I would explicitly prohibit (either by changing the OS or through the review process) modal dialogs not belonging to an application from appearing over another application. This includes built-in OS modal dialogs.
Yes. The focus-stealing modal dialog is one of the dumbest, most egregious UX flaws in common implementations of the WIMP interface. It assumes the user is always looking at the screen and able to immediately shift from interacting with one application to interacting with another. It's a gaping security hole and a patent error in the user interaction model.
Windows solved this in the 1990s with ctrl-alt-del.
And it was already an old idea then. The use of a Secure Attention Key or Secure Attention Sequence goes back at least to the 1970s.
The problem with SAKs, as usual, is users. A SAK is only a useful security feature for a vigilant user who knows to avoid any prompt for credentials that wasn't elicited by a SAK. It's also an awkward user experience when applications need to request credentials for cases where they can't inherit the OS authentication.
Personally, I keep the SAK enabled on my Windows systems. But it doesn't prune much of the credential-stealing attack tree.
A more-comprehensive solution is dedicated hardware (e.g. a smartcard reader), so the user's credentials are never exposed to a general-purpose computer at all. But that requires sweeping changes in how password-based authentication is currently used by popular OSes and applications.
On a slightly serious note, I drive between Michigan and New Mexico several times a year, and now for family reasons nearly always take the Kansas route. (We used to avoid Kansas because five or six hours on I-70 through western Kansas and eastern Colorado is sheer torture.)
The Missouri leg is actually one of the best parts, before you get to the better parts of Colorado. Of course I don't take I-70 in Missouri, because that too is agonizing - almost nothing but traffic and billboards. There are some good places to eat, but the scenery is much less pretty than it is around the Ozarks. St Louis traffic is bad, and while KC is better, there are still some annoying stretches.
So I go north and take US-36 through Missouri, which is quite pleasant. It's mostly rural and much less heavily traveled, and far fewer billboards. It's a leg I look forward to, particularly after the five or six hundred hours on interstates in frickin' Illinois. (Illinois has two modes: Chicago, which has features but ghastly traffic; and everything else, which alternates between nothing and packs of obnoxious drivers.)
Similarly, when going through western Kansas, I now get off I-70 onto US-40 when it splits off, then go through rural Kansas and Colorado. The fewer miles of interstate, the better.
Forget the Missouri hyperloop. What we need is are hyperloops across Illinois and Kansas. Chunnel-style ones that ferry cars.
There's absolutely nothing wrong in starting a sentence with a conjunction, and the armchair pedants who insist on that sort of arbitrary and unhelpful prescription are the real foes of English usage.
Take your ruler, return to your desk, and shut up.
touchpads were inferior, but the modern ones just completely outperform any trackpoint implementation for speed and accuracy
When I think of the two or three seconds I lose each day using the Trackpoint... well, I don't fucking care.
When I think of the substandard accuracy of the Trackpoint - you know what? It's never been a problem. Ever. In the nearly 25 years I've been using them (started circa 1993).
If you hit the touchpad with your thumb while typing, why isn't your thumb on the space bar where it belongs?
"Yes, we put a control right where you'll hit it by accident. It's your fault if you can't learn to avoid it."
at 480p, if you wanted to do a landscape, you had to letterbox, now you only have 300 scan lines or less and you don't get the same quality.
Shrug. I have plenty of VHS videotapes that are letterboxed, and I continue to enjoy the content on them, because it includes good storytelling. As far as I'm concerned, the image only has to be good enough to be discernible.
I don't have a 4K television, and I won't be getting one until the current set dies, and then only if I can only find 4K sets that have the features I actually want. (Or, actually, the features I have to compromise on, since electronics companies insist on providing idiotic "smart" TVs. It'd be nice if I can continue to avoid this crap, but I don't have a lot of hope. May have to buy components and build my own.)
Because my gut tells me that you can probably permute the board and convert partially-filled n-queens problem to a "normalized" form that can be solved as efficiently as the empty board n-queens problem: "can this input be rearranged into a prefix substring of the normalized solution for NxN?"
Interesting idea, but JOTTOMH I'm concerned about the normalization step. Permutation grows super-polynomially, of course, so naive normalization won't get you out of NP-Complete.
Or from the other direction: assuming one-way functions exist, then starting with a normal-form partial-N-queens solution, shuffling into any of the non-normal permutations is poly-time, but reversing that shuffle looks at first glance like it might be orthogonal to some other one-way functions.
But that's based on, like, 30 seconds of thinking about it. So I'm probably missing something obvious.
Will be interesting to see whether Blum or somebody else can fix the proof.
Totally not my area, but my guess is no. The problem with Tardos's function looks pretty fatal. AIUI (from Aaronson's blog and some other sources), Blum was trying to extend superpolynomial lower bounds proofs on monotone Boolean functions to general Boolean functions, and Tardos showed (back in the '80s) that there's at least one such function that meets all the superpoly tests when implemented with a monotone circuit, but has a poly general circuit. So it's a counterexample to what Blum was trying to prove.
Of course I may be hilariously wrong about all of that.
In the same manner every P problem could probably be solved with an NP method
No "probably" about it; NP includes P, so every P-time solution is an NP-time solution. Or to put it another way, a nondeterministic Turing machine can behave deterministically; it just doesn't have to.
If a someone figures out a method to convert an NP problem into a P problem then it is likely that method only works on that particular type of NP problems and not all of them.
The "P vs NP" problem, technically, is whether all of NP is in P. That means a proof has to apply to NP-Complete problems, and any problem in NP can be transformed into an NP-Complete one in poly-time. So if you prove any NP-Complete problem can be solved in poly-time, your proof applies to all of NP. A valid constructive proof would give you a poly-time method for solving all NP problems.
However, as other posters have noted, poly-time doesn't mean "feasible". I remember Matt Skala once wrote about a P-time algorithm he'd invented/discovered[1] that he'd shown had a lower bound with some ridiculous exponent - over 100, anyway. And it was actually useful in theory, just not in practice for any input of significant size. Even a smallish exponent can put P-time solutions out of reach for interesting problems.
[1] Take your pick. I'm not interested in fighting with mathematical realists.
IRC doesn't let you see history without special bots and shit, it doesn't have built-in screen sharing or video, etc. I wish peolple would stop bringing up IRC as if it was really a competitor.
Indeed. These are all reasons why it's better than Slack [he wrote, stroking his grey beard and frowning at the children frolicking perilously close to his lawn].
'round these parts, we use RocketChat, which the Slack fans claim is a pale imitation. But I've seen some Slack demos and don't grasp the attraction. I'd take RC over Slack, IRC over RC, and Usenet over IRC. I've never found live messaging very useful.
Just think how dull Mysteries of Lowell would be with all the characters replaced by robots. (No doubt the well-educated/. readership is intimately acquainted with Mysteries of Lowell.)
It's nice to see the/. community coming together to heap abuse on it, though.
I'm tempted to submit a story. "Noted software developer and occasional author of Slashdot comments Michael Wojcik cut his fingernails last night. 'I suppose I typically cut them once a week or so,' the tech legend reported."
Not to mention "incase" as a compound rather than a phrase, "JavaScript and NodeJS [sic] are single-handedly" (what, one hand between the two of them?), and assorted other infelicities. But then "dubious summary by a marginal writer" is pretty much par for the course.
I remember in the 2000s, few people saw the value of smartphones.
Yes. And now in social situations you have everyone fucking around with their phones instead of paying attention to one another. Progress!
The transition to autonomous cars is likely inevitable, and it will ruin all the best things about driving. Exploring an area by car, for example.
Just a couple of days ago I did an 18-hour drive. In the past month I've put about 5500 miles on my car. And while that's not something I want to do all the time, I'm still glad it was me (or a family member for a small part of it) rather than a machine driving. When I don't want to control my travel, I have plenty of options: flights, taxis[1], trains (in some cases, as rail travel is pretty limited in the US), buses, etc. When I take my car, it's because I want to operate it.
[1] I don't use fake taxi services like Uber, either. I play by the rules and I'm not interested in rewarding those who don't.
My body is heavy, and getting older, so I wanted to be comfortable, so I have a car near 4k lb. Even so, 300HP is still plenty.
Indeed. I have a Volvo XC70 - an AWD wagon - which was purchased based on its suitability for long road trips that include some bad weather and rough roads (logging roads in the mountains and that sort of thing, not hard-core offroading).
Very comfortable seats, 4000lb curb weight, 300 HP. I've never come close to wanting more power, even towing a loaded utility trailer up mountain passes in Colorado and New Mexico. I can easily pass the folks with travel trailers.
I grew up driving little economy cars with less than 100 HP, and liked them just fine. Of course highway speed limits are much higher in the US now; but even so cars of today are hilariously over-powered.
I do wish the Volvo were available in this country with a manual transmission, though - just because I enjoy shifting.
Even gold only has value because people believe it does. Its only real value is as a conductor, but beyond that, it's pretty worthless.
Unlike fiat currencies, precious metals have natural scarcity. A government can't decide to start circulating ten times as much gold (unless little is already in circulation, and the government has huge reserves) and cause hyperinflation.
Note that I'm not an advocate of non-fiat currencies, just as I'm not an advocate of traveling everywhere on horseback. And deflation ain't fun either. But there is a difference between a currency that's reliably grounded in a scarce physical resource and one that simply holds value by consensus.
Now, a currency based on a commodity that is simply scarce and has no other appeal is likely to be very weak. One based on scarcity and other desirable properties - such as an industrial use (conductivity in the case of gold) and an ornamental one (which is very important in the case of gold, and I have to believe you're willfully ignoring it) - is stronger. One that was indispensable (e.g. oxygen as a currency in a space habitat) would have problems with hording, and would likely be used as barter.
All that said, we do have real-world examples of bottom-up fiat currencies, where communities simply come to a popular agreement to regard them as a medium of exchange, etc. It happened in Somalia, for one.
In American English, "three-month" would be the more common form of the adjectival phrase:
I engaged in a three-month English-usage flame war on/., to no avail.
The plural variant is, IME, common in British English, and may indeed be dominant there.
A quick Google Ngram Viewer check suggests the plural form was more common historically, but was displaced by the singular one around 1935, and the latter is now far more popular.
people already eat imitation-crab in their California rolls anyway
I think that's the problem. Between the real stuff at the top (which covers a large range of attributes, including price point) and surimi at the bottom (which is cheap and also comes in a wide variety of textures and flavor variations), there just isn't much room in the market for a middle-tier premium fish product.
The really expensive fish products are Veblen goods, so there's not much point in trying to undercut the price on those. The next tier is often chosen as much for presentation as anything else, so you'd have to make your cultured fish look right (including inedible fish parts) too. After that it's a race to the bottom.
But I could certainly be wrong. And sushi and sashimi have the advantage that buyers often never get to see the whole fish (or anything more than a tiny part of it), so the presentation aspect is displaced onto other aspects of the dish.
If something like vat-grown meat ever takes off, every farm animal in the country will be dead within a few years.
Highly unlikely, I'd say.
Because farmers don't raise cows and pigs and chickens because they enjoy their company, they raise them for income. Once the animals become unsellable, they're going to be exterminated.
There's a healthy market for meat from small, local farms, sold at local butcher shops and farmers' markets and the like. We get nearly all our meat (at both our Michigan and New Mexico homes) from that boutique market. It will not disappear overnight, or even in a generation.
Despite the best efforts of big agribusiness, small family farms continue to be viable in many parts of the country. There are plenty of affluent consumers who are willing to pay a high premium to patronize them. Whether that's rational behavior is debatable; but any behavioral economist will tell you that the economy is not driven by rational agents.
I'm willing to bet that the livestock pens at my local county fair will be just as well populated fifty years from now as they were this year, with animals personally raised by 4H and FFA kids, and then sold to the local butcher shops.
The ready availability of venison in the butcher shops, and even many supermarkets, hasn't slowed down deer hunting one bit. The same goes for fish in the markets and fishing. You think hunting and dressing your own deer, or catching and cleaning your own fish, is economically competitive with the meat counter? Only when you take intangibles into account - and the intangibles culturally attached to foodstuffs are many and potent.
High school economics have very little predictive power in general, and pretty much none when emotional considerations come into play.
Slashdot Mirror
I would explicitly prohibit (either by changing the OS or through the review process) modal dialogs not belonging to an application from appearing over another application. This includes built-in OS modal dialogs.
Yes. The focus-stealing modal dialog is one of the dumbest, most egregious UX flaws in common implementations of the WIMP interface. It assumes the user is always looking at the screen and able to immediately shift from interacting with one application to interacting with another. It's a gaping security hole and a patent error in the user interaction model.
Windows solved this in the 1990s with ctrl-alt-del.
And it was already an old idea then. The use of a Secure Attention Key or Secure Attention Sequence goes back at least to the 1970s.
The problem with SAKs, as usual, is users. A SAK is only a useful security feature for a vigilant user who knows to avoid any prompt for credentials that wasn't elicited by a SAK. It's also an awkward user experience when applications need to request credentials for cases where they can't inherit the OS authentication.
Personally, I keep the SAK enabled on my Windows systems. But it doesn't prune much of the credential-stealing attack tree.
A more-comprehensive solution is dedicated hardware (e.g. a smartcard reader), so the user's credentials are never exposed to a general-purpose computer at all. But that requires sweeping changes in how password-based authentication is currently used by popular OSes and applications.
On a slightly serious note, I drive between Michigan and New Mexico several times a year, and now for family reasons nearly always take the Kansas route. (We used to avoid Kansas because five or six hours on I-70 through western Kansas and eastern Colorado is sheer torture.)
The Missouri leg is actually one of the best parts, before you get to the better parts of Colorado. Of course I don't take I-70 in Missouri, because that too is agonizing - almost nothing but traffic and billboards. There are some good places to eat, but the scenery is much less pretty than it is around the Ozarks. St Louis traffic is bad, and while KC is better, there are still some annoying stretches.
So I go north and take US-36 through Missouri, which is quite pleasant. It's mostly rural and much less heavily traveled, and far fewer billboards. It's a leg I look forward to, particularly after the five or six hundred hours on interstates in frickin' Illinois. (Illinois has two modes: Chicago, which has features but ghastly traffic; and everything else, which alternates between nothing and packs of obnoxious drivers.)
Similarly, when going through western Kansas, I now get off I-70 onto US-40 when it splits off, then go through rural Kansas and Colorado. The fewer miles of interstate, the better.
Forget the Missouri hyperloop. What we need is are hyperloops across Illinois and Kansas. Chunnel-style ones that ferry cars.
Turn it back? The Apple II (well, the //e) was Peak Apple. It never got better than that.
There's absolutely nothing wrong in starting a sentence with a conjunction, and the armchair pedants who insist on that sort of arbitrary and unhelpful prescription are the real foes of English usage.
Take your ruler, return to your desk, and shut up.
touchpads were inferior, but the modern ones just completely outperform any trackpoint implementation for speed and accuracy
When I think of the two or three seconds I lose each day using the Trackpoint... well, I don't fucking care.
When I think of the substandard accuracy of the Trackpoint - you know what? It's never been a problem. Ever. In the nearly 25 years I've been using them (started circa 1993).
If you hit the touchpad with your thumb while typing, why isn't your thumb on the space bar where it belongs?
"Yes, we put a control right where you'll hit it by accident. It's your fault if you can't learn to avoid it."
It's not just incremental, it's a genuine attempt to make video almost indistinguishable from reality.
Well that sounds just awful.
at 480p, if you wanted to do a landscape, you had to letterbox, now you only have 300 scan lines or less and you don't get the same quality.
Shrug. I have plenty of VHS videotapes that are letterboxed, and I continue to enjoy the content on them, because it includes good storytelling. As far as I'm concerned, the image only has to be good enough to be discernible.
I don't have a 4K television, and I won't be getting one until the current set dies, and then only if I can only find 4K sets that have the features I actually want. (Or, actually, the features I have to compromise on, since electronics companies insist on providing idiotic "smart" TVs. It'd be nice if I can continue to avoid this crap, but I don't have a lot of hope. May have to buy components and build my own.)
Because my gut tells me that you can probably permute the board and convert partially-filled n-queens problem to a "normalized" form that can be solved as efficiently as the empty board n-queens problem: "can this input be rearranged into a prefix substring of the normalized solution for NxN?"
Interesting idea, but JOTTOMH I'm concerned about the normalization step. Permutation grows super-polynomially, of course, so naive normalization won't get you out of NP-Complete.
Or from the other direction: assuming one-way functions exist, then starting with a normal-form partial-N-queens solution, shuffling into any of the non-normal permutations is poly-time, but reversing that shuffle looks at first glance like it might be orthogonal to some other one-way functions.
But that's based on, like, 30 seconds of thinking about it. So I'm probably missing something obvious.
Will be interesting to see whether Blum or somebody else can fix the proof.
Totally not my area, but my guess is no. The problem with Tardos's function looks pretty fatal. AIUI (from Aaronson's blog and some other sources), Blum was trying to extend superpolynomial lower bounds proofs on monotone Boolean functions to general Boolean functions, and Tardos showed (back in the '80s) that there's at least one such function that meets all the superpoly tests when implemented with a monotone circuit, but has a poly general circuit. So it's a counterexample to what Blum was trying to prove.
Of course I may be hilariously wrong about all of that.
In the same manner every P problem could probably be solved with an NP method
No "probably" about it; NP includes P, so every P-time solution is an NP-time solution. Or to put it another way, a nondeterministic Turing machine can behave deterministically; it just doesn't have to.
If a someone figures out a method to convert an NP problem into a P problem then it is likely that method only works on that particular type of NP problems and not all of them.
The "P vs NP" problem, technically, is whether all of NP is in P. That means a proof has to apply to NP-Complete problems, and any problem in NP can be transformed into an NP-Complete one in poly-time. So if you prove any NP-Complete problem can be solved in poly-time, your proof applies to all of NP. A valid constructive proof would give you a poly-time method for solving all NP problems.
However, as other posters have noted, poly-time doesn't mean "feasible". I remember Matt Skala once wrote about a P-time algorithm he'd invented/discovered[1] that he'd shown had a lower bound with some ridiculous exponent - over 100, anyway. And it was actually useful in theory, just not in practice for any input of significant size. Even a smallish exponent can put P-time solutions out of reach for interesting problems.
[1] Take your pick. I'm not interested in fighting with mathematical realists.
IRC doesn't let you see history without special bots and shit, it doesn't have built-in screen sharing or video, etc. I wish peolple would stop bringing up IRC as if it was really a competitor.
Indeed. These are all reasons why it's better than Slack [he wrote, stroking his grey beard and frowning at the children frolicking perilously close to his lawn].
'round these parts, we use RocketChat, which the Slack fans claim is a pale imitation. But I've seen some Slack demos and don't grasp the attraction. I'd take RC over Slack, IRC over RC, and Usenet over IRC. I've never found live messaging very useful.
Sure, it can sew - but can it love?
Just think how dull Mysteries of Lowell would be with all the characters replaced by robots. (No doubt the well-educated /. readership is intimately acquainted with Mysteries of Lowell.)
It's nice to see the /. community coming together to heap abuse on it, though.
I'm tempted to submit a story. "Noted software developer and occasional author of Slashdot comments Michael Wojcik cut his fingernails last night. 'I suppose I typically cut them once a week or so,' the tech legend reported."
Not to mention "incase" as a compound rather than a phrase, "JavaScript and NodeJS [sic] are single-handedly" (what, one hand between the two of them?), and assorted other infelicities. But then "dubious summary by a marginal writer" is pretty much par for the course.
I remember in the 2000s, few people saw the value of smartphones.
Yes. And now in social situations you have everyone fucking around with their phones instead of paying attention to one another. Progress!
The transition to autonomous cars is likely inevitable, and it will ruin all the best things about driving. Exploring an area by car, for example.
Just a couple of days ago I did an 18-hour drive. In the past month I've put about 5500 miles on my car. And while that's not something I want to do all the time, I'm still glad it was me (or a family member for a small part of it) rather than a machine driving. When I don't want to control my travel, I have plenty of options: flights, taxis[1], trains (in some cases, as rail travel is pretty limited in the US), buses, etc. When I take my car, it's because I want to operate it.
[1] I don't use fake taxi services like Uber, either. I play by the rules and I'm not interested in rewarding those who don't.
My body is heavy, and getting older, so I wanted to be comfortable, so I have a car near 4k lb. Even so, 300HP is still plenty.
Indeed. I have a Volvo XC70 - an AWD wagon - which was purchased based on its suitability for long road trips that include some bad weather and rough roads (logging roads in the mountains and that sort of thing, not hard-core offroading).
Very comfortable seats, 4000lb curb weight, 300 HP. I've never come close to wanting more power, even towing a loaded utility trailer up mountain passes in Colorado and New Mexico. I can easily pass the folks with travel trailers.
I grew up driving little economy cars with less than 100 HP, and liked them just fine. Of course highway speed limits are much higher in the US now; but even so cars of today are hilariously over-powered.
I do wish the Volvo were available in this country with a manual transmission, though - just because I enjoy shifting.
...They is a good rational for doing this...
No, there isn't.
But is there a good transcendent? Perhaps an imaginary?
It's a temporary blimp
I believe you'll find it's a temporary dirigible, due to its internal structure.
Even gold only has value because people believe it does. Its only real value is as a conductor, but beyond that, it's pretty worthless.
Unlike fiat currencies, precious metals have natural scarcity. A government can't decide to start circulating ten times as much gold (unless little is already in circulation, and the government has huge reserves) and cause hyperinflation.
Note that I'm not an advocate of non-fiat currencies, just as I'm not an advocate of traveling everywhere on horseback. And deflation ain't fun either. But there is a difference between a currency that's reliably grounded in a scarce physical resource and one that simply holds value by consensus.
Now, a currency based on a commodity that is simply scarce and has no other appeal is likely to be very weak. One based on scarcity and other desirable properties - such as an industrial use (conductivity in the case of gold) and an ornamental one (which is very important in the case of gold, and I have to believe you're willfully ignoring it) - is stronger. One that was indispensable (e.g. oxygen as a currency in a space habitat) would have problems with hording, and would likely be used as barter.
All that said, we do have real-world examples of bottom-up fiat currencies, where communities simply come to a popular agreement to regard them as a medium of exchange, etc. It happened in Somalia, for one.
In American English, "three-month" would be the more common form of the adjectival phrase:
I engaged in a three-month English-usage flame war on /., to no avail.
The plural variant is, IME, common in British English, and may indeed be dominant there.
A quick Google Ngram Viewer check suggests the plural form was more common historically, but was displaced by the singular one around 1935, and the latter is now far more popular.
people already eat imitation-crab in their California rolls anyway
I think that's the problem. Between the real stuff at the top (which covers a large range of attributes, including price point) and surimi at the bottom (which is cheap and also comes in a wide variety of textures and flavor variations), there just isn't much room in the market for a middle-tier premium fish product.
The really expensive fish products are Veblen goods, so there's not much point in trying to undercut the price on those. The next tier is often chosen as much for presentation as anything else, so you'd have to make your cultured fish look right (including inedible fish parts) too. After that it's a race to the bottom.
But I could certainly be wrong. And sushi and sashimi have the advantage that buyers often never get to see the whole fish (or anything more than a tiny part of it), so the presentation aspect is displaced onto other aspects of the dish.
CARP (CARP Ain't Restricted or Patented)
Ugh, that's terrible. Someone come up with a better cultured-meat recursive acronym.
If something like vat-grown meat ever takes off, every farm animal in the country will be dead within a few years.
Highly unlikely, I'd say.
Because farmers don't raise cows and pigs and chickens because they enjoy their company, they raise them for income. Once the animals become unsellable, they're going to be exterminated.
There's a healthy market for meat from small, local farms, sold at local butcher shops and farmers' markets and the like. We get nearly all our meat (at both our Michigan and New Mexico homes) from that boutique market. It will not disappear overnight, or even in a generation.
Despite the best efforts of big agribusiness, small family farms continue to be viable in many parts of the country. There are plenty of affluent consumers who are willing to pay a high premium to patronize them. Whether that's rational behavior is debatable; but any behavioral economist will tell you that the economy is not driven by rational agents.
I'm willing to bet that the livestock pens at my local county fair will be just as well populated fifty years from now as they were this year, with animals personally raised by 4H and FFA kids, and then sold to the local butcher shops.
The ready availability of venison in the butcher shops, and even many supermarkets, hasn't slowed down deer hunting one bit. The same goes for fish in the markets and fishing. You think hunting and dressing your own deer, or catching and cleaning your own fish, is economically competitive with the meat counter? Only when you take intangibles into account - and the intangibles culturally attached to foodstuffs are many and potent.
High school economics have very little predictive power in general, and pretty much none when emotional considerations come into play.
One luck guy did ate the burger and it was extremely expensive to make.
I heard it had too much katsup.