My rotation scheme relies on having a different (and unrelated) number of drives than days of the week. Otherwise, Friday's deletion would only affect the Friday Drive, and the other four would slowly fill up without manual care and feeding.
Doing it this way distributes things nicely. For instance, with four drives:
MTWTF....F....F....F....F (day of week) 1234123412341234123412341 (drive number)
On the first Friday, drive #1 gets nuked and rewritten. Next week, #2 gets nuked and rewritten. And so on.
This gives us four weeks worth of daily backup points, while also resisting incorrect (or missed) rotation very well.
If I had 5 drives and 5 rotations per week, this wouldn't work at all and I might actually have to babysit it myself. And I don't: The boss takes the backup home with him when he leaves, or the receptionist does if he's not in that day, and one of them brings another drive back in the morning. The worst-possible rotation screwup results in larger granularity between available backups, but they'll still stretch out a few weeks even then.
Accordingly, I haven't needed to interfere with the daily operation of this system since I implemented it several years ago.
Last year's data is a special case: We pull a drive from rotation at the end of the year and keep it around until the accountant declares that he will certainly not need/want to use a time machine at any point in the coming year, and then it goes back into rotation.
It's not BS. But I didn't claim to be able to go back years. I kind of implied it, which may have been a mistake. Whatever the case, quite simply: Every winter before the accountant "closes out the year" (whatever that means in accountant-speak), we pull a backup disk and set it aside.
Sometime down the road, one of us will ask him if he's really sure if he's done with last year's stuff, and put it back into rotation.
This small point was not an attempt to proclaim that we maintain years' worth of backups, but just to demonstrate that we know the system works because we actually do test it periodically. The above is simply one of the actual scenarios wherein things get tested, and was mentioned simply to reinforce the point that testing actually-fucking-happens (which seems to be a concept that many folks' backup systems lack).
Which, I think, anyone should be able to figure out from the context of my original post. Perhaps I should've been more clear, or perhaps you're just exceptional.
No. Thursday doesn't have an assigned drive, and every single backup drive has an entire copy of the system already on it except for a brief period on Friday afternoon.
That's really not how Acronis works in our use. There is no "problem" to solve with fancy rotation.
Every backup disk has a full backup on it, except for a few minutes on Friday afternoon (when one of them has nothing). Period.
Acronis's idea of an incremental backup consists of looking at what that particular backup disk already contains, comparing it to whatever is on the live system, and copying the changes.
So, for example: Provisioning a new backup disk is as simple as formatting it, plugging it into the USB cable that hangs out of the back of the server, and walking away... on any day of the week, even on days when only an incremental is scheduled to run. When the incremental backup does run, it compares the contents of the backup disk to the live system, sees that everything is different (since the backup disk is empty), and essentially does a full backup.
And the next time that disk is attached (a few days or a week or whenever later), it does an incremental backup against whatever is already present on the backup disk that happens to be attached at that time.
How does it do this in just a few minutes per day? Fuck, I don't know. It's some funky voodoo NTFS stuff that borders on magic, as far as I'm concerned, since it seems to take rsync ages to do what seems like the same thing. But like I've mentioned earlier, we test it periodically with our identical backup hardware, and it always works without drama whether restoring a few files or restoring onto bare metal.
So. In practice, it works like this: If one backup disk gets nuked once per week, and we have n disks, then we have n weeks worth of backups...as long as n is neither an integer factor nor a divisor of 5.
Why 5? Because we rotate disks 5 times per week. If we had 5 disks, the same one would get erased every Friday, while the other 4 would never get erased and eventually would fill up with incrementals. So having 3 disks works. 4 works. 5 doesn't work. 6 works, along with 7, 8, 9, 11[...]. 10, 15, or 20[...] don't work, because it's just the same problem as having 5, assuming things actually get rotated in order. (Using 2 disks works, but isn't very safe since they'd both be on-site [along with the live data!] at the same time.)
The same concept applies for any number of backup points and any number of backup disks, as long as all of the data can fit on the backup media along with whatever incrementals, and as long as either the number of disks or the number of backup points between wipes is odd. (In particular, if they're both even numbers, things don't distribute very well at all.)
That said: We really don't care much about going back in time. If we did, we could easily schedule things to wipe a backup disk every two weeks, or once a month, or whatever. It's more important to us to have multiple recent copies of our data off-site than it is to restore a file that Jerry deleted a few months ago.
First, you're thinking with dedicated media for days of the week, which isn't at all how this is arranged: There's few drives in rotation, and it doesn't matter which one gets used on which day, or if I (or another tech) borrow one for temporary portable storage, or if one of them gets dropped into a sewer, or if we decide to add more of them, or really even if someone forgets to rotate it (except, perhaps, on Friday).
It really just doesn't matter.
Every single backup disk has a full backup already on it. The daily incremental backup is against the particular full backup and any other incremental backups that are also on that disk.
So, any single backup disk can be used to fully restore the machine from bare metal, and at no time would it be advantageous to use more than one backup disk to perform a restore. If the machine (or the building) blows up two minutes after the backup disk gets blanked on Friday, you just go back to Thursday evening's backup. No big deal.
And by wiping a backup disk (whatever disk it may be) only once per week, we get a few weeks of daily historical points that could be restored (either selectively or in whole), which is way more than sufficient for our needs.
This is not a space-efficient way to do backups. However, it is foolproof, reliable, and cheap (at least for our meager server).
The bare drive to read/write that $25 800GB tape will have cost more by itself than our entire backup system (including drives, enclosures, and software), and presents itself as a single point of failure unless we buy more than one of them.
How, then, do you suppose that tape is cheaper or more reliable than a pile of inexpensive redundant hard drives?
I recognize that your needs may be different. Please recognize that mine may be as well.
Indeed. My own [small place of business] does its backups this way with Acronis:
- Every day, a drive is plugged in.
Every evening, an incremental backup is performed on that drive that is consistent with whatever the backup drive already had on it. (It just takes a few minutes for Acronis to get this done.)
After that it is removed and taken off-site. A different drive is plugged in the following day.
Fridays are special, in that the drive gets automagically erased, and a full backup is performed mid-day. But then the evening incremental ritual is the same as any other day of the week. -
Losing a backup disk is, at most, 24 hours worth of loss, and only then if it is coincident with losing the main system.
FWIW: We use the cheapest 2.5" laptop drives available, in the cheapest bus-powered 2.5" USB enclosures we can get our hands on (I think the last round of them cost us $5, each). And we test our backups randomly, whenever the accountant decides he wants to see what things looked like last week/last year. Acronis then cheerfully does a bare-metal restore from [random backup drive] onto our spare server-box, with 100% success.
Now: We're not a media company. We don't have tens of TB of changing data to back up on a daily basis. But most other small(ish) companies don't either...
Who needs tapes?
Re:I propose we Occupy "Occupy"
on
Occupy Flash?
·
· Score: 1
And guess which codec Flash video usually uses?
Let me guess: One that is widely-distributed and used for free (as in beer) which is available pre-packaged from Adobe for a multitude of different platforms in a form that can be installed and utilized without any possible legal/IP implications for the end-user?
I hear far less about "eye strain" today then I did a dozen years ago when almost everyone still had a CRT on their desk, with a disturbingly-large portion of them flickering at 60Hz...
But at the end of the day, can you really say that you'd prefer a device with no means of illuminating its own display over a device which has illumination available for use? The Qualcomm device in question here is of the latter type: Allegedly, it has a display which can be plainly seen in normal ambient light, as well as the ability to illuminate itself when deemed useful.
I often prefer reading things with low contrast and low luminosity, particularly at night, and especially when excess ambient light might bother other people around me.
Furthermore, even in today's civilized world, I don't always have control over the lighting of my immediate surroundings. Sometimes it is simply too dark to read, and I want to read anyway.
I also don't get migraines, which I believe is a (non-)trait that I share with most other people.
Please stop assuming that your own personal problems and preferences are universal amongst others. No amount of pseudo-science or personal anecdote is going to persuade me to believe that having a device with an electronic display that can self-illuminate is worse than having one that cannot.
Driver code-signing excludes ancient drivers. Modern drivers can still be poorly written and cause problems, whether or not they're signed, and are no more likely to follow any particular set of guidelines from MSFT than any other drives from any other vintage.
...and none of them will be searchable, whether with a hacked Siri client, a native Iris client, or the default voice/text search tools. (Unlike an iPhone, which apparently gets at least half of this right.)
Choice is easy: I already have an Android notepad app that I like very well, thanks. Adding its content to the phone's generalized search functionality is hard, though.
You made it personal. I'm just trying to keep it that way.
But anyway, let's suppose that we simply rewind to a point roughly a couple of years ago. It sure looks like their most expensive base ("off the rack") desktop system was, amusingly, about $4k at that time, while their least expensive was about $1200.
So if you've got a point to make based on anything even resembling a fact, I'm still waiting for it.
First, that your argument about money is based on the lie that is a contract subsidy.
Second, that your argument about speed is circular: Faster hardware is not necessary because games do not require faster hardware because faster hardware is not necessary.
Third, you're somehow assuming that I don't need to leave my house to earn a living, but that might be OK even if folks want to watch a movie, because:
Fourth, you're figuring that I want to buy every member of my family a $600 handheld device, because doing so is somehow better/cheaper than a Playstation.
Even though I mentioned it, you still ignore price.
In (just to pick some numbers) 2015, a game console costing $300 will have better graphics abilities than a $300 handheld, simply because non-mobile platforms don't have the same design constraints and goals that handheld platforms do.
As to speed: Yeah, sure. And, by God, I could play Crysis on my 7-year-old single-core Dell laptop, if I wanted to prove a point by doing so. But it's a far more enjoyable (and prettier) experience on my quad-core SLI desktop. Until games become absolutely photorealistic, the machines that run them will never be fast enough, and we're obviously quite a long way from that level of perfection.
And WTF would I want a "docking station" for, except to complicate my life? Who in the hell would want their personal telephone/pocket computer tied up playing games in the livingroom, especially if it takes extra hardware that is otherwise useless to make it worth doing? The only real practical gaming advantage of a dock over actual dedicated hardware (ie: a traditional console) would be having the same saved games available, but isn't that what "The Cloud" is for?
Count me out. When the kids are busy killing a rainy summer day gaming in front of the TV, I still want to be able to use my space-age battery-powered pocket computer/radio telephone to do old-fashioned stuff like, you know, earn a living...and this means not ever having it tied to a television that someone else is finding enjoyment from.
Handhelds are inherently very personal devices geared toward one person's use; a game console, on the other hand, is a social device, and intended to be enjoyed by multiple people at the same time.
The reason is simple:
"Yeah, sorry boss. I'd sure would like to get right on that, but the [boys are playing Tekken]/[wife is watching a movie] on my phone at the moment. Maybe I'll be able to leave and go work on your problem once they finish up." (This problem does not exist with a dedicated console.)
YMMV, but personally I think you're drinking too much of the Kool-Aid.
And there is every indication that going forward, mobile devices will become faster and start replacing consoles as gaming devices for ever more resource intensive games, which creates a serious question as to whether they will ever even make back their initial investment.
I personally believe that mobile gaming is overrated. I've been using mobile devices since Handspring was still around, and have also had my share of experience using relatively slick full-color PalmOS devices, various Apple iWidgets, specialized portable media players, and the like. (Fast forward through a lot of gear, and we conclude with my well-hacked Android device.)
And let me tell you: I don't want to game or watch movies on any of them, at least not in exclusion to playing on the BFT in the living room. When I'm gaming, I want other warm bodies to be involved. Even on single-player games: Whether the boy is helping me search a room in Fallout, or the wife is playing navigator/spotter in GT5, it's more fun with real people.
Playing games on a handheld is like doodling in the corner. It's fun, sometimes, but it's more socially rewarding to have others involved...even if that means doing something other than doodling.
I use my Droid for all sorts of different kinds of stuff (it is a very lovely little portable computer), but gaming on it has so far been limited to waiting at the doctor's office or similar. Gaming on it is a means to kill time, not so much a means of enjoyment. It's handy that I've always got it with me, but not really very much fun for extended periods.
For enjoyment, I want to sit on the couch with the PS3 (or maybe even the Wii) and physically relax a bit, or for more intense gaming I might opt for the superior performance and keyboard/mouse controls of my PC gaming rig.
Same with movies. I could watch movies on my Palm Zire 71 many years ago, and it worked well enough, but I just didn't ever really bother except as an experiment to determine its worth. It could also do a fairly wide array of console emulation, which was interesting, but not very fun as a practical matter: Around that time, I'd much have rather run a similar emulator on my original Xbox, and view it on the big Trinitron CRT in my living room.
I'm not hardcore enough to buy such things on release day, but when the next generation of consoles starts cropping up I'll certainly be looking at them and trying to figure out how to fit at least one of them cleanly into my budget within a year or so.
And, for the money, I'll -always- get more realism out of a console than I will a PC or a handheld. Always. (My Droid was $539 and its graphics suck, my first PS3 was about $400 and it did OK at the time, and my present gaming rig beats the snot out of a PS3 but it also cost $1750.)
And it shouldn't need said, but: As handhelds get faster/better/cheaper, so does everything else. And nothing else has any of the worries about size, heat dissipation, or power consumption like mobiles do. It therefore will -always- be cheaper to build fixed a gaming system than a handheld of equal worth.
Death of consoles? Please. This argument isn't as new as you think it is.
And what difference, really, do "code-signed" driver make except -- perhaps -- to inform the user that "yes, we [the driver people] insist that this driver is indeed our driver"?
That x64 drivers may tend to work more reliably than x86 drivers has nothing to do with whether or not they are signed. It's not some magic stamp of approval by some higher deity, or any indication of any particular worth or merit. It's just a statement that the compiled driver is, indeed, from the same folks who it says its from.
Any other observations made from their use is just correlation.
Slashdot Mirror
Close, but no.
My rotation scheme relies on having a different (and unrelated) number of drives than days of the week. Otherwise, Friday's deletion would only affect the Friday Drive, and the other four would slowly fill up without manual care and feeding.
Doing it this way distributes things nicely. For instance, with four drives:
On the first Friday, drive #1 gets nuked and rewritten. Next week, #2 gets nuked and rewritten. And so on.
This gives us four weeks worth of daily backup points, while also resisting incorrect (or missed) rotation very well.
If I had 5 drives and 5 rotations per week, this wouldn't work at all and I might actually have to babysit it myself. And I don't: The boss takes the backup home with him when he leaves, or the receptionist does if he's not in that day, and one of them brings another drive back in the morning. The worst-possible rotation screwup results in larger granularity between available backups, but they'll still stretch out a few weeks even then.
Accordingly, I haven't needed to interfere with the daily operation of this system since I implemented it several years ago.
Last year's data is a special case: We pull a drive from rotation at the end of the year and keep it around until the accountant declares that he will certainly not need/want to use a time machine at any point in the coming year, and then it goes back into rotation.
It's not BS. But I didn't claim to be able to go back years. I kind of implied it, which may have been a mistake. Whatever the case, quite simply: Every winter before the accountant "closes out the year" (whatever that means in accountant-speak), we pull a backup disk and set it aside.
Sometime down the road, one of us will ask him if he's really sure if he's done with last year's stuff, and put it back into rotation.
This small point was not an attempt to proclaim that we maintain years' worth of backups, but just to demonstrate that we know the system works because we actually do test it periodically. The above is simply one of the actual scenarios wherein things get tested, and was mentioned simply to reinforce the point that testing actually-fucking-happens (which seems to be a concept that many folks' backup systems lack).
Which, I think, anyone should be able to figure out from the context of my original post. Perhaps I should've been more clear, or perhaps you're just exceptional.
*shrug*
No. Thursday doesn't have an assigned drive, and every single backup drive has an entire copy of the system already on it except for a brief period on Friday afternoon.
That's really not how Acronis works in our use. There is no "problem" to solve with fancy rotation.
Every backup disk has a full backup on it, except for a few minutes on Friday afternoon (when one of them has nothing). Period.
Acronis's idea of an incremental backup consists of looking at what that particular backup disk already contains, comparing it to whatever is on the live system, and copying the changes.
So, for example: Provisioning a new backup disk is as simple as formatting it, plugging it into the USB cable that hangs out of the back of the server, and walking away... on any day of the week, even on days when only an incremental is scheduled to run. When the incremental backup does run, it compares the contents of the backup disk to the live system, sees that everything is different (since the backup disk is empty), and essentially does a full backup.
And the next time that disk is attached (a few days or a week or whenever later), it does an incremental backup against whatever is already present on the backup disk that happens to be attached at that time.
How does it do this in just a few minutes per day? Fuck, I don't know. It's some funky voodoo NTFS stuff that borders on magic, as far as I'm concerned, since it seems to take rsync ages to do what seems like the same thing. But like I've mentioned earlier, we test it periodically with our identical backup hardware, and it always works without drama whether restoring a few files or restoring onto bare metal.
So. In practice, it works like this: If one backup disk gets nuked once per week, and we have n disks, then we have n weeks worth of backups...as long as n is neither an integer factor nor a divisor of 5.
Why 5? Because we rotate disks 5 times per week. If we had 5 disks, the same one would get erased every Friday, while the other 4 would never get erased and eventually would fill up with incrementals. So having 3 disks works. 4 works. 5 doesn't work. 6 works, along with 7, 8, 9, 11[...]. 10, 15, or 20[...] don't work, because it's just the same problem as having 5, assuming things actually get rotated in order. (Using 2 disks works, but isn't very safe since they'd both be on-site [along with the live data!] at the same time.)
The same concept applies for any number of backup points and any number of backup disks, as long as all of the data can fit on the backup media along with whatever incrementals, and as long as either the number of disks or the number of backup points between wipes is odd. (In particular, if they're both even numbers, things don't distribute very well at all.)
That said: We really don't care much about going back in time. If we did, we could easily schedule things to wipe a backup disk every two weeks, or once a month, or whatever. It's more important to us to have multiple recent copies of our data off-site than it is to restore a file that Jerry deleted a few months ago.
No. You've got it all wrong.
First, you're thinking with dedicated media for days of the week, which isn't at all how this is arranged: There's few drives in rotation, and it doesn't matter which one gets used on which day, or if I (or another tech) borrow one for temporary portable storage, or if one of them gets dropped into a sewer, or if we decide to add more of them, or really even if someone forgets to rotate it (except, perhaps, on Friday).
It really just doesn't matter.
Every single backup disk has a full backup already on it. The daily incremental backup is against the particular full backup and any other incremental backups that are also on that disk.
So, any single backup disk can be used to fully restore the machine from bare metal, and at no time would it be advantageous to use more than one backup disk to perform a restore. If the machine (or the building) blows up two minutes after the backup disk gets blanked on Friday, you just go back to Thursday evening's backup. No big deal.
And by wiping a backup disk (whatever disk it may be) only once per week, we get a few weeks of daily historical points that could be restored (either selectively or in whole), which is way more than sufficient for our needs.
This is not a space-efficient way to do backups. However, it is foolproof, reliable, and cheap (at least for our meager server).
YMMV.
The bare drive to read/write that $25 800GB tape will have cost more by itself than our entire backup system (including drives, enclosures, and software), and presents itself as a single point of failure unless we buy more than one of them.
How, then, do you suppose that tape is cheaper or more reliable than a pile of inexpensive redundant hard drives?
I recognize that your needs may be different. Please recognize that mine may be as well.
Indeed. My own [small place of business] does its backups this way with Acronis:
-
Every day, a drive is plugged in.
Every evening, an incremental backup is performed on that drive that is consistent with whatever the backup drive already had on it. (It just takes a few minutes for Acronis to get this done.)
After that it is removed and taken off-site. A different drive is plugged in the following day.
Fridays are special, in that the drive gets automagically erased, and a full backup is performed mid-day. But then the evening incremental ritual is the same as any other day of the week.
-
Losing a backup disk is, at most, 24 hours worth of loss, and only then if it is coincident with losing the main system.
FWIW: We use the cheapest 2.5" laptop drives available, in the cheapest bus-powered 2.5" USB enclosures we can get our hands on (I think the last round of them cost us $5, each). And we test our backups randomly, whenever the accountant decides he wants to see what things looked like last week/last year. Acronis then cheerfully does a bare-metal restore from [random backup drive] onto our spare server-box, with 100% success.
Now: We're not a media company. We don't have tens of TB of changing data to back up on a daily basis. But most other small(ish) companies don't either...
Who needs tapes?
Let me guess: One that is widely-distributed and used for free (as in beer) which is available pre-packaged from Adobe for a multitude of different platforms in a form that can be installed and utilized without any possible legal/IP implications for the end-user?
Am I close?
I hear far less about "eye strain" today then I did a dozen years ago when almost everyone still had a CRT on their desk, with a disturbingly-large portion of them flickering at 60Hz...
But at the end of the day, can you really say that you'd prefer a device with no means of illuminating its own display over a device which has illumination available for use? The Qualcomm device in question here is of the latter type: Allegedly, it has a display which can be plainly seen in normal ambient light, as well as the ability to illuminate itself when deemed useful.
There is no spoon.
I often prefer reading things with low contrast and low luminosity, particularly at night, and especially when excess ambient light might bother other people around me.
Furthermore, even in today's civilized world, I don't always have control over the lighting of my immediate surroundings. Sometimes it is simply too dark to read, and I want to read anyway.
I also don't get migraines, which I believe is a (non-)trait that I share with most other people.
Please stop assuming that your own personal problems and preferences are universal amongst others. No amount of pseudo-science or personal anecdote is going to persuade me to believe that having a device with an electronic display that can self-illuminate is worse than having one that cannot.
Did it say which side they're on?
I think this is why the brightness of such displays is generally adjustable.
I think you've still got it wrong:
Driver code-signing excludes ancient drivers. Modern drivers can still be poorly written and cause problems, whether or not they're signed, and are no more likely to follow any particular set of guidelines from MSFT than any other drives from any other vintage.
I call it a grossly overbearing and network-centric alternative to a simple textual notepad that also didn't come with my phone. :)
That is, indeed, hilarious. Perhaps it's even intentional: "You want a notepad application? Here's a manual, kid: Go figure it out."
Which, you know, I might just do. It's as good an excuse as any to start getting my teeth cut on coding Android apps. Thanks.
...and none of them will be searchable, whether with a hacked Siri client, a native Iris client, or the default voice/text search tools. (Unlike an iPhone, which apparently gets at least half of this right.)
Choice is easy: I already have an Android notepad app that I like very well, thanks. Adding its content to the phone's generalized search functionality is hard, though.
Hmm.
Interesting.
Last I checked, my Android phone didn't even include a notepad. :-/
I've only used Siri a few times on borrowed iPhones...but what do you mean that it "fails so often due to network issues?"
Is this anything like when my Android phone tells me to "Speak Again?"
If so:
*shock* *horror*
Sure. Why not?
You made it personal. I'm just trying to keep it that way.
But anyway, let's suppose that we simply rewind to a point roughly a couple of years ago. It sure looks like their most expensive base ("off the rack") desktop system was, amusingly, about $4k at that time, while their least expensive was about $1200.
So if you've got a point to make based on anything even resembling a fact, I'm still waiting for it.
A subsidy comes with a contract. I'm unwilling to entertain a contract for a $600 device that is easily broken or lost.
I buy my kids phones, sure: But they're not $600. They're just phones.
So. Let me see if I've got this right:
First, that your argument about money is based on the lie that is a contract subsidy.
Second, that your argument about speed is circular: Faster hardware is not necessary because games do not require faster hardware because faster hardware is not necessary.
Third, you're somehow assuming that I don't need to leave my house to earn a living, but that might be OK even if folks want to watch a movie, because:
Fourth, you're figuring that I want to buy every member of my family a $600 handheld device, because doing so is somehow better/cheaper than a Playstation.
Did I get all of that right?
Even though I mentioned it, you still ignore price.
In (just to pick some numbers) 2015, a game console costing $300 will have better graphics abilities than a $300 handheld, simply because non-mobile platforms don't have the same design constraints and goals that handheld platforms do.
As to speed: Yeah, sure. And, by God, I could play Crysis on my 7-year-old single-core Dell laptop, if I wanted to prove a point by doing so. But it's a far more enjoyable (and prettier) experience on my quad-core SLI desktop. Until games become absolutely photorealistic, the machines that run them will never be fast enough, and we're obviously quite a long way from that level of perfection.
And WTF would I want a "docking station" for, except to complicate my life? Who in the hell would want their personal telephone/pocket computer tied up playing games in the livingroom, especially if it takes extra hardware that is otherwise useless to make it worth doing? The only real practical gaming advantage of a dock over actual dedicated hardware (ie: a traditional console) would be having the same saved games available, but isn't that what "The Cloud" is for?
Count me out. When the kids are busy killing a rainy summer day gaming in front of the TV, I still want to be able to use my space-age battery-powered pocket computer/radio telephone to do old-fashioned stuff like, you know, earn a living...and this means not ever having it tied to a television that someone else is finding enjoyment from.
Handhelds are inherently very personal devices geared toward one person's use; a game console, on the other hand, is a social device, and intended to be enjoyed by multiple people at the same time.
The reason is simple:
"Yeah, sorry boss. I'd sure would like to get right on that, but the [boys are playing Tekken]/[wife is watching a movie] on my phone at the moment. Maybe I'll be able to leave and go work on your problem once they finish up." (This problem does not exist with a dedicated console.)
YMMV, but personally I think you're drinking too much of the Kool-Aid.
I personally believe that mobile gaming is overrated. I've been using mobile devices since Handspring was still around, and have also had my share of experience using relatively slick full-color PalmOS devices, various Apple iWidgets, specialized portable media players, and the like. (Fast forward through a lot of gear, and we conclude with my well-hacked Android device.)
And let me tell you: I don't want to game or watch movies on any of them, at least not in exclusion to playing on the BFT in the living room. When I'm gaming, I want other warm bodies to be involved. Even on single-player games: Whether the boy is helping me search a room in Fallout, or the wife is playing navigator/spotter in GT5, it's more fun with real people.
Playing games on a handheld is like doodling in the corner. It's fun, sometimes, but it's more socially rewarding to have others involved...even if that means doing something other than doodling.
I use my Droid for all sorts of different kinds of stuff (it is a very lovely little portable computer), but gaming on it has so far been limited to waiting at the doctor's office or similar. Gaming on it is a means to kill time, not so much a means of enjoyment. It's handy that I've always got it with me, but not really very much fun for extended periods.
For enjoyment, I want to sit on the couch with the PS3 (or maybe even the Wii) and physically relax a bit, or for more intense gaming I might opt for the superior performance and keyboard/mouse controls of my PC gaming rig.
Same with movies. I could watch movies on my Palm Zire 71 many years ago, and it worked well enough, but I just didn't ever really bother except as an experiment to determine its worth. It could also do a fairly wide array of console emulation, which was interesting, but not very fun as a practical matter: Around that time, I'd much have rather run a similar emulator on my original Xbox, and view it on the big Trinitron CRT in my living room.
I'm not hardcore enough to buy such things on release day, but when the next generation of consoles starts cropping up I'll certainly be looking at them and trying to figure out how to fit at least one of them cleanly into my budget within a year or so.
And, for the money, I'll -always- get more realism out of a console than I will a PC or a handheld. Always. (My Droid was $539 and its graphics suck, my first PS3 was about $400 and it did OK at the time, and my present gaming rig beats the snot out of a PS3 but it also cost $1750.)
And it shouldn't need said, but: As handhelds get faster/better/cheaper, so does everything else. And nothing else has any of the worries about size, heat dissipation, or power consumption like mobiles do. It therefore will -always- be cheaper to build fixed a gaming system than a handheld of equal worth.
Death of consoles? Please. This argument isn't as new as you think it is.
And what difference, really, do "code-signed" driver make except -- perhaps -- to inform the user that "yes, we [the driver people] insist that this driver is indeed our driver"?
That x64 drivers may tend to work more reliably than x86 drivers has nothing to do with whether or not they are signed. It's not some magic stamp of approval by some higher deity, or any indication of any particular worth or merit. It's just a statement that the compiled driver is, indeed, from the same folks who it says its from.
Any other observations made from their use is just correlation.