Potentially. Japan, unlike China, never had the problem of 90%+ of their population being completely uneducated and living in 3rd world hellhole, though. Yes, they had to recover from WWII, but they did it quickly, as a national effort. And in a much smaller area than China.
Not saying China can't do it, or won't do it. They just have a lot more hurdles to jump. India, more still.
Sometimes, and usually, yes. But cell phones - any particular model - is usually not produced in very high numbers. (For instance, I'm pretty much in a Verizon-only area. I have not once seen my current phone used by someone else, and only once have I seen any of my old phones. There are a lot of variants out there.
They could easily, easily cut the cost of production by simply producing a higher number of a better phone, and putting pretty color face plates on it (and being Verizon, locking down 90% of the functionality in software).
A mobile phone with a large screen? That'd be awesome! Even the shitty mobile phones from Verizon (the free ones, likely not costing much more than $50 to make, I suppose) likely have chips operating around 200MHz, either ARM or MIPS. Probably faster/newer than that (ie multicore chips w/ video decoding), if it's going to have a non-CLI interface. I'd say at least 200MHz, unless their WiFi implementation doesn't require any CPU utilization. They'll also likely have at least 32Mb RAM due to the requirements of such things, and additional memory overhead, and I'd guess 64Mb+ CF for storage. We're basically talking about the rough specifications of something like a circa-2000 HPC - something like the NEC Mobilepro 780, or a similar era Jornada - but with last-generation CPU architecture (which has improved substantially - if you think x86 has improved a lot since 2000, take a look at ARM!)
High-end phones would be more than capable of handling 100% of most people's computing needs - in the US - as they clock at 600Mhz+ with specialized chips for a/v decoding. Their biggest limitations in doing so, to date, are a lack of host USB/ubiquitous and functional BT input devices, and the common desktop software most people tend to use (a full-featured web browser and flash plugins, basically).
And yes, I think we can expect these computers to have at least these specs, provided the device isn't complete vaporware/not coming out at $10. Mass production can cut a LOT off the cost of a device. Hell, you can get some pretty nice, and likely much more functional, "handheld kits" from a small Israeli company for ~$200 each for a lot of 1000 (with your own custom board options). I imagine the state government of India can reduce fab costs a lot more than Israel can.
OK; worst case scenario, it's meant "public, sold beta". However, most of these applications are at least usable. KDE 4.0 wasn't even feature-complete to the point of being usable, never mind having half the functionality of the previous version.
Off the top of my head, usable #.0 releases are:
OpenOffice.org XFCE GIMP WINE Firefox Thunderbird Paint Shop Pro Samba Apache MySQL Pidgin/gAIM Xchat Apple OS X 10.0 (no experience with prior versions) or for that matter, all 10.x.# releases (as it's roughly synonymous)
And those are just the applications I'm familiar with/use with some degree of regularity. I'm sure there are many others.
It's one thing to not reach a point release and call it stable (such as calling version 0.9 a stable release, or what have you) or to never making the distinction/calling everything a development release (see: Enlightenment); it's another thing entirely to positively mislead people to thinking
x.0 releases have always meant "this software is ready for general user consumption". Whether it's stable or not, in absolute terms, is mostly inconsequential; the fact is that it's considered "stable" by the developers. To advertise the major point release and not denote it as alpha-quality equivalent is irresponsible and reckless.
Sort of like getting on an interstate with a "this road requires you to drive on the left side" sign on the side of the road, in small letters, behind a bush. It's entirely against long-lived convention (at least in the US), goes against common sense, and is dangerous if not foolish.
Anyone who's used a computer for more than a week knows that "point release means it's the new stable release", or at least reasonably close to one. If they intended it to be otherwise, it should have been BETA (or some other versioning scheme, like what the Linux kernel used to use back when you could reasonably download a kernel and have every module included work).
Hell, even Microsoft did this with W7. It went API stable, and then they released a beta. it was very obviously a beta, because they're calling it that.
Meh, it might be a collaborative one, but to me that just sounds like marketing speak. From where I sit, Expose is a bolted-on feature to make up for existential design flaws (and lack of customization ability) in the user interface. It is an inprecise and somewhat awkward, albeit pretty, way to relate to one's computer.
As far as being a sum of parts... Expose was an afterthought when they realized the Dock was fundamentally flawed and limiting. Sure, it's integrated and 'collaborative' now, but just barely.
No telling when it'll be fixed; I looked about for half an hour or so the other day to try to find current efforts to fix it, but could find nothing definitive. There have been numerous bug reports which seem to all come back to the same I/O problem, so I can't provide a specific/definitive one, but here are two as they pertain to Ubuntu:
There was a frontpage post on slashdot about the bug being discovered about a week or so ago as well, FWIW.
I did update Ubuntu (8.10) shortly ago, and while there was no kernel upgrade in the update in the apt-get, I/O performance does appear to be marginally improved after the reboot today (though that might simply be X behaving better after a restart).
You could pick up Cuckoo's Egg by Cliff Stroll. It's a real-to-life first-hand account of the Morris Worm. It's well-written and entertaining - one of the few novels I've actually bought after getting out of the library (I now have two copies).
Just as the people who were exploited are responsible for having been exploited. Their own damn fault.
No, that is not a contradiction. You'd get fired if you, as a security professional, were responsible for the network and it got taken down, would you not? Same kind of thing with those he exploited. Responsibility is responsibility, regardless of scale.
(Note, legal fault/responsibility is different than personal responsibility, obviously.)
That depends on whether you consider the fault on her part having been raped, or failing to try to prevent it, I suppose.
I wouldn't say it's entirely her fault for being raped, but it is her fault for not taking the appropriate precautions to inhibit or prevent the rape. In so far as she does not take precautions, she is (at least) liable for the rape (to the same degree that someone wearing fishnets down a slum alley after dark would be, but to a lesser degree). No, nobody ever "asked for it" - that's the extreme, and so far off on right field that it holds no validity. However, that does not diminish the fact that it hits upon a sentiment (albeit, entirely too strongly) which is appropriate.
In a just world with self-aware, prudent women, that would mean that said potential-rape victim would go about armed so as to inhibit and discourage a rape - and such behavior would be not only acceptable but expected. Much in the same way that it should be acceptable and expected for a person to keep their operating system and software up to date.
Ignorance and naivety have never been excusable traits. Forgivable, certainly.
Regardless of whether or not one admires botmasters' motives (and I don't) crafting botnets on a large scale has a certain "cool" factor, since there is quite a lot of work, skill and even artistry involved in setting them up.
Absolutely. Just for shits and giggles, a friend and I crafted an autonomous botnet engine, where each node could (and did) serve as a limited master, automatically negotiating master in a similar fashion to how human group psychology would dictate. It has a very low network signature due to how it communicates and infects, and the primary 'design goal' was to make a worm which could propagate (ie no payload).
We never released it, but we did it because it was fun. It was like playing chess, in a sense.
Honestly, if you want to provide your grandchildren a picture of the past - antiques and things they aren't likely to see in the future - you're best bet would be to seal up something like a dictionary and/or a firearm. Those items will either be politically/socially passe by the time they're in school, or all but non-existent.
And, worst case senario, they'd be able to utilize the firearm to protect themselves.
There is nothing you could do to make the old hardware functional in 50 years. You will run into a slew of problems.
The first problem you will run into is the hard drives not working in a scant decade. They are not designed to last that long and, particularly after years of use, won't. The polymer lubricants used to keep the things turning will be dry and the drive will either not spin up or spin up and quickly self-destruction.
The second problem is data longevity. Bit rot will set in fairly quickly. While I have no direct evidence as to how quickly, I guess it'll happen within the first 10 years to the point where the data is inaccessible/irretrievable. (Think: a lot of software only 10 years old is no longer accessible; you need to continually bit-refresh medium, or replace it outright.)
I suspect even the best CDR media will not last 20 years. Tape might, but you'd run into bitrot with that, as well. CD pressed media might do the trick.
I'd say your best bet would be to provide media with emulation software. In 50 years, the entire paradigm of computing is likely to have changed. Think, for a moment, about the computing devices your parents had available to them in the 1960s. That's a lot of change, and we're still undergoing said change. A ~15-year standard like VGA/d-sub connectors or USB isn't likely to mean anything by then - and what if a USB mouse or monitor doesn't work, while the rest of the system does? The system as a whole provides a further problem, because a failure of anything else on the system is likely to result in a fairly obvious
Don't put a computer in a school time capsule unless you expect it to merely be a curiosity. Even the 5.25" floppies and Apple II system put in the ground when I was in 2nd grade (20 years ago) is now likely inoperable and poses nothing much more than a curiosity.
The likely cause of the I/O slowdown you describe in Ubuntu is likely due to a well-known kernel bug which has been present since at least 2.6.18. If you were cutting your teeth with Ubuntu 7.10 chances are you don't know better, but Linux never used to do nonsense like this. I've run run right up to the ceiling of RAM use (and dipping into swap, even) while doing things like compiling and the UI is still responsive.
That said, X does tend to get a bit sluggish after a couple weeks. It's gotten better recently, however.
Yes, but at least the ability to do so is present in Windows (which I do not personally use, I should note). This isn't the case with MacOS X.
The problem is that the "mac way" has never worked at a conceptual level for people who are more computer oriented. I don't see "windows as documents" so the application-centric model of MacOS is fundamentally broken; I see windows (more or less) as a process managed by another process to allow me to manage them more clearly. The Apple Way is an additional abstraction, making the user go through additional hoops.
RAM is cheap. It's like $30 a gigabyte or less, and systems can take 8 to 32 of said gigabytes these days. That is more than enough RAM to handle strap-on apps to enhance OS functionality. Unfortunately, OS X pretty much inhibits fixing things in such a manner due to its fundamental design.
You can have an "application centric" model in Windows, if you want (task grouping). It's very similar. The same can not be said of OS X and a window centric model. You are irrevocably shackled to the magic application window which continually interferes with input focus and window switching.
Yet, Spaces does not fix the fundamental flaw: an application centric focus and task management model.
There are still a slew of commonly-performed tasks which are impossible due to the anachronistic app-centric model, all because pre-OSX, MacOS did not suitably perform multiprocess functionality. THAT is the real excuse behind the "Application centric" focus in Apple's operating systems.
I understand the difference, but it still makes working with two separate applications very, very irritating. Just try opening a PDF as a reference point while you're trying to write a document in NeoOffice... you'll pull your hair out.
Expose is very nice and, IMO, it's a necessary feature due to the fact that the Dock is so crippling when it comes to window switching. OS X would be almost unusable for me if Expose wasn't there.
The simple fact is that people do not usually think in terms of applications. They think "I need to work on x, y, and z documents" or "I want to open my resume" or "I want to listen to music while I surf the web". This precludes an "application centric" type focus, and is much closer to the "window centric" method of windowing.
Expose is necessary because the Dock does not scale. Yes, it is very useful within a limited context - ie, there are only a half dozen windows/applications open with a document or two each. Once you start opening 3 or so documents, or need to use two applications side-by-side/concurrently, the Dock (and the Application centric method in general) is a hindrance more than a feature.
Unfortunately, Expose does not scale that well either - at least not as well as a window-centric taskbar. It reaches its limitation at around, I'd say, 8 or so windows, necessitating the use of Spaces instead of just making it a "nice feature to have".
I should note that from my limited exposure, W7 is likely going to be a bit more of a PITA than XP has been, in terms of dealing with large numbers of applications/windows.
What do you mean, this isn't a big deal in the USA job market? Are you saying it's not a big deal to be laid off in the US, or that it's not a big deal because Americans don't have families? Or something else?
As an IT worker who has been laid off twice in the last year due to downturns who has a family, let me tell you, it is a fucking big deal. HR types still very much have the mentality of "oh you were laid off? There must be something wrong with you" and won't even consider someone who's got multiple short stints of employment in a row.
Microsoft is most certainly improving its image with Windows 7. They appear to be getting a lot of things right. They've improved system latency due to I/O over what was present in even XP, and the system is surprisingly stable (for a beta, of course).
Couple this with the fact that the Linux I/O scheduler appears to have moved away from a model which works well on the Linux desktop. For about the last year or so, Linux kernels have resulted in very latent desktop utility during even moderate burst-type I/O (programs/files loading, access of swap - not prologued disk writes). This may or may not be related to the bug supposedly introduced into the kernel in 2.6.18 - I don't know, I haven't personally tested it. But what I do know is that this behavior has become progressively more evident over the past 8 years: I blame the server-centric development focus in the kernel (2.2 and prior were blindingly responsive on the desktop).
With the fact that Linux desktop performance is somewhat lackluster these days giving it a perceived performance more on par with what Vista is capable of, I can see how it would sour people in preference for Windows 7, when Windows 7 appears to implement things properly - or, at least in a way which works to user expectations.
I should note that I've been personally using Linux (mostly Debian, some Ubuntu and OpenSuse) almost exclusively since around 2000. I don't make these criticisms lightly, and personally say it more as an admonishment of the Linux developers/community than I do as a proponent of W7. Whether it's a good product or not, I can not ethically approve of vendor lock in to the extent that MS software use encourages.
(Side note: has anyone noticed how W7's window effects/widgets (to the exception of the "MS-specific blurry/imperfect glass semi-transparent menus) looks shockingly like the bastard child of KDE 4 and OS X 10.5? I thought the first W7 screenshot I saw actually was KDE4 with a 'lookalike' theme.)
I think the biggest thing is that other sources of information are becoming more available - ie, digital sources. Five, 10 years ago, there wasn't much for online reference material on emerging tech (and the breadth of emerging tech was smaller), yet this was still the case.
Keep an eye on traditional media - printed and multimedia both. Especially printed publishing: they're all going to take a hard hit in this recession. Hopefully, it will mean a little respite in the amount of junk mail in the mailbox and the number of high-gloss printed fliers and such we have to put up with now.
Slashdot Mirror
Potentially. Japan, unlike China, never had the problem of 90%+ of their population being completely uneducated and living in 3rd world hellhole, though. Yes, they had to recover from WWII, but they did it quickly, as a national effort. And in a much smaller area than China.
Not saying China can't do it, or won't do it. They just have a lot more hurdles to jump. India, more still.
Sometimes, and usually, yes. But cell phones - any particular model - is usually not produced in very high numbers. (For instance, I'm pretty much in a Verizon-only area. I have not once seen my current phone used by someone else, and only once have I seen any of my old phones. There are a lot of variants out there.
They could easily, easily cut the cost of production by simply producing a higher number of a better phone, and putting pretty color face plates on it (and being Verizon, locking down 90% of the functionality in software).
I imagine you could expect it to at least be 256 gradient monochrome at a fairly high resolution - say, 640x320 or such. Maybe even 256 color.
A mobile phone with a large screen? That'd be awesome! Even the shitty mobile phones from Verizon (the free ones, likely not costing much more than $50 to make, I suppose) likely have chips operating around 200MHz, either ARM or MIPS. Probably faster/newer than that (ie multicore chips w/ video decoding), if it's going to have a non-CLI interface. I'd say at least 200MHz, unless their WiFi implementation doesn't require any CPU utilization. They'll also likely have at least 32Mb RAM due to the requirements of such things, and additional memory overhead, and I'd guess 64Mb+ CF for storage. We're basically talking about the rough specifications of something like a circa-2000 HPC - something like the NEC Mobilepro 780, or a similar era Jornada - but with last-generation CPU architecture (which has improved substantially - if you think x86 has improved a lot since 2000, take a look at ARM!)
High-end phones would be more than capable of handling 100% of most people's computing needs - in the US - as they clock at 600Mhz+ with specialized chips for a/v decoding. Their biggest limitations in doing so, to date, are a lack of host USB/ubiquitous and functional BT input devices, and the common desktop software most people tend to use (a full-featured web browser and flash plugins, basically).
And yes, I think we can expect these computers to have at least these specs, provided the device isn't complete vaporware/not coming out at $10. Mass production can cut a LOT off the cost of a device. Hell, you can get some pretty nice, and likely much more functional, "handheld kits" from a small Israeli company for ~$200 each for a lot of 1000 (with your own custom board options). I imagine the state government of India can reduce fab costs a lot more than Israel can.
OK; worst case scenario, it's meant "public, sold beta". However, most of these applications are at least usable. KDE 4.0 wasn't even feature-complete to the point of being usable, never mind having half the functionality of the previous version.
Off the top of my head, usable #.0 releases are:
OpenOffice.org
XFCE
GIMP
WINE
Firefox
Thunderbird
Paint Shop Pro
Samba
Apache
MySQL
Pidgin/gAIM
Xchat
Apple OS X 10.0 (no experience with prior versions) or for that matter, all 10.x.# releases (as it's roughly synonymous)
And those are just the applications I'm familiar with/use with some degree of regularity. I'm sure there are many others.
It's one thing to not reach a point release and call it stable (such as calling version 0.9 a stable release, or what have you) or to never making the distinction/calling everything a development release (see: Enlightenment); it's another thing entirely to positively mislead people to thinking
x.0 releases have always meant "this software is ready for general user consumption". Whether it's stable or not, in absolute terms, is mostly inconsequential; the fact is that it's considered "stable" by the developers. To advertise the major point release and not denote it as alpha-quality equivalent is irresponsible and reckless.
My recollection might be off, but KDE 3.0 was pretty stable. Much more stable than KDE 4.0. I don't recall KDE 2.0, though.
And yet it is (was) largely known that minor version releases with odd numbers for Linux were unstable/development trees:
major: first number
minor: middle number
point: last number
Been a while since I've really looked into it, but my understanding is that it's somewhat more confusing than that, now.
Why would I give away the rest of the story to a potential reader? :P That's like a movie's back-cover review telling you the hero dies.
Got me on the novel bit.
They told me, really?
Sort of like getting on an interstate with a "this road requires you to drive on the left side" sign on the side of the road, in small letters, behind a bush. It's entirely against long-lived convention (at least in the US), goes against common sense, and is dangerous if not foolish.
Anyone who's used a computer for more than a week knows that "point release means it's the new stable release", or at least reasonably close to one. If they intended it to be otherwise, it should have been BETA (or some other versioning scheme, like what the Linux kernel used to use back when you could reasonably download a kernel and have every module included work).
Hell, even Microsoft did this with W7. It went API stable, and then they released a beta. it was very obviously a beta, because they're calling it that.
Meh, it might be a collaborative one, but to me that just sounds like marketing speak. From where I sit, Expose is a bolted-on feature to make up for existential design flaws (and lack of customization ability) in the user interface. It is an inprecise and somewhat awkward, albeit pretty, way to relate to one's computer.
As far as being a sum of parts... Expose was an afterthought when they realized the Dock was fundamentally flawed and limiting. Sure, it's integrated and 'collaborative' now, but just barely.
No telling when it'll be fixed; I looked about for half an hour or so the other day to try to find current efforts to fix it, but could find nothing definitive. There have been numerous bug reports which seem to all come back to the same I/O problem, so I can't provide a specific/definitive one, but here are two as they pertain to Ubuntu:
https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.22/+bug/131094
https://launchpad.net/ubuntu/+bug/1
There was a frontpage post on slashdot about the bug being discovered about a week or so ago as well, FWIW.
I did update Ubuntu (8.10) shortly ago, and while there was no kernel upgrade in the update in the apt-get, I/O performance does appear to be marginally improved after the reboot today (though that might simply be X behaving better after a restart).
You could pick up Cuckoo's Egg by Cliff Stroll. It's a real-to-life first-hand account of the Morris Worm. It's well-written and entertaining - one of the few novels I've actually bought after getting out of the library (I now have two copies).
Oh, he's fully responsible for his own actions.
Just as the people who were exploited are responsible for having been exploited. Their own damn fault.
No, that is not a contradiction. You'd get fired if you, as a security professional, were responsible for the network and it got taken down, would you not? Same kind of thing with those he exploited. Responsibility is responsibility, regardless of scale.
(Note, legal fault/responsibility is different than personal responsibility, obviously.)
That depends on whether you consider the fault on her part having been raped, or failing to try to prevent it, I suppose.
I wouldn't say it's entirely her fault for being raped, but it is her fault for not taking the appropriate precautions to inhibit or prevent the rape. In so far as she does not take precautions, she is (at least) liable for the rape (to the same degree that someone wearing fishnets down a slum alley after dark would be, but to a lesser degree). No, nobody ever "asked for it" - that's the extreme, and so far off on right field that it holds no validity. However, that does not diminish the fact that it hits upon a sentiment (albeit, entirely too strongly) which is appropriate.
In a just world with self-aware, prudent women, that would mean that said potential-rape victim would go about armed so as to inhibit and discourage a rape - and such behavior would be not only acceptable but expected. Much in the same way that it should be acceptable and expected for a person to keep their operating system and software up to date.
Ignorance and naivety have never been excusable traits. Forgivable, certainly.
Regardless of whether or not one admires botmasters' motives (and I don't) crafting botnets on a large scale has a certain "cool" factor, since there is quite a lot of work, skill and even artistry involved in setting them up.
Absolutely. Just for shits and giggles, a friend and I crafted an autonomous botnet engine, where each node could (and did) serve as a limited master, automatically negotiating master in a similar fashion to how human group psychology would dictate. It has a very low network signature due to how it communicates and infects, and the primary 'design goal' was to make a worm which could propagate (ie no payload).
We never released it, but we did it because it was fun. It was like playing chess, in a sense.
Precisely.
Honestly, if you want to provide your grandchildren a picture of the past - antiques and things they aren't likely to see in the future - you're best bet would be to seal up something like a dictionary and/or a firearm. Those items will either be politically/socially passe by the time they're in school, or all but non-existent.
And, worst case senario, they'd be able to utilize the firearm to protect themselves.
There is nothing you could do to make the old hardware functional in 50 years. You will run into a slew of problems.
The first problem you will run into is the hard drives not working in a scant decade. They are not designed to last that long and, particularly after years of use, won't. The polymer lubricants used to keep the things turning will be dry and the drive will either not spin up or spin up and quickly self-destruction.
The second problem is data longevity. Bit rot will set in fairly quickly. While I have no direct evidence as to how quickly, I guess it'll happen within the first 10 years to the point where the data is inaccessible/irretrievable. (Think: a lot of software only 10 years old is no longer accessible; you need to continually bit-refresh medium, or replace it outright.)
I suspect even the best CDR media will not last 20 years. Tape might, but you'd run into bitrot with that, as well. CD pressed media might do the trick.
I'd say your best bet would be to provide media with emulation software. In 50 years, the entire paradigm of computing is likely to have changed.
Think, for a moment, about the computing devices your parents had available to them in the 1960s. That's a lot of change, and we're still undergoing said change. A ~15-year standard like VGA/d-sub connectors or USB isn't likely to mean anything by then - and what if a USB mouse or monitor doesn't work, while the rest of the system does? The system as a whole provides a further problem, because a failure of anything else on the system is likely to result in a fairly obvious
Don't put a computer in a school time capsule unless you expect it to merely be a curiosity. Even the 5.25" floppies and Apple II system put in the ground when I was in 2nd grade (20 years ago) is now likely inoperable and poses nothing much more than a curiosity.
The likely cause of the I/O slowdown you describe in Ubuntu is likely due to a well-known kernel bug which has been present since at least 2.6.18. If you were cutting your teeth with Ubuntu 7.10 chances are you don't know better, but Linux never used to do nonsense like this. I've run run right up to the ceiling of RAM use (and dipping into swap, even) while doing things like compiling and the UI is still responsive.
That said, X does tend to get a bit sluggish after a couple weeks. It's gotten better recently, however.
Yes, but at least the ability to do so is present in Windows (which I do not personally use, I should note). This isn't the case with MacOS X.
The problem is that the "mac way" has never worked at a conceptual level for people who are more computer oriented. I don't see "windows as documents" so the application-centric model of MacOS is fundamentally broken; I see windows (more or less) as a process managed by another process to allow me to manage them more clearly. The Apple Way is an additional abstraction, making the user go through additional hoops.
RAM is cheap. It's like $30 a gigabyte or less, and systems can take 8 to 32 of said gigabytes these days. That is more than enough RAM to handle strap-on apps to enhance OS functionality. Unfortunately, OS X pretty much inhibits fixing things in such a manner due to its fundamental design.
You can have an "application centric" model in Windows, if you want (task grouping). It's very similar. The same can not be said of OS X and a window centric model. You are irrevocably shackled to the magic application window which continually interferes with input focus and window switching.
Yet, Spaces does not fix the fundamental flaw: an application centric focus and task management model.
There are still a slew of commonly-performed tasks which are impossible due to the anachronistic app-centric model, all because pre-OSX, MacOS did not suitably perform multiprocess functionality. THAT is the real excuse behind the "Application centric" focus in Apple's operating systems.
I understand the difference, but it still makes working with two separate applications very, very irritating. Just try opening a PDF as a reference point while you're trying to write a document in NeoOffice... you'll pull your hair out.
Expose is very nice and, IMO, it's a necessary feature due to the fact that the Dock is so crippling when it comes to window switching. OS X would be almost unusable for me if Expose wasn't there.
The simple fact is that people do not usually think in terms of applications. They think "I need to work on x, y, and z documents" or "I want to open my resume" or "I want to listen to music while I surf the web". This precludes an "application centric" type focus, and is much closer to the "window centric" method of windowing.
Expose is necessary because the Dock does not scale. Yes, it is very useful within a limited context - ie, there are only a half dozen windows/applications open with a document or two each. Once you start opening 3 or so documents, or need to use two applications side-by-side/concurrently, the Dock (and the Application centric method in general) is a hindrance more than a feature.
Unfortunately, Expose does not scale that well either - at least not as well as a window-centric taskbar. It reaches its limitation at around, I'd say, 8 or so windows, necessitating the use of Spaces instead of just making it a "nice feature to have".
I should note that from my limited exposure, W7 is likely going to be a bit more of a PITA than XP has been, in terms of dealing with large numbers of applications/windows.
What do you mean, this isn't a big deal in the USA job market? Are you saying it's not a big deal to be laid off in the US, or that it's not a big deal because Americans don't have families? Or something else?
As an IT worker who has been laid off twice in the last year due to downturns who has a family, let me tell you, it is a fucking big deal. HR types still very much have the mentality of "oh you were laid off? There must be something wrong with you" and won't even consider someone who's got multiple short stints of employment in a row.
Microsoft is most certainly improving its image with Windows 7. They appear to be getting a lot of things right. They've improved system latency due to I/O over what was present in even XP, and the system is surprisingly stable (for a beta, of course).
Couple this with the fact that the Linux I/O scheduler appears to have moved away from a model which works well on the Linux desktop. For about the last year or so, Linux kernels have resulted in very latent desktop utility during even moderate burst-type I/O (programs/files loading, access of swap - not prologued disk writes). This may or may not be related to the bug supposedly introduced into the kernel in 2.6.18 - I don't know, I haven't personally tested it. But what I do know is that this behavior has become progressively more evident over the past 8 years: I blame the server-centric development focus in the kernel (2.2 and prior were blindingly responsive on the desktop).
With the fact that Linux desktop performance is somewhat lackluster these days giving it a perceived performance more on par with what Vista is capable of, I can see how it would sour people in preference for Windows 7, when Windows 7 appears to implement things properly - or, at least in a way which works to user expectations.
I should note that I've been personally using Linux (mostly Debian, some Ubuntu and OpenSuse) almost exclusively since around 2000. I don't make these criticisms lightly, and personally say it more as an admonishment of the Linux developers/community than I do as a proponent of W7. Whether it's a good product or not, I can not ethically approve of vendor lock in to the extent that MS software use encourages.
(Side note: has anyone noticed how W7's window effects/widgets (to the exception of the "MS-specific blurry/imperfect glass semi-transparent menus) looks shockingly like the bastard child of KDE 4 and OS X 10.5? I thought the first W7 screenshot I saw actually was KDE4 with a 'lookalike' theme.)
I think the biggest thing is that other sources of information are becoming more available - ie, digital sources. Five, 10 years ago, there wasn't much for online reference material on emerging tech (and the breadth of emerging tech was smaller), yet this was still the case.
Keep an eye on traditional media - printed and multimedia both. Especially printed publishing: they're all going to take a hard hit in this recession. Hopefully, it will mean a little respite in the amount of junk mail in the mailbox and the number of high-gloss printed fliers and such we have to put up with now.