Why exactly can't you make me a "real" laptop inexpensively. Give me a cheaper, but laptop-friendly CPU.... VIA C3 or low-speed mobile chip.
It drives me nuts that I can't even choose to make that compromise, because it's a choice I'd gladly make. In the bottom tier of laptops, I can either buy an overpowered machine that slows down on battery (Dell, Toshiba, etc.), an overpowered machine without a battery (ECS Desknote), or a used machine from Uncle Slimeball's Computers and 100 Big Shrimp with a battery that appears to have been trampled by a wild hamster stampede.
A very similar UI exists with the awful BASIC in the TRS 80 Model 1, Level 1. This is the little unpleasant 4k machine that requires a propriatery B&W monitor and looks like a Commodore dipped in paint. I guess the intent was to seem human-like-familiar and chatty, but it just led to frustration for me. The box is sitting somewhere.
>I mean seriously: The screen is fuzzy, it's not >significantly cheaper than a full-blown PC, and >running it on the XBOX adds nothing to running a >normal PC with it.
Look at all the people buying cute little C3 boards to stick in tiny little cases. The XBox is essentially the same thing, but pre-tested, pre-assembled and ready to go, and sells for a decent price.
I can see a very nice application for a console running arbitrary code-- the "Custom TV show". Videotex-style displays of info I want from the Web (Weather, latest news, finance reports), perhaps augmented with stuff brought in from sensors ("It is now 24.1C in the dining room and 31.1C in the computer room"), and perhaps a relay-based system (x10?) to control devices around the house.
I think it's a 'poison-pill' clause, intended to discourage a certain behaviour.
Software patents make it theoretically possible to have software that's libre free (in terms of the copyrighted material) but still restricted (in terms of patents). Big example: GIF, MP3, and now JPEG patents.
A clause like this makes it expensive for a patent-holder to try to enforce software patents against Free Software. They might get a few thousand bucks in settlements, assuming their claims aren't thrown out, but at the cost of losing the right to redistribute software they may need for their business, or even products that make up part of their product line.
I like it, in a way. The Free Software movement can't really offer cross-licencing agreements, or even really buy patent licences, but now they have something they can use for leverage.
1. The usual "Okay, we'll get it right this time" mentality you can get from a new architecture.
2. The mention of stability in the press release. Nice to see that. Am I the only one who cares about the machine remaining up?
3. Likely low price, cool running (if it's on a small process and relatively low-performance).
4. Cool name. Why do so few processors have cool names? The K6-III had a brilliant codename ('Sharptooth' IIRC) and a thoroughly boring market name.
5. It may oppress me politically, but at least it probably won't give me a fuss about ripping MP3s.:)
However, I have to say to the marketers: Give me an interesting market niche. It's almost overkill for most embedded needs, and underkill for a "normal" desktop. Perhaps they can deliver a sub-$8000 Internet Fridge (don't laugh, LG makes one) Or given their server market play, a new supply for when all the old P1xx machines you would have turned into light-duty servers get sold off.
Right now, the IP industry is based on *recovery* of costs... create the software/CD/film, then try to reclaim your expenses in sales. How about some more proactive approaches?
"We're looking to do something innovative. Who wants to cough up a Hamilton to ensure it happens?"
"At the design phase of our project, we can still inject any changes people want, for a price... Want the Evil Basilisk of Death to look suspiciously like your 7th-grade English teacher? 50 bucks."
Ideally, your udon noodles and Athlon XPs should be paid for before you put the last } in.
There are times when you don't want to cut out the tumor until you're sure it won't grow into a pair of wings.
The current application of P2P may be considered offensive to some, but who knows what the future holds? I can see two extremely valuable uses for P2P technology if allowed to flourish and expand its use.
1. The difficult-to-silence distributed source of legitimate, but "touchy" material. Wonder if putting the Bill of Rights on Gnutella would get a lot of downloads from.cn sites? 2. Distributed backup. I'll trade off a few Gb of my hard disc, and in exchange, I can get my data encrypted and stored remotely on multiple redundant machines.
Not quite true. What stops you from putting the credit card info on a box on your INTERNAL network. Then the machine connected to the Internet can only pick the card numbers off through a set of constrained methods. I'm thinking either a system that logs every transaction with pertinent info (He's requesting the card number for site.com account No. 50601 with a 'send to' address of '550 Foo Street, Bazville') or one that just fails the request if it lacks correct tokens ("He wants the card number of Joe Blow, site.com account number 2606, but his account number is actually 5081-- fail the request and possibly log it!)
The problem with the 'alcohol in prohibition' analogue is that monitoring doesn't scale up well.
First, consider if you have five million tractor-trailers a day going down a stretch of distribution network highway. Now try to match the contents of every single trailer against a list of forbidden items of indefinite length and that changes daily. And note that the trailers marked "Ace Illegal Booze Limited" may actually contain bottles of soda, and vice versa.
Our rights are affected because the "solutions" proposed will throw out babies (legitimate applications and fair use) with bathwaters.
For point 2, the basic idea is that they license the code GPL, but still hold onto the trademarks such as the programme name. Much like you'd probably get in trouble if you hacked Mozilla and tried to pass it off as Netscape, why couldn't Epson just go after people distributing modified versions in a manner designed to confuse them with the originals. I used Info-ZIP as an example because their licence explicitly mentioned not applying their product names to modified versions, but I think it could be done with ordinary trademark-protection practices.
On your rebuttal to point 3, the difficulty of modifying a binary-only programme varies widely, depending on how you want to cause problems. It doesn't have to be subtle to cause grief, and in fact, if your hack isn't readily detectable (for example, crippling the 2400dpi mode isn't a biggie if you can't discern above 600), it would defeat its purpose. You could spread a nice bit of FUD just by making a package with garbled or virus-infected binaries, or one lacking key parts of the product.
1. If you want Epson's drivers, wouldn't you probably get them from a reliable source like, oh... Epson? Or your distribution supplier of choice?
2. If nothing else, they could licence the trademarks for the unmodified version only so that modified versions can't be confused with the original (I believe there's a line like this in the Info-ZIP licence.)
3. What's to stop someone from mangling a binary-only driver with the old search 'n replace, and creating similar problems? A quick change to a "set-resolution" or "change mode" command could cause just as much hassle.
>Each processor will have its own, unique >public/private key pair. Each processor's public >key will be digitally signed by a manufacturer's >private key, whose public key will be signed by >Microsoft's private key.
That's not where the weak link lies. The weak link is that, when you get down far enough, the key has to be represented in some tangible format (the bridges/fuses/PROM).
With regards to the keyspace size, I will refer to the similar cases with Ethernet MAC numbers and IP addresses... large blocks are allocated to manufacturers/firms, and their internal subpartitioning can eat up a lot more space.
The value of any size key drops if you're burning bits on things like manufacturer and model IDs.
Final thought: The ability to kill valid keys on the fly seems like a recipe for massive infighting. How long before Company X tries to get Company Y's keys revoked as a form of industrial sabotague? Or worse yet, the wholesale confusion if a block of 50,000 processors used in retail machines are suddenly marked as 'compromised', preventing Joe Sixpack from using Spyware Player Deluxe 2007?
I see a minor flaw in this-- how will these chips be manufactured?
Method 1: All chips use the same private key because they're identical internally. No benefit.
Method 2: A set of bridges or pins set the key at final-assembly-time. We all know how well that works to lock AMD multipliers.
Method 3: There are internal connections set (I think this is how the P4 multiplier lock is done) This is the most troubling one, but probably still possible to decipher... particularly if someone can throw a lot of effort and money into the system. Perhaps this explains all the job postings here: "Hardware engineer wanted. Must speak fluent Arabic, be willing to relocate, have long beard.":)
Also, I can see two logistical problems to derail the system:
1. The "AMD mole" angle. "You're kidding! Every single one of our stepping-DB0 processors has an embedded key of 0xdeadbeef? That was a mistake. Sorry."
2. The finite keyspace angle. What happens when you simply have no more keys, and all of a sudden, some guy gets access to all sorts of goodies because his key has been reused on another processor.
Re:Used Printers? You get what you pay for.
on
Printer Makers' Ploys
·
· Score: 2, Insightful
Actually, a good place to look is a thrift shop... you can often plug them in and run the test sequence to see if all is well.
I purchased a Brother HL-645M (very dirty output) for $2.50 at a rummage sale, a nice NEC Silentwriter 95 for $10 at a thrift shop, and an Okidata OL-400e (with really bad feed problems) for $15... together tiding me over until I could get a new Samsung 1210 on sale for USD 100 after rebate.
Simple question: if everyone subscribes to the simple "one media bill that covers access to everything", how long before they say "Oh... to serve our bankbooks better, let's charge you $50 a month instead of $10."
If nothing else, the "collection" of music/tangible media phenomenon only allows them to charge you once per piece of content obtained.
OTOH, I would pay into a pool that provided stipends for professional artists to release their work into the public domain.
Even those working perfectly within the law suffer. For example, the "rights manglement" systems often work by locking the content to a specific player. If you want to let your friend borrow it, it won't work on his machine, or even worse, it won't play on your replacement for a broken/obsolete machine. Combined with attempts to limit the ability to backup content, or place it in a more reliable/flexible/transferrable form, the goal is clear: create demand for new product that was never there before by making the product un-resellable and fragile.
Furthermore, there's the negative repercussion of a central licencing clearinghouse... most people don't want to deal with getting authorization before using content (and, how exactly will you activate my.mp5s if I'm in the middle of Nunavut?), and it opens a wide range of potential for monitoring.
Why can't they have the voter swipe their driver's licence (or even a voter's registration card with an added magnetic stripe?) at the voting kiosk before voting? This could then be cross-referenced with a list of legitimate (registered, alive, not convicted of felony) voters for this voting site (stored on the local machine), and then against a list of people who have already voted at that site (by putting each of the machines on a local network).
Disadvantadge: you would have trouble voting at a non-local precinct, but this is a problem with all voting systems.
They're saying that they aren't buying as many or more discs. That doesn't mean they're buying NONE. Important distinction.
If 6.27% of the market is buying less discs due to downloading, the total sales will drop.0627 * X
where X is the proportion of CDs they're not buying that they would have otherwise. If X is 0.1 or 0.5 even, it's insufficient to blame the big nasty Internet boogeyman for this.
2. Create an infrastructure to buy creative work by volume, and offer it at cost of reproduction. I have to assume this will be a gov't project, with a similar principle to the free-library concept-- supply information to the masses.
Because of part 1, nobody will ever be able to sell more than one copy of a work reliably, so they may as well sell it to the distribution infrastructure instead of trying to find buyer No. 1.
I have been informed that a variant of such a system exists in some countries (.nl IIRC) -- the state is a buyer of last resort.
I wouldn't be so confident in professional certification. While it may work better in gems, a similar story was promoted for coins and other investment/collectibles, and resulted in the following:
-Dubious certification services appear; in some cases, these services were owned/controlled by the very people whose goods they were certifying. In some cases, you'd have a harder time selling an item certified by one of those authorities than if the dealer alone quoted a grade.
-Standards shifted over time, even with well-regarded services. The coin you got certified as "MS-65" and worth $200 in 1992 is now a "MS-67" worth $4000, or a "MS-63" worth $20. This problem is particularly bad when an item is between two levels with significantly different values.
-Technical grades didn't always accomodate eye-appeal well. One "XF-40" might look an order of magnitude better than another due to differences in the damage. If pricing or trading is supposed to be done sight-unseen using the grades, this can be trouble. I suspect this will be the biggie with stones.
The problem wasn't the 8086-- as you said, there's nothing you can do. The problem comes up when you move to a CPU that allows more than 1Mb. Because people bypassed the BIOS and OS routines with great regularity, and due to cards with ROM and/or registers that mapped into the 640-1024k region, you can't just move the ROMs and device memory up (the most desirable option).
What surprises me is that we still haven't developed some sort of alternative... even if that means discarding the lowest 1Mb of a 512Mb system because of its complex internals.
Just because one resolution is the prettiest doesn't mean the others are useless. Many times, apps that have a naturally small display can use the resolution set down to make them fill the screen. Good examples are emulators and TV tuner apps (fundamentally limited to the 525 lines of the NTSC standard), and some games (nobody will ever want to play the original Linux Doom binaries at > 640x480!)
If it's supposed to limit distribution, it's a remarkably ineffective way to go about it.
First, aren't there a relatively small number of regions? (8?) That doesn't make for very fine-grained control, plus, once the regions are set, you're stuck with them, so it can't be used to say "This release goes to countries A, B, and C, but this one goes to A and C and perhaps D but not B." Unless the industry knew exactly what regions would be needed for all time, this will not be a suitable solution.
Second, the region control ties it not to a specific country (could we have that technology-- perhaps, by putting a GPS in every player), but a specific set of players. If you're desperate enough to watch a Region X disc in Region Y, what stops you from importing a Region X player (and, if necessary, a compatible TV?) I believe this has been done occasionally to bridge the TV standard barrier with video tape.
Third, I may be mistaken, but isn't the region code not honoured by all players? I'm thinking some older players didn't handle it.
I feel like these people are trying to say "We locked the bank vault" when they applied duct tape and a shrinkwrap licence to the door.
Just because Windows/Outlook/whatever $LATEST_VERSION is secure/reliable/whatever, doesn't mean a whole lot. The massive installed base means anything broken in an old version will stay a problem for a long time. Not to neglect that any security setting can always be turned down by a user who is greeted by a message that says "I am Ana Kournikova, I'm naked, and I sent you this document to have your advice, so please turn down your security settings so you can see it properly"
Slashdot Mirror
Why exactly can't you make me a "real" laptop inexpensively. Give me a cheaper, but laptop-friendly CPU.... VIA C3 or low-speed mobile chip.
It drives me nuts that I can't even choose to make that compromise, because it's a choice I'd gladly make. In the bottom tier of laptops, I can either buy an overpowered machine that slows down on battery (Dell, Toshiba, etc.), an overpowered machine without a battery (ECS Desknote), or a used machine from Uncle Slimeball's Computers and 100 Big Shrimp with a battery that appears to have been trampled by a wild hamster stampede.
A very similar UI exists with the awful BASIC in the TRS 80 Model 1, Level 1. This is the little unpleasant 4k machine that requires a propriatery B&W monitor and looks like a Commodore dipped in paint. I guess the intent was to seem human-like-familiar and chatty, but it just led to frustration for me. The box is sitting somewhere.
>I mean seriously: The screen is fuzzy, it's not >significantly cheaper than a full-blown PC, and >running it on the XBOX adds nothing to running a >normal PC with it.
Look at all the people buying cute little C3 boards to stick in tiny little cases. The XBox is essentially the same thing, but pre-tested, pre-assembled and ready to go, and sells for a decent price.
I can see a very nice application for a console running arbitrary code-- the "Custom TV show". Videotex-style displays of info I want from the Web (Weather, latest news, finance reports), perhaps augmented with stuff brought in from sensors ("It is now 24.1C in the dining room and 31.1C in the computer room"), and perhaps a relay-based system (x10?) to control devices around the house.
They don't want to be held to anything now, so as to avoid setting precedent.
If they said today "Okay... want to stream, send a cheque for $x to PO Box y...", they'd have a hard time defending it if they said, later:
"We've decided the appropriate streaming fee is $3x"
"We're losing $2x per unauthorised stream occuring"
or
"They're all pirating b*****s", which is promptly responded to with the display of a cancelled cheque for $x.
I think it's a 'poison-pill' clause, intended to discourage a certain behaviour.
Software patents make it theoretically possible to have software that's libre free (in terms of the copyrighted material) but still restricted (in terms of patents). Big example: GIF, MP3, and now JPEG patents.
A clause like this makes it expensive for a patent-holder to try to enforce software patents against Free Software. They might get a few thousand bucks in settlements, assuming their claims aren't thrown out, but at the cost of losing the right to redistribute software they may need for their business, or even products that make up part of their product line.
I like it, in a way. The Free Software movement can't really offer cross-licencing agreements, or even really buy patent licences, but now they have something they can use for leverage.
Things I like from what I've seen:
:)
1. The usual "Okay, we'll get it right this time" mentality you can get from a new architecture.
2. The mention of stability in the press release. Nice to see that. Am I the only one who cares about the machine remaining up?
3. Likely low price, cool running (if it's on a small process and relatively low-performance).
4. Cool name. Why do so few processors have cool names? The K6-III had a brilliant codename ('Sharptooth' IIRC) and a thoroughly boring market name.
5. It may oppress me politically, but at least it probably won't give me a fuss about ripping MP3s.
However, I have to say to the marketers: Give me an interesting market niche. It's almost overkill for most embedded needs, and underkill for a "normal" desktop. Perhaps they can deliver a sub-$8000 Internet Fridge (don't laugh, LG makes one) Or given their server market play, a new supply for when all the old P1xx machines you would have turned into light-duty servers get sold off.
Perhaps you need a different business model.
Right now, the IP industry is based on *recovery* of costs... create the software/CD/film, then try to reclaim your expenses in sales. How about some more proactive approaches?
"We're looking to do something innovative. Who wants to cough up a Hamilton to ensure it happens?"
"At the design phase of our project, we can still inject any changes people want, for a price... Want the Evil Basilisk of Death to look suspiciously like your 7th-grade English teacher? 50 bucks."
Ideally, your udon noodles and Athlon XPs should be paid for before you put the last } in.
There are times when you don't want to cut out the tumor until you're sure it won't grow into a pair of wings.
.cn sites?
The current application of P2P may be considered offensive to some, but who knows what the future holds? I can see two extremely valuable uses for P2P technology if allowed to flourish and expand its use.
1. The difficult-to-silence distributed source of legitimate, but "touchy" material. Wonder if putting the Bill of Rights on Gnutella would get a lot of downloads from
2. Distributed backup. I'll trade off a few Gb of my hard disc, and in exchange, I can get my data encrypted and stored remotely on multiple redundant machines.
Not quite true. What stops you from putting the credit card info on a box on your INTERNAL network. Then the machine connected to the Internet can only pick the card numbers off through a set of constrained methods. I'm thinking either a system that logs every transaction with pertinent info (He's requesting the card number for site.com account No. 50601 with a 'send to' address of '550 Foo Street, Bazville') or one that just fails the request if it lacks correct tokens ("He wants the card number of Joe Blow, site.com account number 2606, but his account number is actually 5081-- fail the request and possibly log it!)
The problem with the 'alcohol in prohibition' analogue is that monitoring doesn't scale up well.
First, consider if you have five million tractor-trailers a day going down a stretch of distribution network highway. Now try to match the contents of every single trailer against a list of forbidden items of indefinite length and that changes daily. And note that the trailers marked "Ace Illegal Booze Limited" may actually contain bottles of soda, and vice versa.
Our rights are affected because the "solutions" proposed will throw out babies (legitimate applications and fair use) with bathwaters.
For point 2, the basic idea is that they license the code GPL, but still hold onto the trademarks such as the programme name. Much like you'd probably get in trouble if you hacked Mozilla and tried to pass it off as Netscape, why couldn't Epson just go after people distributing modified versions in a manner designed to confuse them with the originals. I used Info-ZIP as an example because their licence explicitly mentioned not applying their product names to modified versions, but I think it could be done with ordinary trademark-protection practices.
On your rebuttal to point 3, the difficulty of modifying a binary-only programme varies widely, depending on how you want to cause problems. It doesn't have to be subtle to cause grief, and in fact, if your hack isn't readily detectable (for example, crippling the 2400dpi mode isn't a biggie if you can't discern above 600), it would defeat its purpose. You could spread a nice bit of FUD just by making a package with garbled or virus-infected binaries, or one lacking key parts of the product.
Let's see:
1. If you want Epson's drivers, wouldn't you probably get them from a reliable source like, oh... Epson? Or your distribution supplier of choice?
2. If nothing else, they could licence the trademarks for the unmodified version only so that modified versions can't be confused with the original (I believe there's a line like this in the Info-ZIP licence.)
3. What's to stop someone from mangling a binary-only driver with the old search 'n replace, and creating similar problems? A quick change to a "set-resolution" or "change mode" command could cause just as much hassle.
>Each processor will have its own, unique >public/private key pair. Each processor's public >key will be digitally signed by a manufacturer's >private key, whose public key will be signed by >Microsoft's private key.
That's not where the weak link lies. The weak link is that, when you get down far enough, the key has to be represented in some tangible format (the bridges/fuses/PROM).
With regards to the keyspace size, I will refer to the similar cases with Ethernet MAC numbers and IP addresses... large blocks are allocated to manufacturers/firms, and their internal subpartitioning can eat up a lot more space.
The value of any size key drops if you're burning bits on things like manufacturer and model IDs.
Final thought: The ability to kill valid keys on the fly seems like a recipe for massive infighting. How long before Company X tries to get Company Y's keys revoked as a form of industrial sabotague? Or worse yet, the wholesale confusion if a block of 50,000 processors used in retail machines are suddenly marked as 'compromised', preventing Joe Sixpack from using Spyware Player Deluxe 2007?
I see a minor flaw in this-- how will these chips be manufactured?
:)
Method 1: All chips use the same private key because they're identical internally. No benefit.
Method 2: A set of bridges or pins set the key at final-assembly-time. We all know how well that works to lock AMD multipliers.
Method 3: There are internal connections set (I think this is how the P4 multiplier lock is done) This is the most troubling one, but probably still possible to decipher... particularly if someone can throw a lot of effort and money into the system. Perhaps this explains all the job postings here: "Hardware engineer wanted. Must speak fluent Arabic, be willing to relocate, have long beard."
Also, I can see two logistical problems to derail the system:
1. The "AMD mole" angle. "You're kidding! Every single one of our stepping-DB0 processors has an embedded key of 0xdeadbeef? That was a mistake. Sorry."
2. The finite keyspace angle. What happens when you simply have no more keys, and all of a sudden, some guy gets access to all sorts of goodies because his key has been reused on another processor.
Actually, a good place to look is a thrift shop... you can often plug them in and run the test sequence to see if all is well.
I purchased a Brother HL-645M (very dirty output) for $2.50 at a rummage sale, a nice NEC Silentwriter 95 for $10 at a thrift shop, and an Okidata OL-400e (with really bad feed problems) for $15... together tiding me over until I could get a new Samsung 1210 on sale for USD 100 after rebate.
Simple question: if everyone subscribes to the simple "one media bill that covers access to everything", how long before they say "Oh... to serve our bankbooks better, let's charge you $50 a month instead of $10."
If nothing else, the "collection" of music/tangible media phenomenon only allows them to charge you once per piece of content obtained.
OTOH, I would pay into a pool that provided stipends for professional artists to release their work into the public domain.
Even those working perfectly within the law suffer. For example, the "rights manglement" systems often work by locking the content to a specific player. If you want to let your friend borrow it, it won't work on his machine, or even worse, it won't play on your replacement for a broken/obsolete machine. Combined with attempts to limit the ability to backup content, or place it in a more reliable/flexible/transferrable form, the goal is clear: create demand for new product that was never there before by making the product un-resellable and fragile.
.mp5s if I'm in the middle of Nunavut?), and it opens a wide range of potential for monitoring.
Furthermore, there's the negative repercussion of a central licencing clearinghouse... most people don't want to deal with getting authorization before using content (and, how exactly will you activate my
Why can't they have the voter swipe their driver's licence (or even a voter's registration card with an added magnetic stripe?) at the voting kiosk before voting? This could then be cross-referenced with a list of legitimate (registered, alive, not convicted of felony) voters for this voting site (stored on the local machine), and then against a list of people who have already voted at that site (by putting each of the machines on a local network).
Disadvantadge: you would have trouble voting at a non-local precinct, but this is a problem with all voting systems.
They're saying that they aren't buying as many or more discs. That doesn't mean they're buying NONE. Important distinction.
.0627 * X
If 6.27% of the market is buying less discs due to downloading, the total sales will drop
where X is the proportion of CDs they're not buying that they would have otherwise. If X is 0.1 or 0.5 even, it's insufficient to blame the big nasty Internet boogeyman for this.
How would I change things?
1. Eliminate all restrictions on distribution.
2. Create an infrastructure to buy creative work by volume, and offer it at cost of reproduction. I have to assume this will be a gov't project, with a similar principle to the free-library concept-- supply information to the masses.
Because of part 1, nobody will ever be able to sell more than one copy of a work reliably, so they may as well sell it to the distribution infrastructure instead of trying to find buyer No. 1.
I have been informed that a variant of such a system exists in some countries (.nl IIRC) -- the state is a buyer of last resort.
I wouldn't be so confident in professional certification. While it may work better in gems, a similar story was promoted for coins and other investment/collectibles, and resulted in the following:
-Dubious certification services appear; in some cases, these services were owned/controlled by the very people whose goods they were certifying. In some cases, you'd have a harder time selling an item certified by one of those authorities than if the dealer alone quoted a grade.
-Standards shifted over time, even with well-regarded services. The coin you got certified as "MS-65" and worth $200 in 1992 is now a "MS-67" worth $4000, or a "MS-63" worth $20. This problem is particularly bad when an item is between two levels with significantly different values.
-Technical grades didn't always accomodate eye-appeal well. One "XF-40" might look an order of magnitude better than another due to differences in the damage. If pricing or trading is supposed to be done sight-unseen using the grades, this can be trouble. I suspect this will be the biggie with stones.
The problem wasn't the 8086-- as you said, there's nothing you can do. The problem comes up when you move to a CPU that allows more than 1Mb. Because people bypassed the BIOS and OS routines with great regularity, and due to cards with ROM and/or registers that mapped into the 640-1024k region, you can't just move the ROMs and device memory up (the most desirable option).
What surprises me is that we still haven't developed some sort of alternative... even if that means discarding the lowest 1Mb of a 512Mb system because of its complex internals.
Just because one resolution is the prettiest doesn't mean the others are useless. Many times, apps that have a naturally small display can use the resolution set down to make them fill the screen. Good examples are emulators and TV tuner apps (fundamentally limited to the 525 lines of the NTSC standard), and some games (nobody will ever want to play the original Linux Doom binaries at > 640x480!)
If it's supposed to limit distribution, it's a remarkably ineffective way to go about it.
First, aren't there a relatively small number of regions? (8?) That doesn't make for very fine-grained control, plus, once the regions are set, you're stuck with them, so it can't be used to say "This release goes to countries A, B, and C, but this one goes to A and C and perhaps D but not B." Unless the industry knew exactly what regions would be needed for all time, this will not be a suitable solution.
Second, the region control ties it not to a specific country (could we have that technology-- perhaps, by putting a GPS in every player), but a specific set of players. If you're desperate enough to watch a Region X disc in Region Y, what stops you from importing a Region X player (and, if necessary, a compatible TV?) I believe this has been done occasionally to bridge the TV standard barrier with video tape.
Third, I may be mistaken, but isn't the region code not honoured by all players? I'm thinking some older players didn't handle it.
I feel like these people are trying to say "We locked the bank vault" when they applied duct tape and a shrinkwrap licence to the door.
Just because Windows/Outlook/whatever $LATEST_VERSION is secure/reliable/whatever, doesn't mean a whole lot. The massive installed base means anything broken in an old version will stay a problem for a long time. Not to neglect that any security setting can always be turned down by a user who is greeted by a message that says "I am Ana Kournikova, I'm naked, and I sent you this document to have your advice, so please turn down your security settings so you can see it properly"