The other question, of course, is whether the "On Macs" requirement is technical in some serious sense(any one of the modern virtualization tricks where you pass as much as possible through to the hardware, rather than trapping it and crunching it in software emulation depending somehow on EFI, or the particular chipsets of Macs, or something of that nature), or whether it is a purely artificial constraint, that exists to keep Oracle out of range of Steve Jobs' eye lasers...
If the former, turning it into a general purpose "Virtualbox virtualizes OSX" will require some nontrivial work. If the latter, I'm assuming that there will be a third party build of VirtualBox(large swaths of which are, after all, FOSS) supporting OSX on arbitrary hosts floating around within a matter of days, or even hours.
The problem for Microsoft and tablet support is that, while in strict engineering terms, tablet support is "something you add"(ie. wacom drivers, handwriting/ink support, some touch gestures, a demo app or two), in terms of design, UI, and pleasantness of user experience, tablet support is all about what you remove. It's like the old notion of "burning the boats" to inspire your army.
As long as MS approaches tablet support as just a few optional features, that can be added as a superset of their primary OS, they may well be technically competent(I've heard that their handwriting recognition is actually pretty good, for instance); but they will, outside of tech-demo-ware and highly specific custom applications, never escape the massive gravitational pull of the gigantic install base of the touchless OS. At worst, their superset offering will be completely ignored. At best, it will find a few niches, and a reasonably broad adoption in the form of "pen=mouse" ports of existing applications. Since these applications won't be all that comfortable, manufacturers will back off from bold all-tablet designs, and just start churning out "convertibles", which are just laptops with a wobbly single hinge and a screen that looks like crap because of the digitizer layer.
This is one of MS's major strength/weakness combinations. They have the resources(and some genuinely good people) to relentlessly add interlocking feature-set after interlocking feature-set to their products. However, because of their enterprise orientation, they are not good at the exotic, or the starkly cut down. Any innovation has to be capable of being tacked on to the gigantic interlocking feature mass. Any cut-down subset has to alienate as few 3rd parties and legacy customers as possible, and integrate with the feature mass as much as possible. On the plus side, this means that their stuff makes it relatively easy(if not wise) to build a towering enterprise stack, and then have it supported for years and years. On the minus side, it pretty much stomps on innovation, even where technologically possible.
The problems arise with applications that make heavy use of the GPU...
Near-native CPU performance inside VMs has been a solved problem for a while now(outside of old or deliberately crippled chips); but the work is just beginning for system peripherals. Some of the nicer Real Serious Server VM setups allow you to do what amounts to "PCI passthrough", where you can actually graft specific devices on the PCIe bus on to VMs(so, for example, an Astrix VM could talk to a physical POTS interface card); but that hasn't trickled down to general desktop use yet.
The task of sharing a single GPU between host and guest(s) is even trickier. Some VM software has rudimentary support for 3D inside the VM; but the performance penalties are quite serious.
Curious. I just tried it last night on a desktop with an RT2860 and it worked normally. Something about device IDs, or the difference between PCI and miniPCIe, or what?
There is no way that it would be worth 1.2 billion just for a couple of ok-but-not-thrilling phone bodies(and the patents, while nice, probably won't change HP's world too much; because they are already big enough to be locked in the Patent Cold War with all the other major players who have both patents and products).
If they didn't want WebOS, they wouldn't have bothered.
Conveniently, situations that involve absolutely titanic amounts of power being generated in one place, and needing to be shipped somewhere else, are exactly the sort of thing where using superconductive lines becomes economically practical...
Depending on how much direct purchasing and/or persuasive power you have, you might seriously wish to consider substantially increasing the number of images you keep.
"Nearline" storage is crazy cheap these days. Say you have a user who won't be fully productive for a week following a re-image, and will have IT hovering around him during that week. Unless you are both being paid in peanuts, it would almost certainly be cheaper to have a machine/user-specific image. Assuming a modestly compressed image format, that's what, maybe 50GB on the outside, with a huge load of custom software, most likely down in the 5-10GB range?
Even if you have virtually no purchasing pull, you should be able to get 2TB or more of space(RAID1 for protection against drive failure) to store user-specific images for your toughest cases for under $500(basically $200 each for 2TB drives, another $100 for a cheap RAID card if the discarded desktop from last year's refresh doesn't have one). If you have more serious pull, you could get something with data de-duplication and rack-rails for a few thousand.
Failing that, or in addition to that, I'd make a policy of(whenever you have time), trying to look into ways of scripting/automating things that you find yourself doing. It sounds like you are in an MS environment, so that probably means getting familiar with WMI. It is pretty byzantine; but crazy powerful. Combine, for instance, a script that runs daily and dumps each user's list of desktop shortcuts to a SVN or similar version control system with a second script that can restore the desktop shortcuts from any particular revision. Should be totally doable with a bit of WMI-fu, should occupy miminimal space(since the list of desktop shortcuts is a tiny text file that doesn't' change much), and should save you a fair amount of time.
I suspect that that is something you would do for basically the same reason that evidence gathered by illicit means isn't allowed to be used at trial. Not because you actually want it to be(who would want to throw out evidence or pay unemployment to the office asshole?); but because its presence curtails worse behavior.
If being fired "for disagreement or dispute with boss or co-worker" is grounds for denial of benefits, one strongly suspects that a lot of people who were on the layoff track will magically sprout really disagreeable personalities and simply unbearable attitudes... Instant savings.
Arguably, hourly pay(or salary, which, unless you are somehow skating out the door at lunch half the time, or your boss is really screwing you on the unpaid overtime, is basically equivalent to hourly but more stable), can function as a kind of insurance structure. Similar to the "lawyering up when needed" vs. "having lawyer on retainer".
In internal IT/operations, say, demand ebbs and flows. Some days the shit really hits the fan, some days are merely full, some days everything actually works properly, with minimal intervention.
The company has basically two options: They can try to be all dynamic, bringing in pricey white-knight consultants when the situation gets busy, and scraping by on the core "minimal intervention" staffing level the rest of the time. Or, they can hire an average number, knowing that they'll be a bit over-busy some days, and have some definite downtime other days(which, in an ideal world, can go into research/self-training/experimentation on process improvement).
In many(though not all) cases, the latter actually works out to be cheaper. Even genuinely brilliant people take a little while to get their bearings, and the facts on the ground, when plopped into a new situation. That, plus the premium that consultants command for their flexibility, can make them extremely expensive. Even if they are "working" for all the hours you are paying them, "getting up to speed on your environment" is work; but isn't useful to you. Permanent people only have to do it once, plus, people demand less money in stable positions.
If its wikipedia entry is to be believed, FITS is extraordinarily flexible. (whether or not this is a good thing in a file format is rather less clear).
Luckily, the number of non-geeks who care about media container formats, beyond the level of "will it play/won't it play", is approximately zero, so any failure to style your communications for the layman is largely unimportant...
The situation isn't entirely bleak(see Lexmark vs. Static Control); but I don't think that there is much cause for optimism about that "not much of a stretch" actually being made.
As you say, you usually can't get around first sale just by doing something trivial like printing a scary license agreement on the back cover of a book. The buyer can still resell the book, not too much you can do about it. However, objects without clear physical manifestation, which are becoming more common, don't have that defense. If you buy a book through your Kindle(or an app on your iPod, or some downloadable content, or a Steam game...), you cannot transfer it to somebody else(for free or for a consideration of any kind) without either hacking the system(at which point you would draw fire from the DMCA) or with Amazon's consent(if they added a "sell this book" button to the program). Because circumvention is hard for an average person of no special technical skill(and, even for the skilled, attacking closed appliances isn't getting any easier), code is de facto law. If the system won't let you, you can't. Because of the DMCA and its ilk, code is, in many cases, actual law.
Aside from purely digital items, the other main threat is physical items with substantial digital logic embedded in them. This is because, while first sale protects your right to sell something, it has not(historically) faced the case of property that can simply refuse to function if resold, or otherwise used against the manufacturer's wishes.
Quite the contrary. Black markets often suffer from relatively weak, corrupt, or inconsistent governance(gang warfare isn't going to find its way into the dictionary under "due process" any time soon); but they virtually always have nontrivial regulatory systems.
They tend toward being somewhat decentralized, socially structured, and based on direct violence(rather than centralized, bureaucratic, and based on potential violence); but they are definitely there.
Highly unlikely that that would make much difference.
There is a class of activity that consists more or less exclusively of making bets about the relative movements of currencies, and there is probably some behavior in "real" markets that is affected by currency movements; but most of this stuff is just about price discrimination, which will be economically rational for sellers as long as there are multiple buyers willing to pay different amounts more than the marginal cost of production.
National borders happen to be one convenient mechanism for price discrimination; but there are many others. The idea that such techniques deserve the protection of law, against the doctrine of first sale, though, is disgusting.
It's even nastier than that, in some cases, because some "real and tangible" property is now complex enough to enforce(against all but extremely sophisticated individuals) its own rules about use.
CSS is pitifully weak; but it was perhaps the first demonstration of this concept that gained huge market traction. Thanks to CSS licensing requirements, adding technologically enforced region coding became trivial.
As the cost of computing power continues to fall, and the number of devices that have embedded firmware and/or unique serial numbers continues to increase, there is virtually no area of "real property" over which the DMCA and copyright law will not eventually exert de facto control.
Of course not. When a corporation needs cheap labor/materials/lax laws/no taxes, then "Free Trade" is what makes the world go round.
If you wish to engage in a little arbitrage of your own, then "grey market" is about the most polite term. It is Vital to uphold Safety Standards and Preserve Market Stability, after all...
Architecturally, whitelisting is a great solution. In closed environments with fairly static requirements(eg. corporate) you can do it Right Now, if you want. And, while it won't save you from truly subtle attacks on the kernel or services, it blocks a fair percentage of common stuff good and hard.
The trouble begins when you try to implement it in the real world. Being the "they" who gets to bless all good programs is both a gigantic pain in the ass(requiring a massive staff of analysts and sophisticated techniques to keep up with the stream of software being produced, not to mention the problem of in-house and bespoke/private stuff) and a truly ghastly temptation, with which no entity can easily be trusted. He who gets to bless "good" software is the gatekeeper, a position of incredible market power, from which most any agenda(financial, political, aesthetic, or moral) may easily be advanced.
Software whitelisting is to computer security what dictatorship is to governance. In principle, if you could find a benevolent and wise dictator, it'd beat the hell out of the chaos and inefficiency of other institutions. Obtaining such, though, is a bit of a trick.
I'm with you on being done with supporting home users of Windows; but minis start at $700, with 2GB of RAM and no monitor. Dell will furnish you with a (big, ugly) box with triple the RAM, a 1TB HDD(rather than 160GB), and a 20 inch flat panel for the same money...(getting a 2.8GHz Phenom X4 instead of a 2.3GHz Core2 duo is just icing).
The mini is cuter, certainly, and if you have to have OSX you have to have OSX; but the pricing is hardly equivalent for anybody willing to run linux or shove their computer under their desk.
The, perhaps more accurate; but equally troubling, fact is that Powerpoint allows us to continue being stupid.
When you go to write something, an essay or a brief or something of that sort, you generally start stupid. You have a dubiously coherent mass of questionably formed notions. It's ghastly. However, because of the way the essay format works, you will have a very hard time getting away with that. If you don't clean your ideas up, think things through, force them into some semblance of coherent order, your essay will be transparently worthless.
Powerpoint, though, makes it a relatively simple and mechanical exercise(and let us not even mention the dreaded "autocontent wizard". The mere existence of such a monster should tell you everything you need to know about the epistemological framework in which Powerpoint is operated...) to transfer the confused morass of half-formed ideas that you start with straight to the page.
Back in the day, I had a Laserjet 4L that was basically immortal(it was 12 years old when I got it, and did all my printing in college no problem). Toner cartridges were big and cheap, and there was no driver shit to deal with because it spoke some well-supported PCL dialect.
These days, they all have integrated wireless and multifunction scanners and whatnot; but the driver install weighs in at 700MB(and doesn't work), the consumables last about as long as it took you to get them out of the package, and the firmware seems to know more about throwing cryptic errors than printing.
Assuming that you are the incumbent(whether The top dog, or just one of the people for whom the status quo is working quite well, thanks), every day you successfully delay change is another day of profit rather than loss, and risk rather than security. There may be a point where you cut your own throat by resisting change(either the cost of resistance simply becomes too high, and consumes all your profits, or your resistance actively precludes your taking advantage of certain options in the changed future); but until you reach that point, a rearguard action is totally rational, even if it is inevitably doomed on the medium to long timescale. The degree to which rearguard actions are logical is increased if you have access to overt or covert subsidies. In the media case, they've been very effective in lobbying for copyright infringement, and its tools, to be ever more criminalized and, once criminalized, made a greater law enforcement priority. Fighting change is always cost effective when you are using somebody else's money...
Second is that change is only really inevitable in hindsight. Many changes have been successfully fought, even though their proponents were convinced of their inevitability. Incumbents who don't fight change don't remain incumbents for as long as incumbents who do; because almost any change, unless it is truly structurally unsound, can push you over unless you push back; but only a relative few changes are irresistible(and, even in those cases, see point 1).
On the minus side, I would be rather more surprised to see a net positive value in change resistance("net positive" in the "overall value across a society" sense from econ). Incumbents, by virtue of being incumbents, so very often have access to other people's money with which to fight change. Therefore, it is logical to suspect that(because of that effective subsidy) a greater-than-socially-optimal amount of change-resistance is generated. Further, all but the most dramatic innovations have a period of manifest inferiority to existing, well-polished, methods. During this period, they can be smothered in the cradle at comparatively low cost.
Arguably, in the past decade or so, HP moved from being a major driver of printing to being a major force driving people away from printing...
Their cheap seats aren't even worthy of execration at this point, and even their nicer stuff isn't what it used to be(Hey guys, y'know what was a great idea? Releasing a print driver that crashes the print spooler service if somebody prints a PDF...)
Unfortunately for them, Microsoft's "Online Services Division" has a deeper fundamental problem than merely losing money.
That is, much of the money that they do manage to make, they make by cannibalizing MS server products sales. Now, I'm sure that they'd rather cannibalize their own server product sales than have Google/Amazon/assorted 3rd party penguin swarm datacenters eat them, cannibalism beats starvation after all; but that is still sort of a depressing mandate.
Their only "greenfield", so to speak, revenue opportunities are search(at which they are fairly tepid) or in providing "the-first-hit-is-really-cheap, also granular" access to various MS services(Exchange, Sharepoint, MS SQL, etc.) to tiny outfits that can't afford to do them in house(and, given SKUs like Small Business Server, we are talking pretty small outfits).
Slashdot Mirror
The other question, of course, is whether the "On Macs" requirement is technical in some serious sense(any one of the modern virtualization tricks where you pass as much as possible through to the hardware, rather than trapping it and crunching it in software emulation depending somehow on EFI, or the particular chipsets of Macs, or something of that nature), or whether it is a purely artificial constraint, that exists to keep Oracle out of range of Steve Jobs' eye lasers...
If the former, turning it into a general purpose "Virtualbox virtualizes OSX" will require some nontrivial work. If the latter, I'm assuming that there will be a third party build of VirtualBox(large swaths of which are, after all, FOSS) supporting OSX on arbitrary hosts floating around within a matter of days, or even hours.
The problem for Microsoft and tablet support is that, while in strict engineering terms, tablet support is "something you add"(ie. wacom drivers, handwriting/ink support, some touch gestures, a demo app or two), in terms of design, UI, and pleasantness of user experience, tablet support is all about what you remove. It's like the old notion of "burning the boats" to inspire your army.
As long as MS approaches tablet support as just a few optional features, that can be added as a superset of their primary OS, they may well be technically competent(I've heard that their handwriting recognition is actually pretty good, for instance); but they will, outside of tech-demo-ware and highly specific custom applications, never escape the massive gravitational pull of the gigantic install base of the touchless OS. At worst, their superset offering will be completely ignored. At best, it will find a few niches, and a reasonably broad adoption in the form of "pen=mouse" ports of existing applications. Since these applications won't be all that comfortable, manufacturers will back off from bold all-tablet designs, and just start churning out "convertibles", which are just laptops with a wobbly single hinge and a screen that looks like crap because of the digitizer layer.
This is one of MS's major strength/weakness combinations. They have the resources(and some genuinely good people) to relentlessly add interlocking feature-set after interlocking feature-set to their products. However, because of their enterprise orientation, they are not good at the exotic, or the starkly cut down. Any innovation has to be capable of being tacked on to the gigantic interlocking feature mass. Any cut-down subset has to alienate as few 3rd parties and legacy customers as possible, and integrate with the feature mass as much as possible. On the plus side, this means that their stuff makes it relatively easy(if not wise) to build a towering enterprise stack, and then have it supported for years and years. On the minus side, it pretty much stomps on innovation, even where technologically possible.
The problems arise with applications that make heavy use of the GPU...
Near-native CPU performance inside VMs has been a solved problem for a while now(outside of old or deliberately crippled chips); but the work is just beginning for system peripherals. Some of the nicer Real Serious Server VM setups allow you to do what amounts to "PCI passthrough", where you can actually graft specific devices on the PCIe bus on to VMs(so, for example, an Astrix VM could talk to a physical POTS interface card); but that hasn't trickled down to general desktop use yet.
The task of sharing a single GPU between host and guest(s) is even trickier. Some VM software has rudimentary support for 3D inside the VM; but the performance penalties are quite serious.
Curious. I just tried it last night on a desktop with an RT2860 and it worked normally. Something about device IDs, or the difference between PCI and miniPCIe, or what?
There is no way that it would be worth 1.2 billion just for a couple of ok-but-not-thrilling phone bodies(and the patents, while nice, probably won't change HP's world too much; because they are already big enough to be locked in the Patent Cold War with all the other major players who have both patents and products).
If they didn't want WebOS, they wouldn't have bothered.
Conveniently, situations that involve absolutely titanic amounts of power being generated in one place, and needing to be shipped somewhere else, are exactly the sort of thing where using superconductive lines becomes economically practical...
Depending on how much direct purchasing and/or persuasive power you have, you might seriously wish to consider substantially increasing the number of images you keep.
"Nearline" storage is crazy cheap these days. Say you have a user who won't be fully productive for a week following a re-image, and will have IT hovering around him during that week. Unless you are both being paid in peanuts, it would almost certainly be cheaper to have a machine/user-specific image. Assuming a modestly compressed image format, that's what, maybe 50GB on the outside, with a huge load of custom software, most likely down in the 5-10GB range?
Even if you have virtually no purchasing pull, you should be able to get 2TB or more of space(RAID1 for protection against drive failure) to store user-specific images for your toughest cases for under $500(basically $200 each for 2TB drives, another $100 for a cheap RAID card if the discarded desktop from last year's refresh doesn't have one). If you have more serious pull, you could get something with data de-duplication and rack-rails for a few thousand.
Failing that, or in addition to that, I'd make a policy of(whenever you have time), trying to look into ways of scripting/automating things that you find yourself doing. It sounds like you are in an MS environment, so that probably means getting familiar with WMI. It is pretty byzantine; but crazy powerful. Combine, for instance, a script that runs daily and dumps each user's list of desktop shortcuts to a SVN or similar version control system with a second script that can restore the desktop shortcuts from any particular revision. Should be totally doable with a bit of WMI-fu, should occupy miminimal space(since the list of desktop shortcuts is a tiny text file that doesn't' change much), and should save you a fair amount of time.
I suspect that that is something you would do for basically the same reason that evidence gathered by illicit means isn't allowed to be used at trial. Not because you actually want it to be(who would want to throw out evidence or pay unemployment to the office asshole?); but because its presence curtails worse behavior.
If being fired "for disagreement or dispute with boss or co-worker" is grounds for denial of benefits, one strongly suspects that a lot of people who were on the layoff track will magically sprout really disagreeable personalities and simply unbearable attitudes... Instant savings.
Arguably, hourly pay(or salary, which, unless you are somehow skating out the door at lunch half the time, or your boss is really screwing you on the unpaid overtime, is basically equivalent to hourly but more stable), can function as a kind of insurance structure. Similar to the "lawyering up when needed" vs. "having lawyer on retainer".
In internal IT/operations, say, demand ebbs and flows. Some days the shit really hits the fan, some days are merely full, some days everything actually works properly, with minimal intervention.
The company has basically two options: They can try to be all dynamic, bringing in pricey white-knight consultants when the situation gets busy, and scraping by on the core "minimal intervention" staffing level the rest of the time. Or, they can hire an average number, knowing that they'll be a bit over-busy some days, and have some definite downtime other days(which, in an ideal world, can go into research/self-training/experimentation on process improvement).
In many(though not all) cases, the latter actually works out to be cheaper. Even genuinely brilliant people take a little while to get their bearings, and the facts on the ground, when plopped into a new situation. That, plus the premium that consultants command for their flexibility, can make them extremely expensive. Even if they are "working" for all the hours you are paying them, "getting up to speed on your environment" is work; but isn't useful to you. Permanent people only have to do it once, plus, people demand less money in stable positions.
If its wikipedia entry is to be believed, FITS is extraordinarily flexible. (whether or not this is a good thing in a file format is rather less clear).
Luckily, the number of non-geeks who care about media container formats, beyond the level of "will it play/won't it play", is approximately zero, so any failure to style your communications for the layman is largely unimportant...
The situation isn't entirely bleak(see Lexmark vs. Static Control); but I don't think that there is much cause for optimism about that "not much of a stretch" actually being made.
As you say, you usually can't get around first sale just by doing something trivial like printing a scary license agreement on the back cover of a book. The buyer can still resell the book, not too much you can do about it. However, objects without clear physical manifestation, which are becoming more common, don't have that defense. If you buy a book through your Kindle(or an app on your iPod, or some downloadable content, or a Steam game...), you cannot transfer it to somebody else(for free or for a consideration of any kind) without either hacking the system(at which point you would draw fire from the DMCA) or with Amazon's consent(if they added a "sell this book" button to the program). Because circumvention is hard for an average person of no special technical skill(and, even for the skilled, attacking closed appliances isn't getting any easier), code is de facto law. If the system won't let you, you can't. Because of the DMCA and its ilk, code is, in many cases, actual law.
Aside from purely digital items, the other main threat is physical items with substantial digital logic embedded in them. This is because, while first sale protects your right to sell something, it has not(historically) faced the case of property that can simply refuse to function if resold, or otherwise used against the manufacturer's wishes.
Quite the contrary. Black markets often suffer from relatively weak, corrupt, or inconsistent governance(gang warfare isn't going to find its way into the dictionary under "due process" any time soon); but they virtually always have nontrivial regulatory systems.
They tend toward being somewhat decentralized, socially structured, and based on direct violence(rather than centralized, bureaucratic, and based on potential violence); but they are definitely there.
Highly unlikely that that would make much difference.
There is a class of activity that consists more or less exclusively of making bets about the relative movements of currencies, and there is probably some behavior in "real" markets that is affected by currency movements; but most of this stuff is just about price discrimination, which will be economically rational for sellers as long as there are multiple buyers willing to pay different amounts more than the marginal cost of production.
National borders happen to be one convenient mechanism for price discrimination; but there are many others. The idea that such techniques deserve the protection of law, against the doctrine of first sale, though, is disgusting.
It's even nastier than that, in some cases, because some "real and tangible" property is now complex enough to enforce(against all but extremely sophisticated individuals) its own rules about use.
CSS is pitifully weak; but it was perhaps the first demonstration of this concept that gained huge market traction. Thanks to CSS licensing requirements, adding technologically enforced region coding became trivial.
As the cost of computing power continues to fall, and the number of devices that have embedded firmware and/or unique serial numbers continues to increase, there is virtually no area of "real property" over which the DMCA and copyright law will not eventually exert de facto control.
Of course not. When a corporation needs cheap labor/materials/lax laws/no taxes, then "Free Trade" is what makes the world go round.
If you wish to engage in a little arbitrage of your own, then "grey market" is about the most polite term. It is Vital to uphold Safety Standards and Preserve Market Stability, after all...
Who, exactly, is "they"?
Architecturally, whitelisting is a great solution. In closed environments with fairly static requirements(eg. corporate) you can do it Right Now, if you want. And, while it won't save you from truly subtle attacks on the kernel or services, it blocks a fair percentage of common stuff good and hard.
The trouble begins when you try to implement it in the real world. Being the "they" who gets to bless all good programs is both a gigantic pain in the ass(requiring a massive staff of analysts and sophisticated techniques to keep up with the stream of software being produced, not to mention the problem of in-house and bespoke/private stuff) and a truly ghastly temptation, with which no entity can easily be trusted. He who gets to bless "good" software is the gatekeeper, a position of incredible market power, from which most any agenda(financial, political, aesthetic, or moral) may easily be advanced.
Software whitelisting is to computer security what dictatorship is to governance. In principle, if you could find a benevolent and wise dictator, it'd beat the hell out of the chaos and inefficiency of other institutions. Obtaining such, though, is a bit of a trick.
I'm with you on being done with supporting home users of Windows; but minis start at $700, with 2GB of RAM and no monitor. Dell will furnish you with a (big, ugly) box with triple the RAM, a 1TB HDD(rather than 160GB), and a 20 inch flat panel for the same money...(getting a 2.8GHz Phenom X4 instead of a 2.3GHz Core2 duo is just icing).
The mini is cuter, certainly, and if you have to have OSX you have to have OSX; but the pricing is hardly equivalent for anybody willing to run linux or shove their computer under their desk.
Apparently you aren't familiar with how the "alternative" (sometimes "complementary" or "holistic") medical industry operates...
Because AntiVirus 2010 has just detected dozens or even hundreds of critical security threats that your existing AV has missed!
What upgrade could be more sensible?
The, perhaps more accurate; but equally troubling, fact is that Powerpoint allows us to continue being stupid.
When you go to write something, an essay or a brief or something of that sort, you generally start stupid. You have a dubiously coherent mass of questionably formed notions. It's ghastly. However, because of the way the essay format works, you will have a very hard time getting away with that. If you don't clean your ideas up, think things through, force them into some semblance of coherent order, your essay will be transparently worthless.
Powerpoint, though, makes it a relatively simple and mechanical exercise(and let us not even mention the dreaded "autocontent wizard". The mere existence of such a monster should tell you everything you need to know about the epistemological framework in which Powerpoint is operated...) to transfer the confused morass of half-formed ideas that you start with straight to the page.
It's too bad, really.
Back in the day, I had a Laserjet 4L that was basically immortal(it was 12 years old when I got it, and did all my printing in college no problem). Toner cartridges were big and cheap, and there was no driver shit to deal with because it spoke some well-supported PCL dialect.
These days, they all have integrated wireless and multifunction scanners and whatnot; but the driver install weighs in at 700MB(and doesn't work), the consumables last about as long as it took you to get them out of the package, and the firmware seems to know more about throwing cryptic errors than printing.
Oh, definitely. Two main things:
Assuming that you are the incumbent(whether The top dog, or just one of the people for whom the status quo is working quite well, thanks), every day you successfully delay change is another day of profit rather than loss, and risk rather than security. There may be a point where you cut your own throat by resisting change(either the cost of resistance simply becomes too high, and consumes all your profits, or your resistance actively precludes your taking advantage of certain options in the changed future); but until you reach that point, a rearguard action is totally rational, even if it is inevitably doomed on the medium to long timescale. The degree to which rearguard actions are logical is increased if you have access to overt or covert subsidies. In the media case, they've been very effective in lobbying for copyright infringement, and its tools, to be ever more criminalized and, once criminalized, made a greater law enforcement priority. Fighting change is always cost effective when you are using somebody else's money...
Second is that change is only really inevitable in hindsight. Many changes have been successfully fought, even though their proponents were convinced of their inevitability. Incumbents who don't fight change don't remain incumbents for as long as incumbents who do; because almost any change, unless it is truly structurally unsound, can push you over unless you push back; but only a relative few changes are irresistible(and, even in those cases, see point 1).
On the minus side, I would be rather more surprised to see a net positive value in change resistance("net positive" in the "overall value across a society" sense from econ). Incumbents, by virtue of being incumbents, so very often have access to other people's money with which to fight change. Therefore, it is logical to suspect that(because of that effective subsidy) a greater-than-socially-optimal amount of change-resistance is generated. Further, all but the most dramatic innovations have a period of manifest inferiority to existing, well-polished, methods. During this period, they can be smothered in the cradle at comparatively low cost.
Arguably, in the past decade or so, HP moved from being a major driver of printing to being a major force driving people away from printing...
Their cheap seats aren't even worthy of execration at this point, and even their nicer stuff isn't what it used to be(Hey guys, y'know what was a great idea? Releasing a print driver that crashes the print spooler service if somebody prints a PDF...)
Unfortunately for them, Microsoft's "Online Services Division" has a deeper fundamental problem than merely losing money.
That is, much of the money that they do manage to make, they make by cannibalizing MS server products sales. Now, I'm sure that they'd rather cannibalize their own server product sales than have Google/Amazon/assorted 3rd party penguin swarm datacenters eat them, cannibalism beats starvation after all; but that is still sort of a depressing mandate.
Their only "greenfield", so to speak, revenue opportunities are search(at which they are fairly tepid) or in providing "the-first-hit-is-really-cheap, also granular" access to various MS services(Exchange, Sharepoint, MS SQL, etc.) to tiny outfits that can't afford to do them in house(and, given SKUs like Small Business Server, we are talking pretty small outfits).