications, that most people had come to consider them standard.
So now when you buy a Microsoft car, you get all the basics, plus a GPS, fancy stereo, ABS, airbags, 18" rims, spoiler, fart pipe, air conditioning, leather seats and dozens of other little optional extras that you used to have to go out and buy yourself after you'd picked the car up - just like when you buy an Apple car or an Ubuntu car. Why ? Because that's what y'all have told them you want.
No Microsoft does not deserve any kind of forgiveness for shipping crap, no business does.
They don't ship "crap", by objective, rational and sane measurements (unless you genuinely think most software is "crap").
If you do compare them Windows look rather poor.
By what metric ? How are you normalising that metric to account for different market shares, user demographics and system capabilities ?