People are actually using email in two quite distinct ways; but one way is faster at doing what we thought everyone used email for, and is therefor better? Cool.
Frankly, this sounds like a challenge for team search: computers are very good indeed, even with the quite basic desktop search mechanisms, not the fancy search engine stuff, at assorted glorified greps. You want all the emails that mention project X, or were sent by Mr. Y? No problem. You want to know when project X needs to be finished? Well, get all emails mentioning project X and start exploring the exciting universe of different natural language ways of suggesting that project X needs to be finished. Search isn't completely useless; but you've basically gone back to filing...
I've seen a few hints of this in Gmail, which will pick out emails that appear to obviously be appointments or date/time combinations and offer to add them to your calendar; but further expansion would be nice. Aside from the people who are just conceptually crippled, it seems unlikely that users are sorting their emails into folders just because doing electronic shit work is all fun and giggles. They are likely doing it because search can't(or the advanced search features that can, they can't use) organize their email for them in the way they prefer it to be. Let's see a software agent that starts picking out salient topics, and piecing together a slightly creepy knowledge of it by watching your mailstream(and FFS, let's make it client side, or based on servers you control, not some you are a peon in the cloud plantation shit...)
~$300, won't handle the fancy card graphics and embossed numerals("Magnetic Stripe Card Embossers" are used for that, also perfectly licit off-the-shelf items); but will turn a card blank into something that an automated POS won't bat an eye at(and, in most cases, re-using a bank-issued card, even if the number on the card doesn't match the one on the stripe, should probably escape a retail employee's notice).
Magnetic card stock is also a legitimate off-the-shelf item, as are printers that will dump an arbitrary color image onto blanks(entirely non-suspicious, any organization that issues mag-stripe IDs probably has such a printer on the shelf somewhere.) Getting a card-stock supplier to do a large print run of cards identical to bank blanks would probably raise some eyebrows; so you would presumably have to steal or print your own.
Everything you need to produce fully functional magnetic stripe cards is fully licit, available off the shelf, and not particularly expensive. The only "secret" is the name and number prominently displayed on actual issued credit cards, and handed over during each transaction. The "chip and PIN" stuff is horribly broken; but at least it pretends to be concerned about card cloning...
I despise how these cases get treated as "identity theft" rather than "bank/CC fraud with a side of impersonation". An "identity" as it is presently constructed for financial purposes, is basically all public, or near-public information(much of it is public record, the rest is simultaneously treated as Super Secret Proof, and demanded, all the time, by basically everybody, because it is Super Secret Proof, which of course means that it is basically public, like SSNs and CC numbers...) It isn't the person whose "identity" is used to perpetrate a given frauds fault that financial institutions can't be bothered to actually verify transactions properly, although the poor bastards often get stuck with years of hassle for it anyway.
The notion of "identity theft" seems like nothing more than a cynical way to shift responsibility away from the responsible parties, and the parties who could do something about it(hey, Visa, don't want my CC getting cloned by anybody who manages to obtain the numbers visible in plaintext on the card, which have to be used to perform a transaction? Try cryptography...) and onto the suckers at the bottom of the food chain who, realistically, have very little control over the 'security' that a bunch of nearly public information connected to them is given by the large number of people who have access to it.
Hardware write-blockers are pretty much standard for any hands-on forensics not conducted by utter amateurs. The trouble is, of course, that you can only use those after you seize the hardware, and the feds want something they can use before they seize the hardware.
I don't know if analogous US malware tools(Magic Lantern, CIPAV, possibly others) have been studied in as much detail; and they may or may not be of higher quality; but anything that has to run on the live host system isn't going to be fundamentally less capable of modifying that system.
The problem is that, once you've had malware on the system, all the write-blocker can do is assure you that nothing was tampered with during the forensic investigation, which provides no protection against tampering by the malware, and likely makes anything planted by it appear to be much higher quality evidence...
The piece of incompetence that I find really striking is not so much the general shoddiness; but the fact that the malware is using a proxy setup in the US to avoid having its traffic traced back to the German police entity using it. Even if they know nothing about the tech side of things, surely exporting the evidence outside of the state, country, and EU, to some random datacenter in the US, would mean a hairy pile of privacy and chain-of-custody problems for the chaps in legal?
While your general point is valid: against targeted attackers the ratios for "desktops cracked, by platform" are pretty irrelevant"; there is more to it:
A game console, many smartphones, tivos, etc. do checks of the OSes they run. If the signature doesn't check, the device doesn't boot. Better implemenations(newer xbox360s, for instance, pretty much have to be voltage glitched to get past that.
If you are going to be strapping some hellfire missiles to something, you really, really shouldn't be running an OS/architecture so stock that desktop or corporate penetration and bug numbers are terribly relevant...
Loyalty is considered to be a virtue for a reason(hominid life, up until the past few thousand years, and still in many places up to the present day has basically been a case of 'iterative prisoner's dilemma'. As it turns out, being a good guy by default, and only shafting the other guy if he has a history of shafting people works out for everyone fairly well.)
So, here's the question: if you job, by virtue of size/holding structure of company, psychological profile of leader figures, etc. is still small enough that its behavior is largely governed by "human" heuristics, loyalty can pay off. They will know who their loyal people are, value that, and your long-term payoff(especially if the product launch goes well) is likely to be good.
If the company is larger than a certain size, run by sociopaths, or otherwise no longer governed by conventional human logic, the management will still recognize "loyal" employees; but by "loyal" they mean "sucker who will stay around for more punishment, for illogical emotional reasons, until we suck him dry and throw his husk away". Bad situation...
That's the real trick. Being loyal to people is usually a pretty good idea. Being loyal to an organization or sociopath who considers you a "human resource" and your "loyalty" to be a form of primitive emotional weakness that makes you easier to exploit is always a terrible, terrible plan.
If your employer would(hypothetically), tell you to clean out your desk and instruct security not to let the door hit your worthless ass on the way out if you were to get sick and be expected to be less productive because of treatment/recovery for a period of time, then it is a fairly safe bet that you are just an "input" to them. If so, fuck-em. They'd fuck you over for money, and it looks like you've been handed the change to do unto them before they do unto you.
If, in that same hypothetical situation, they would exhibit care, understanding, concern, accomodation, etc, it is probable that they are the sort of entity that will recognize, value, and reciprocate loyalty...
It's terrible, I tell you, somebody isbiased against having their personal information sold to who-the-fuck-knows who, without their consent, unless they manage to respond to an email, likely sent to a nonexistent address, that will probably be written in a style that is calculated to throw spamassasin into a killing rage, because he may or may not have implicitly 'agreed' to a contract of adhesion at some point in time.
Why can't somebody be more fair and balanced?
Oh, wait, of course; because there is no "other side" to this story, just customers getting shafted.
Less of an obstacle than it might be...(assuming there is some way of testing tweaks in 'ugly mode' rather than recompiling every time you want to make any change...)
If 128GB is needed for
Such kit, while a bit excessive if you just want to game on the cheap, is not exactly science fiction anymore, and may well be an "I already have one, incremental cost $0" item for a gaming enthusiast.
For those with less patience, a 16-core(2 socket) Poweredge R715 with 128GB of RAM will run ~$5,500. If you need the cores, an R815 with 32 cores across four sockets is only ~$7,300 with 128GB(rising to a slightly alarming ~$25,000 if you go with top end CPUs and 512GB of RAM, those 16GB RDIMMS are Not Cheap).
Obviously, most prospective modders aren't going to have a ~$5,500 server kicking around; but 8-24GB is "enthusiast gaming machine" material these days, and 128GB, two sockets, is well under $10,000(particularly if redundancy, uptime, and all that irksome enterprise stuff can be ignored because its a hobby machine.) It does cut down on the pool of potential modders a bit; but people have spent substantially more on hobbies no less weird...
That's one of the things that gives it such a spammy feel:
Marketing scumdroids constantly think, and talk, about people in terms of the potentially-'monetized' demographic slice they are part of. Actual humans tend to talk about things that they are or aren't interested in. These two views are not strictly incommensurate with one another(a group of people with the same interests and non interests is a demographic slice, and a member of a demographic slice will have interests and non interests); but it reeks of marketing weasel when somebody says "X is a must have for aspirational Gen-Xers attempting to recapture their fading youth", rather than just expressing love of red convertibles...
There almost certainly are people who, by some fair measures, are "web browser fanatics"; but nobody who is would use the phrase like that.
For the right price, I'm sure we could work something out; but it might cause some difficulty in meeting your desired 'impulse purchase' price point...
In the field of pixel-addressable e-ink screened devices, $80 is very aggressive pricing. It would appear that they have chosen to go with the "make it cheaper" option, (which advances in technology generally provide as an alternative to "make it better"), for this particular kindle.
Only the sales figures will say for sure whether it was a good idea; but encheapening the hell out of the lowest-end dedicated conduit to your gigantic electronic store doesn't seem like an obviously crazy strategy...
It wouldn't be a terribly helpful permission, for the user-knowledge/caring reasons you rightly point out.
Architecturally, though, I think that there would be a case to be made that "localhost only" and "internet only, excluding localhost" are logical subdivisions(not mutually exclusive, an application could request both); because many applications need only communication with remote hosts, and aren't necessarily to be trusted crawling over localhost behind the firewall, and others might have a legitimate need for some sort of IPC; but needn't be trusted with the ability to exfiltrate whatever they are doing.
None of this, of course, would solve the immediate problem, which is that HTC decided to ship a diagnostic application with root access and a homebrew access mechanism a zillion times worse than rlogin(seriously, HTC, ssh is free, and supports keypair authentication...)
More generally, though, I'd say that I'm in favor of more rather than less granularity in permissions structures: If you have a very, very, fine-grained permission system, and need to make it simpler for the end user, 'bundling' groups of highly-granular rights into user-friendly lumps is comparatively easy(and you can still take full advantage of the granularity in situations where you want to get fancy with the control exercised over portions of the system that the end user shouldn't need to mess with). If you have a very coarse permission system, and it turns out that you need something tighter, tacking that on after the fact is not necessarily going to be pretty...
It is quite possible that the "Facebook Beacon" 'feature' that debuted I-don't-even-remember-how-many privacy controversies ago, was simply the public face of the installation of the eye-pyramids in each location...
Short of discovering that each of Facebook's datacenters is actually a vast, nearly empty, mausoleum, lit only by the unhallowed glow of Masonic runes drawn in the blood of innocents and the blinding glare of the all-seeing-eye atop the pyramid in the center; could there be any revelation about their privacy practices worse than those that can already be inferred from prior activity?
Obviously, with his newfound UV-vision powers, he is ready to decode(or manipulate) the hidden coloration used by plants to attract bees...
As HiveLord, numberless swarms of eusocial attack insects will bend to his will! The crops of man shall be bounteous, or wither unpollinated, by his hand! His amazonian suicide warriors will throw themselves at all foes, laying down their lives that the Swarm's venom may find its target!
Arguably there is a problem with "the permissions"; but not in a narrowly technical sense(well, strictly speaking, it might be nice if Android broke network permissions down a little further, so that you could allow an application to access internet resources; but forbid it from connecting to anything on localhost, or allow something to connect to one or more ports on localhost; but not the outside...)
A major vendor is shipping a 'diagnostic' application so fucked that it might as well be a rootkit on a large-but-not-precisely-known number of devices expected to be connected to the internet and in possession of relatively juicy information for most of their operational lives, and nobody in the chain decided that this was maybe a bad idea until 3rd parties discovered it and wrote it up...
This suggests that HTC's "Sense" team might not have any.
Slashdot Mirror
In the future, your primitive "filters" will be replaced by an IBM Watson site licence...
People are actually using email in two quite distinct ways; but one way is faster at doing what we thought everyone used email for, and is therefor better? Cool.
Frankly, this sounds like a challenge for team search: computers are very good indeed, even with the quite basic desktop search mechanisms, not the fancy search engine stuff, at assorted glorified greps. You want all the emails that mention project X, or were sent by Mr. Y? No problem. You want to know when project X needs to be finished? Well, get all emails mentioning project X and start exploring the exciting universe of different natural language ways of suggesting that project X needs to be finished. Search isn't completely useless; but you've basically gone back to filing...
I've seen a few hints of this in Gmail, which will pick out emails that appear to obviously be appointments or date/time combinations and offer to add them to your calendar; but further expansion would be nice. Aside from the people who are just conceptually crippled, it seems unlikely that users are sorting their emails into folders just because doing electronic shit work is all fun and giggles. They are likely doing it because search can't(or the advanced search features that can, they can't use) organize their email for them in the way they prefer it to be. Let's see a software agent that starts picking out salient topics, and piecing together a slightly creepy knowledge of it by watching your mailstream(and FFS, let's make it client side, or based on servers you control, not some you are a peon in the cloud plantation shit...)
Why am I not surprised to hear that PayPal is doing something evil, again?
Wrong search terms: "Magnetic Stripe Encoders" are what you are looking for.
~$300, won't handle the fancy card graphics and embossed numerals("Magnetic Stripe Card Embossers" are used for that, also perfectly licit off-the-shelf items); but will turn a card blank into something that an automated POS won't bat an eye at(and, in most cases, re-using a bank-issued card, even if the number on the card doesn't match the one on the stripe, should probably escape a retail employee's notice).
Magnetic card stock is also a legitimate off-the-shelf item, as are printers that will dump an arbitrary color image onto blanks(entirely non-suspicious, any organization that issues mag-stripe IDs probably has such a printer on the shelf somewhere.) Getting a card-stock supplier to do a large print run of cards identical to bank blanks would probably raise some eyebrows; so you would presumably have to steal or print your own.
Everything you need to produce fully functional magnetic stripe cards is fully licit, available off the shelf, and not particularly expensive. The only "secret" is the name and number prominently displayed on actual issued credit cards, and handed over during each transaction. The "chip and PIN" stuff is horribly broken; but at least it pretends to be concerned about card cloning...
I despise how these cases get treated as "identity theft" rather than "bank/CC fraud with a side of impersonation". An "identity" as it is presently constructed for financial purposes, is basically all public, or near-public information(much of it is public record, the rest is simultaneously treated as Super Secret Proof, and demanded, all the time, by basically everybody, because it is Super Secret Proof, which of course means that it is basically public, like SSNs and CC numbers...) It isn't the person whose "identity" is used to perpetrate a given frauds fault that financial institutions can't be bothered to actually verify transactions properly, although the poor bastards often get stuck with years of hassle for it anyway.
The notion of "identity theft" seems like nothing more than a cynical way to shift responsibility away from the responsible parties, and the parties who could do something about it(hey, Visa, don't want my CC getting cloned by anybody who manages to obtain the numbers visible in plaintext on the card, which have to be used to perform a transaction? Try cryptography...) and onto the suckers at the bottom of the food chain who, realistically, have very little control over the 'security' that a bunch of nearly public information connected to them is given by the large number of people who have access to it.
Hardware write-blockers are pretty much standard for any hands-on forensics not conducted by utter amateurs. The trouble is, of course, that you can only use those after you seize the hardware, and the feds want something they can use before they seize the hardware.
I don't know if analogous US malware tools(Magic Lantern, CIPAV, possibly others) have been studied in as much detail; and they may or may not be of higher quality; but anything that has to run on the live host system isn't going to be fundamentally less capable of modifying that system.
The problem is that, once you've had malware on the system, all the write-blocker can do is assure you that nothing was tampered with during the forensic investigation, which provides no protection against tampering by the malware, and likely makes anything planted by it appear to be much higher quality evidence...
The piece of incompetence that I find really striking is not so much the general shoddiness; but the fact that the malware is using a proxy setup in the US to avoid having its traffic traced back to the German police entity using it. Even if they know nothing about the tech side of things, surely exporting the evidence outside of the state, country, and EU, to some random datacenter in the US, would mean a hairy pile of privacy and chain-of-custody problems for the chaps in legal?
There had better be a very good reason why 'in-app purchases from 3rd parties' are somehow not similar to buying stuff through a web browser...
While your general point is valid: against targeted attackers the ratios for "desktops cracked, by platform" are pretty irrelevant"; there is more to it:
A game console, many smartphones, tivos, etc. do checks of the OSes they run. If the signature doesn't check, the device doesn't boot. Better implemenations(newer xbox360s, for instance, pretty much have to be voltage glitched to get past that.
If you are going to be strapping some hellfire missiles to something, you really, really shouldn't be running an OS/architecture so stock that desktop or corporate penetration and bug numbers are terribly relevant...
Loyalty is considered to be a virtue for a reason(hominid life, up until the past few thousand years, and still in many places up to the present day has basically been a case of 'iterative prisoner's dilemma'. As it turns out, being a good guy by default, and only shafting the other guy if he has a history of shafting people works out for everyone fairly well.)
So, here's the question: if you job, by virtue of size/holding structure of company, psychological profile of leader figures, etc. is still small enough that its behavior is largely governed by "human" heuristics, loyalty can pay off. They will know who their loyal people are, value that, and your long-term payoff(especially if the product launch goes well) is likely to be good.
If the company is larger than a certain size, run by sociopaths, or otherwise no longer governed by conventional human logic, the management will still recognize "loyal" employees; but by "loyal" they mean "sucker who will stay around for more punishment, for illogical emotional reasons, until we suck him dry and throw his husk away". Bad situation...
That's the real trick. Being loyal to people is usually a pretty good idea. Being loyal to an organization or sociopath who considers you a "human resource" and your "loyalty" to be a form of primitive emotional weakness that makes you easier to exploit is always a terrible, terrible plan.
If your employer would(hypothetically), tell you to clean out your desk and instruct security not to let the door hit your worthless ass on the way out if you were to get sick and be expected to be less productive because of treatment/recovery for a period of time, then it is a fairly safe bet that you are just an "input" to them. If so, fuck-em. They'd fuck you over for money, and it looks like you've been handed the change to do unto them before they do unto you.
If, in that same hypothetical situation, they would exhibit care, understanding, concern, accomodation, etc, it is probable that they are the sort of entity that will recognize, value, and reciprocate loyalty...
"I'm Seth. Just... Seth. From God, to Kain, to Seth. I am his right hand and I have a task for you."
It's terrible, I tell you, somebody isbiased against having their personal information sold to who-the-fuck-knows who, without their consent, unless they manage to respond to an email, likely sent to a nonexistent address, that will probably be written in a style that is calculated to throw spamassasin into a killing rage, because he may or may not have implicitly 'agreed' to a contract of adhesion at some point in time.
Why can't somebody be more fair and balanced?
Oh, wait, of course; because there is no "other side" to this story, just customers getting shafted.
Less of an obstacle than it might be...(assuming there is some way of testing tweaks in 'ugly mode' rather than recompiling every time you want to make any change...)
If 128GB is needed for
Such kit, while a bit excessive if you just want to game on the cheap, is not exactly science fiction anymore, and may well be an "I already have one, incremental cost $0" item for a gaming enthusiast.
For those with less patience, a 16-core(2 socket) Poweredge R715 with 128GB of RAM will run ~$5,500. If you need the cores, an R815 with 32 cores across four sockets is only ~$7,300 with 128GB(rising to a slightly alarming ~$25,000 if you go with top end CPUs and 512GB of RAM, those 16GB RDIMMS are Not Cheap).
Obviously, most prospective modders aren't going to have a ~$5,500 server kicking around; but 8-24GB is "enthusiast gaming machine" material these days, and 128GB, two sockets, is well under $10,000(particularly if redundancy, uptime, and all that irksome enterprise stuff can be ignored because its a hobby machine.) It does cut down on the pool of potential modders a bit; but people have spent substantially more on hobbies no less weird...
Gaming press interprets tech demo as game, responds accordingly.
Sounds like the perfect chip upon which to build a modernized version of the coolest looking computer ever
Has this guy considered a career with the DEA?
That's one of the things that gives it such a spammy feel:
Marketing scumdroids constantly think, and talk, about people in terms of the potentially-'monetized' demographic slice they are part of. Actual humans tend to talk about things that they are or aren't interested in. These two views are not strictly incommensurate with one another(a group of people with the same interests and non interests is a demographic slice, and a member of a demographic slice will have interests and non interests); but it reeks of marketing weasel when somebody says "X is a must have for aspirational Gen-Xers attempting to recapture their fading youth", rather than just expressing love of red convertibles...
There almost certainly are people who, by some fair measures, are "web browser fanatics"; but nobody who is would use the phrase like that.
For the right price, I'm sure we could work something out; but it might cause some difficulty in meeting your desired 'impulse purchase' price point...
In the field of pixel-addressable e-ink screened devices, $80 is very aggressive pricing. It would appear that they have chosen to go with the "make it cheaper" option, (which advances in technology generally provide as an alternative to "make it better"), for this particular kindle.
Only the sales figures will say for sure whether it was a good idea; but encheapening the hell out of the lowest-end dedicated conduit to your gigantic electronic store doesn't seem like an obviously crazy strategy...
It wouldn't be a terribly helpful permission, for the user-knowledge/caring reasons you rightly point out.
Architecturally, though, I think that there would be a case to be made that "localhost only" and "internet only, excluding localhost" are logical subdivisions(not mutually exclusive, an application could request both); because many applications need only communication with remote hosts, and aren't necessarily to be trusted crawling over localhost behind the firewall, and others might have a legitimate need for some sort of IPC; but needn't be trusted with the ability to exfiltrate whatever they are doing.
None of this, of course, would solve the immediate problem, which is that HTC decided to ship a diagnostic application with root access and a homebrew access mechanism a zillion times worse than rlogin(seriously, HTC, ssh is free, and supports keypair authentication...)
More generally, though, I'd say that I'm in favor of more rather than less granularity in permissions structures: If you have a very, very, fine-grained permission system, and need to make it simpler for the end user, 'bundling' groups of highly-granular rights into user-friendly lumps is comparatively easy(and you can still take full advantage of the granularity in situations where you want to get fancy with the control exercised over portions of the system that the end user shouldn't need to mess with). If you have a very coarse permission system, and it turns out that you need something tighter, tacking that on after the fact is not necessarily going to be pretty...
It is quite possible that the "Facebook Beacon" 'feature' that debuted I-don't-even-remember-how-many privacy controversies ago, was simply the public face of the installation of the eye-pyramids in each location...
"Another great read, a must for Web browser fanatics!"
Seriously? Could you sound any more astroturfy if you tried?
Short of discovering that each of Facebook's datacenters is actually a vast, nearly empty, mausoleum, lit only by the unhallowed glow of Masonic runes drawn in the blood of innocents and the blinding glare of the all-seeing-eye atop the pyramid in the center; could there be any revelation about their privacy practices worse than those that can already be inferred from prior activity?
Obviously, with his newfound UV-vision powers, he is ready to decode(or manipulate) the hidden coloration used by plants to attract bees...
As HiveLord, numberless swarms of eusocial attack insects will bend to his will! The crops of man shall be bounteous, or wither unpollinated, by his hand! His amazonian suicide warriors will throw themselves at all foes, laying down their lives that the Swarm's venom may find its target!
Arguably there is a problem with "the permissions"; but not in a narrowly technical sense(well, strictly speaking, it might be nice if Android broke network permissions down a little further, so that you could allow an application to access internet resources; but forbid it from connecting to anything on localhost, or allow something to connect to one or more ports on localhost; but not the outside...)
A major vendor is shipping a 'diagnostic' application so fucked that it might as well be a rootkit on a large-but-not-precisely-known number of devices expected to be connected to the internet and in possession of relatively juicy information for most of their operational lives, and nobody in the chain decided that this was maybe a bad idea until 3rd parties discovered it and wrote it up...
This suggests that HTC's "Sense" team might not have any.