It is so very convenient, is it not, that consumer electronics are one of the relatively few categories of goods that are getting cheaper faster than you('you' in an average sense, individual 'you's may and do vary) are getting poorer. The real cost of food, housing, medical care, education, and petrochemicals may be rising relentlessly against wages(if you have a job, employment insecurity, permatemping, and other fun are all up too!); but you have a big TV so you must be a damn lazy welfare queen...
Even if we accept the article's premise(that the 'great society' collection of programs was a failure), the best that that proves is that some contemporary critics of the Apollo program chose dubious grounds for criticism. As we have learned(and, incidentally, only by trying) social engineering is one of the trickier flavors of engineering.
Where TFA seems to go off the rails a bit is the jump from 'people who think we should have spent the money on 'great society' were wrong because great society failed' to 'Apollo program: Vindicated!'. If you want to assess the worth of a spaceflight R&D program, compare it to other possible spaceflight R&D programs(or to non-spaceflight R&D programs designed to produce interesting technologies: variations on the 'well, set the grad students loose to do basic research' are pretty cheap...)
As with any sufficiently large engineering project, there were some side effects. Somebody had to build the thing, and certain technological advances had to be made or perfected to get it working; but the same would be true of building a sufficiently large bridge to nowhere. If you actually want to vindicate a space program, you either have to admit that you are doing it because space is pretty cool, or seriously examine it against other possible technology programs, rather than digging up some overt failure to run against...
Insurance makes a degree of sense(as hard as insurers try to change this) when dealing with situations where risk is either inevitable(ie. your body, with all its potential for horrible and expensive mishaps, comes standard and you'll need one until you die) or a fairly large chunk of your net worth(most homeowning, say) or where your potential to hurt others potentially far exceeds your personal ability to compensate them and insurance is therefore mandated(car insurance on the consumer level, potentially various other flavors among venue operators and the like).
On cheap consumer devices, it just doesn't make much sense. The insurer has to make a profit in order to continue offering the insurance, so you know that(on average) purchasing the insurance is a bad deal compared to self-insuring, and you also know that the potential costs are bounded(ie. there is nothing that could happen to my cellphone that could possibly cost more than a new cellphone. There are plenty of diseases and/or accidents that could happen at any time that could run into an unpredictable but very large number that I don't even necessarily have a good way of estimating).
Just put the price of the insurance wherever you usually put money for storage, let the warranty handle any material defects/abnormal failures, and maybe buy a case if you are a bit of klutz. Unless you murder your phone both brutally and swiftly, you'll probably be able to get a refurb for the money you saved by not buying insurance, plus the deductable you would have paid, by the time your phone eventually does bite it. Worst case, a used or refurbed iphone 4/4S will cost peanuts if you kill your present one and really can't afford a replacement.
Given that, on average, buyers of insurance lose money, you should really only be buying it on things that are at the outer envelope of affordability; but that you must have for one reason or another.
Given that 'bored and/or drunk hicks shooting at the fiber' is an actual cause of outages in rural runs, I'd be overwhelmingly surprised to hear that people aren't using hunting rifles on speed traps...
I don't doubt that somebody will break out the anti-materiel rifle that they normally use for hunting armored moose; but the more plausible response would seem to involve paintball gear.
It's cheap, it's close to silent, BATF doesn't give a damn about it, so its relatively anonymous, and it coats visible light optics just fine, freeing you to come closer and give your target a dose of tire-iron surprise.
...into the Appelsk walled garden that Windows 8 appears to be heralding in (Windows Store only apps, "for your own security, comfort and ease of use", coming to you in Windows 8.5/9). Last thing our walled gardeners want is an alternative OS weed like Linux, working perfectly on the same hardware...
That does seem to be the plan these days; but the CPU level seems an odd place to do it. If you've got a TPM and the right crypto sauce in your bootloader(which, along with everything and the kitchen sink, UEFI certainly can and frequently does) you can lock unsigned code out without any expensive and time consuming(and potentially buggy) dicking around with the instruction set, or memory layout, or connector geometry, or anything like that.
Really, while they still can work, control-through-churn is an obsolete strategy. Churning your software, APIs, instruction sets, etc. hurts your own developers and leads to time and bugs, and churning your proprietary connectors only keeps the chinese clone shops at bay for about a week. The future belongs to cryptographically enforced incompatibility laid on top of as much cheap, reliable, standardized stuff as you can get away with using...
What I find odd about Intel's phrasing is that they seem to be (voluntarily, no less) throwing water on their chances of selling this particular silicon to somebody who wants to run linux on it...
Unless there is something seriously whacked about this part compared to a conventional x86 part, it should require, at most, modest tweaks to get Linux up and going. Now, support for the fancy new power management features and whatnot may or may not be there(just as the world is infested with motherboards with ACPI so broken that all OSes that aren't the version of Windows they were designed to sell with might as well just not touch ACPI at all to the degree possible); but that is quite different from 'cannot run'.
It seems particularly odd given that Intel is fairly likely to, for simple reasons of efficiency, try to use similar power management tricks across their CPU lineup, from SoCs to high end Xeons, and it'll be a cold day in hell before the bulk Xeon buyers of the world accept a chip that won't run Linux, or vmkernel, or the assorted other non-windows things that are entirely routine in server land.
Also, isn't this the same Intel that fails to understand that ARM is going to be very important in the future? AFAIK they're the only ones that aren't licensing the technology.
Intel purchased 'StrongARM' from DEC ages ago(back when DEC still had things you could purchase), took it through a few generations under that name and then as 'Xscale', and then sold it to Marvell 6ish years ago(with the possible exception of one flavor that they use on their RAID boards, I can't remember).
They still have an ARM license, they just aren't terribly motivated to use it. x86 doesn't have too many friends; but it certainly has a lot of customers, and Intel has somewhat... limited... incentive to march into the business of being yet another SoC shover as long as they can get away with the margins on their x86s parts and supporting silicon.
iPod Touches definitely are iOS devices, just without the cellular modems and with a light dusting of contempt from Apple when it comes to adding the punchier new hardware... Estimates are that 40-ish percent of iOS devices are ipod touches.
what fragmentation? there are three phones supported by ios6
Which means that there are a few tens to hundreds of millions of actively used phones that will never even run the current OS... That's sort of what 'fragmentation' is all about...
To be sure, Android is substantially more fragmented even if you only look at 'currently-sold-and-supported-by-people-you've-heard-of', since there are multiple hardware OEMs shoving handsets out; but unless you start killswitching all your products, 'fragmentation' inevitably happens whenever people don't stop using the older ones and you introduce something new.
The chart in TFA is entirely useless, since it focuses on end user features that apple has or hasn't included on various models; but there is the not-so-minor matter of spec changes(TFA's chart doesn't even touch them; but 'keeping the core APIs consistent' also doesn't address them). There are some pretty significant differences in CPU and GPU power, and how quickly the OS will run out of RAM and quietly start memory-managing you, between those models.
If your 'app' is just some lousy re-implementation of a website that you really wanted to flog through the app store for some reason, it probably isn't a big deal; but anybody who really needs the punch provided by running native can't necessarily ignore that.
Helium is light enough that it doesn't persist very well in the atmosphere(unlike the heavier noble gasses, that you can just distill out if the price gets high enough to pay for the energy needed), and it is only replenished quite slowly by alpha decay of assorted radioactives in the crust.
The only significant source is natural gas wells in proximity to suitable minerals over geologic time and equipped to capture the helium when the product is brought to the surface.
The Michigan proposal involved some assorted fancy-materials-science tricks(inks with very atypical optical properties and other stuff that the anti-counterfieting guys have been poking around at to raise the cost and required sophistication of producing a convincing fake) in addition to QR codes. If anything, the QR part seemed like something of a trend-crazed afterthought.
(Incididentally, the one thing that cryptography can do for physical items like currency is make it impossible for forgers to produce novel forgeries: If, for instance, the bill has a data field that is its serial number/place/time of manufacturer/etc. signed with a treasury private key, that doesn't stop me from just photocopying it; but it does prevent me from producing any bills that aren't direct copies of official bills. If combined with a reasonably effecient automated scanning system, to identify duplicates, this makes it more difficult to pass large numbers of copies of a single legitimate bill, and makes it impossible to produce anything other than copies of bills that were actually produced by the private key holder. Whether this is actually useful depends on what your mechanisms are for weeding out duplicates and tracing them back to their originators; but it can be done.)
In pure performance? Not particularly. In 'just supposed to work, and ships on some enormous percentage of all x86s churned out every year'? That's Intel.
which would make excessively fast blinks useless as indicators for humans to read. You could probably build something to read them. Assuming the LED's emitting the data have a fast enough rising and falling time.
That's basically how all optical networking not classy enough to afford laser diodes works(TOSLINK is probably the biggest example, wouldn't surprise me if somebody sold cheap-n-nasty GBICs that worked the same way at some point; but laser diodes have gotten pretty cheap...) However, since the indicator lights are intended for humans, apparently the chips driving higher speed interfaces(10mb ethernet and up) deliberately lengthen the on an off cycles(coincidentially cratering the bandwidth of the channel) in order to allow the human to distinguish between 'solid on' and 'blinking'. If the designs didn't intentionally do that, it is quite plausible that faster interfaces would leak through their indicators as well.
The RONJA guys coaxed 10mb out of unexceptional red LEDs across several hundred meters of free air, so the rise/fall times are apparently not insurmountable...
In that vein, the demo unit appears to be using red LEDs. I wonder if the phosphor-blobbed white ones are severely bandwidth constrained by the residual glow from the phosphor layer, or if the ghastly framerates of the cheapy cameras that intel proposes to use as receivers for these signals is the limiting factor?
It was demonstrated on low speed modems, some of which directly tied the blinkenlights to the serial traffic(including, if memory serves, an embarassing case where some model of fancy 'encrypted' modem tied the blinkenlights to the serial traffic before the encryption stage...); but the author concluded that status indicator lights on higher speed stuff were(either in response to the possibility of attack, or just because humans can't distinguish between ultrafast blinking and 'on', which would make excessively fast blinks useless as indicators) not usefully coupled to the data channel for anything more than vague inference about traffic volume...
I am a trifle surprised that nobody seems to have banged out a USB/bluetooth 'calculator keyboard' peripheral(external numpads are a dime a dozen; but I've never seen one with a scientific calculator's complement of operator symbols and things), since building HID devices to spit out whatever keycodes burned into their dinky little processors isn't a wildly expensive process; and would make phone or computer-based calculating a bit more comfortable.
As for dedicated calculators, though, it seems most unlikely: the second you get good enough to help little timmy skip learning in favor of data entry, you can kiss the educational market goodbye, and that is close to being the entire market for calculators, in volume terms... Test administrators can get touchy enough about the pathetic internal storage of the TI-83 and similar, they aren't likely to warm to something that has a radio data link and could trivially be farming out the problems on your math test over the internet to somewhere poor enough that "take a lazy first-worlder's SAT for $50" sounds like a really good idea even to those with actual math skills...
Some software(Photoshop being the big name; but not exclusive to them) also includes this 'feature'. If you manipulate an image of a major world currency in excessive detail, a neat little binary module included with photoshop will snag you and direct you to this rather bland organization.
The difference with digital camera watermarking is that EXIF is a (not always obvious depending on the UI, and sometimes less standard that it ought to be) standardized metadata storage system. The internet is rife with amusing mistakes made by people who don't know about exif and upload anyway; but that's a UI/user problem. The fields are well known, easily viewed and edited with commonly available software, and not designed to be covert or strip-resistant in any way. Some imaging devices are, quite arguably, excessively chatty by default, and that is a legitimate concern given user ignorance; but there isn't anything sneaky about the technology.
Watermarks, at least in this incarnation, are designed to be covert, strip-resistant, and are not intended for the creator of the image to be aware of.
This is a 'prisons and fortresses share certain architectural similarties; but do not share purposes' situation...
If memory serves, it isn't actually a factor of printer manufacturer(and/or re-badger); but of the OEM behind the color laser print engine. Apparently there are relatively few of those, and some, thanks to a little leaning from Uncle Sam the details of which have never come to light, include the watermarking 'feature' in all their print engines. Since printer manufacturers can, and sometimes do, switch parts suppliers between models, a given manufacturer might have both bugged and clean hardware on offer at a given time.
Blizzard actually poisons the kernel entropy pool so cleverly that 'random' behaviors by the computer end up leaking identifiable information. Very sneaky of them...
CALCnet allows networking of TI-83 and similar calculators with relatively simple external hardware.
With that detail out of the way, you are free to implement a display-wall and/or the most powerful z80 cluster computer in the known universe.
Extra credit, of course, will be awarded if you succeed in writing an xorg driver that can treat an MxN array of networked calculators as a greyscale display of appropriate resolution.
Slashdot Mirror
"the poor have color televisions":
It is so very convenient, is it not, that consumer electronics are one of the relatively few categories of goods that are getting cheaper faster than you('you' in an average sense, individual 'you's may and do vary) are getting poorer. The real cost of food, housing, medical care, education, and petrochemicals may be rising relentlessly against wages(if you have a job, employment insecurity, permatemping, and other fun are all up too!); but you have a big TV so you must be a damn lazy welfare queen...
Even if we accept the article's premise(that the 'great society' collection of programs was a failure), the best that that proves is that some contemporary critics of the Apollo program chose dubious grounds for criticism. As we have learned(and, incidentally, only by trying) social engineering is one of the trickier flavors of engineering.
Where TFA seems to go off the rails a bit is the jump from 'people who think we should have spent the money on 'great society' were wrong because great society failed' to 'Apollo program: Vindicated!'. If you want to assess the worth of a spaceflight R&D program, compare it to other possible spaceflight R&D programs(or to non-spaceflight R&D programs designed to produce interesting technologies: variations on the 'well, set the grad students loose to do basic research' are pretty cheap...)
As with any sufficiently large engineering project, there were some side effects. Somebody had to build the thing, and certain technological advances had to be made or perfected to get it working; but the same would be true of building a sufficiently large bridge to nowhere. If you actually want to vindicate a space program, you either have to admit that you are doing it because space is pretty cool, or seriously examine it against other possible technology programs, rather than digging up some overt failure to run against...
Insurance makes a degree of sense(as hard as insurers try to change this) when dealing with situations where risk is either inevitable(ie. your body, with all its potential for horrible and expensive mishaps, comes standard and you'll need one until you die) or a fairly large chunk of your net worth(most homeowning, say) or where your potential to hurt others potentially far exceeds your personal ability to compensate them and insurance is therefore mandated(car insurance on the consumer level, potentially various other flavors among venue operators and the like).
On cheap consumer devices, it just doesn't make much sense. The insurer has to make a profit in order to continue offering the insurance, so you know that(on average) purchasing the insurance is a bad deal compared to self-insuring, and you also know that the potential costs are bounded(ie. there is nothing that could happen to my cellphone that could possibly cost more than a new cellphone. There are plenty of diseases and/or accidents that could happen at any time that could run into an unpredictable but very large number that I don't even necessarily have a good way of estimating).
Just put the price of the insurance wherever you usually put money for storage, let the warranty handle any material defects/abnormal failures, and maybe buy a case if you are a bit of klutz. Unless you murder your phone both brutally and swiftly, you'll probably be able to get a refurb for the money you saved by not buying insurance, plus the deductable you would have paid, by the time your phone eventually does bite it. Worst case, a used or refurbed iphone 4/4S will cost peanuts if you kill your present one and really can't afford a replacement.
Given that, on average, buyers of insurance lose money, you should really only be buying it on things that are at the outer envelope of affordability; but that you must have for one reason or another.
Given that 'bored and/or drunk hicks shooting at the fiber' is an actual cause of outages in rural runs, I'd be overwhelmingly surprised to hear that people aren't using hunting rifles on speed traps...
I don't doubt that somebody will break out the anti-materiel rifle that they normally use for hunting armored moose; but the more plausible response would seem to involve paintball gear.
It's cheap, it's close to silent, BATF doesn't give a damn about it, so its relatively anonymous, and it coats visible light optics just fine, freeing you to come closer and give your target a dose of tire-iron surprise.
...into the Appelsk walled garden that Windows 8 appears to be heralding in (Windows Store only apps, "for your own security, comfort and ease of use", coming to you in Windows 8.5/9). Last thing our walled gardeners want is an alternative OS weed like Linux, working perfectly on the same hardware...
That does seem to be the plan these days; but the CPU level seems an odd place to do it. If you've got a TPM and the right crypto sauce in your bootloader(which, along with everything and the kitchen sink, UEFI certainly can and frequently does) you can lock unsigned code out without any expensive and time consuming(and potentially buggy) dicking around with the instruction set, or memory layout, or connector geometry, or anything like that.
Really, while they still can work, control-through-churn is an obsolete strategy. Churning your software, APIs, instruction sets, etc. hurts your own developers and leads to time and bugs, and churning your proprietary connectors only keeps the chinese clone shops at bay for about a week. The future belongs to cryptographically enforced incompatibility laid on top of as much cheap, reliable, standardized stuff as you can get away with using...
What I find odd about Intel's phrasing is that they seem to be (voluntarily, no less) throwing water on their chances of selling this particular silicon to somebody who wants to run linux on it...
Unless there is something seriously whacked about this part compared to a conventional x86 part, it should require, at most, modest tweaks to get Linux up and going. Now, support for the fancy new power management features and whatnot may or may not be there(just as the world is infested with motherboards with ACPI so broken that all OSes that aren't the version of Windows they were designed to sell with might as well just not touch ACPI at all to the degree possible); but that is quite different from 'cannot run'.
It seems particularly odd given that Intel is fairly likely to, for simple reasons of efficiency, try to use similar power management tricks across their CPU lineup, from SoCs to high end Xeons, and it'll be a cold day in hell before the bulk Xeon buyers of the world accept a chip that won't run Linux, or vmkernel, or the assorted other non-windows things that are entirely routine in server land.
Also, isn't this the same Intel that fails to understand that ARM is going to be very important in the future? AFAIK they're the only ones that aren't licensing the technology.
Intel purchased 'StrongARM' from DEC ages ago(back when DEC still had things you could purchase), took it through a few generations under that name and then as 'Xscale', and then sold it to Marvell 6ish years ago(with the possible exception of one flavor that they use on their RAID boards, I can't remember).
They still have an ARM license, they just aren't terribly motivated to use it. x86 doesn't have too many friends; but it certainly has a lot of customers, and Intel has somewhat... limited... incentive to march into the business of being yet another SoC shover as long as they can get away with the margins on their x86s parts and supporting silicon.
iPod Touches definitely are iOS devices, just without the cellular modems and with a light dusting of contempt from Apple when it comes to adding the punchier new hardware... Estimates are that 40-ish percent of iOS devices are ipod touches.
what fragmentation? there are three phones supported by ios6
Which means that there are a few tens to hundreds of millions of actively used phones that will never even run the current OS... That's sort of what 'fragmentation' is all about...
To be sure, Android is substantially more fragmented even if you only look at 'currently-sold-and-supported-by-people-you've-heard-of', since there are multiple hardware OEMs shoving handsets out; but unless you start killswitching all your products, 'fragmentation' inevitably happens whenever people don't stop using the older ones and you introduce something new.
The chart in TFA is entirely useless, since it focuses on end user features that apple has or hasn't included on various models; but there is the not-so-minor matter of spec changes(TFA's chart doesn't even touch them; but 'keeping the core APIs consistent' also doesn't address them). There are some pretty significant differences in CPU and GPU power, and how quickly the OS will run out of RAM and quietly start memory-managing you, between those models.
If your 'app' is just some lousy re-implementation of a website that you really wanted to flog through the app store for some reason, it probably isn't a big deal; but anybody who really needs the punch provided by running native can't necessarily ignore that.
Likely to be a permanent condition.
Helium is light enough that it doesn't persist very well in the atmosphere(unlike the heavier noble gasses, that you can just distill out if the price gets high enough to pay for the energy needed), and it is only replenished quite slowly by alpha decay of assorted radioactives in the crust.
The only significant source is natural gas wells in proximity to suitable minerals over geologic time and equipped to capture the helium when the product is brought to the surface.
The Michigan proposal involved some assorted fancy-materials-science tricks(inks with very atypical optical properties and other stuff that the anti-counterfieting guys have been poking around at to raise the cost and required sophistication of producing a convincing fake) in addition to QR codes. If anything, the QR part seemed like something of a trend-crazed afterthought.
(Incididentally, the one thing that cryptography can do for physical items like currency is make it impossible for forgers to produce novel forgeries: If, for instance, the bill has a data field that is its serial number/place/time of manufacturer/etc. signed with a treasury private key, that doesn't stop me from just photocopying it; but it does prevent me from producing any bills that aren't direct copies of official bills. If combined with a reasonably effecient automated scanning system, to identify duplicates, this makes it more difficult to pass large numbers of copies of a single legitimate bill, and makes it impossible to produce anything other than copies of bills that were actually produced by the private key holder. Whether this is actually useful depends on what your mechanisms are for weeding out duplicates and tracing them back to their originators; but it can be done.)
Yeah yeah, it's got all kind of sciencey uses. Asteroids and stuff like that. But what are the military applications???
I'll let you ponder all the other fun stuff that you can track in the vicinity of the earth with a high-performance radar system...
In pure performance? Not particularly. In 'just supposed to work, and ships on some enormous percentage of all x86s churned out every year'? That's Intel.
Hey, cut them some slack. Lying in public is one of the few pleasures of having a customer base that consists of people who don't know better...
which would make excessively fast blinks useless as indicators for humans to read.
You could probably build something to read them. Assuming the LED's emitting the data have a fast enough rising and falling time.
That's basically how all optical networking not classy enough to afford laser diodes works(TOSLINK is probably the biggest example, wouldn't surprise me if somebody sold cheap-n-nasty GBICs that worked the same way at some point; but laser diodes have gotten pretty cheap...) However, since the indicator lights are intended for humans, apparently the chips driving higher speed interfaces(10mb ethernet and up) deliberately lengthen the on an off cycles(coincidentially cratering the bandwidth of the channel) in order to allow the human to distinguish between 'solid on' and 'blinking'. If the designs didn't intentionally do that, it is quite plausible that faster interfaces would leak through their indicators as well.
The RONJA guys coaxed 10mb out of unexceptional red LEDs across several hundred meters of free air, so the rise/fall times are apparently not insurmountable...
In that vein, the demo unit appears to be using red LEDs. I wonder if the phosphor-blobbed white ones are severely bandwidth constrained by the residual glow from the phosphor layer, or if the ghastly framerates of the cheapy cameras that intel proposes to use as receivers for these signals is the limiting factor?
It was demonstrated on low speed modems, some of which directly tied the blinkenlights to the serial traffic(including, if memory serves, an embarassing case where some model of fancy 'encrypted' modem tied the blinkenlights to the serial traffic before the encryption stage...); but the author concluded that status indicator lights on higher speed stuff were(either in response to the possibility of attack, or just because humans can't distinguish between ultrafast blinking and 'on', which would make excessively fast blinks useless as indicators) not usefully coupled to the data channel for anything more than vague inference about traffic volume...
I am a trifle surprised that nobody seems to have banged out a USB/bluetooth 'calculator keyboard' peripheral(external numpads are a dime a dozen; but I've never seen one with a scientific calculator's complement of operator symbols and things), since building HID devices to spit out whatever keycodes burned into their dinky little processors isn't a wildly expensive process; and would make phone or computer-based calculating a bit more comfortable.
As for dedicated calculators, though, it seems most unlikely: the second you get good enough to help little timmy skip learning in favor of data entry, you can kiss the educational market goodbye, and that is close to being the entire market for calculators, in volume terms... Test administrators can get touchy enough about the pathetic internal storage of the TI-83 and similar, they aren't likely to warm to something that has a radio data link and could trivially be farming out the problems on your math test over the internet to somewhere poor enough that "take a lazy first-worlder's SAT for $50" sounds like a really good idea even to those with actual math skills...
Some software(Photoshop being the big name; but not exclusive to them) also includes this 'feature'. If you manipulate an image of a major world currency in excessive detail, a neat little binary module included with photoshop will snag you and direct you to this rather bland organization.
The difference with digital camera watermarking is that EXIF is a (not always obvious depending on the UI, and sometimes less standard that it ought to be) standardized metadata storage system. The internet is rife with amusing mistakes made by people who don't know about exif and upload anyway; but that's a UI/user problem. The fields are well known, easily viewed and edited with commonly available software, and not designed to be covert or strip-resistant in any way. Some imaging devices are, quite arguably, excessively chatty by default, and that is a legitimate concern given user ignorance; but there isn't anything sneaky about the technology.
Watermarks, at least in this incarnation, are designed to be covert, strip-resistant, and are not intended for the creator of the image to be aware of.
This is a 'prisons and fortresses share certain architectural similarties; but do not share purposes' situation...
If memory serves, it isn't actually a factor of printer manufacturer(and/or re-badger); but of the OEM behind the color laser print engine. Apparently there are relatively few of those, and some, thanks to a little leaning from Uncle Sam the details of which have never come to light, include the watermarking 'feature' in all their print engines. Since printer manufacturers can, and sometimes do, switch parts suppliers between models, a given manufacturer might have both bugged and clean hardware on offer at a given time.
Blizzard actually poisons the kernel entropy pool so cleverly that 'random' behaviors by the computer end up leaking identifiable information. Very sneaky of them...
CALCnet allows networking of TI-83 and similar calculators with relatively simple external hardware.
With that detail out of the way, you are free to implement a display-wall and/or the most powerful z80 cluster computer in the known universe.
Extra credit, of course, will be awarded if you succeed in writing an xorg driver that can treat an MxN array of networked calculators as a greyscale display of appropriate resolution.