...a bill that is crazy priced and has a full color Goatse under complaint dept...
I'd think the Goatse guy would be a great choice for a customer complaints person. He's obviously wide open and accepting of criticism, can clearly handle anything you throw at him, and you know he can't help but give a shit, whether he wants to or not...
I think that'd be a terrible idea. The reason Debian works as a distribution base is because it's a complete, centralised, internally consistent and _working_ distribution guided by principles of end-user (admin) choice and unusually strict guidelines for package quality and interoperability, which make it fairly easy to drop new packages into the mix from random sources, mix-and-match packages already in the pool, and pre-decide things for your specific user base.
And you most certainly can build packages that'll work on every Debian-derived system, with a bit of care, and providing you don't need to rely on any packages whose ABIs have been forked downstream or have simply aged and been discarded from the repositories.
Sure, if you build something on Debian/unstable, you have no expectation that it'll work on Debian/stable, but if you build on the oldest distribution you care to support, a litte bit of effort with your dependancies should get you working on nearly-as-old distributions from other distributors, and you'll only be stopped from working on newer distributions when they don't provide a library ABI you need, in which case you need to build a version that works with _that_ ABI and so on.
Except where downstream distributions have created their own rules for shared library package naming, of course.
I don't believe that people and computers will be able to safe share roads. Here's my reasoning.
The major cause of accidents on the roads is surprise. If everyone drives the way everyone else expects then the road is a very safe place to be indeed.
Hence, our road safety largely relies upon drivers being able to successfully predict what other drivers (and technically, landmarks and animals) are going to do.
It's one of the main principles of driving: Don't surprise others, don't be surprised by others. If you don't know what someone is going to do, keep your distance.
You don't (as I've heard people be advised) assume that everyone around you is an idiot just waiting to pull a U-turn and drive head-on into you, because there's no way to drive safely under such a set of assumptions.
It's been statistically shown (i.e. I've been shown graphs) that going 10km/h faster or slower than the surrounding traffic makes you 10 times more likely to be involved in a multi-vehicle collision. Which makes sense if you think about it. If everyone's going in the same speed and the same direction, you cannot physically hit anyone else, your vectors do not intersect.
Now to the computers. The basic reasoning is this: Take your general population. Consider all those who are competent drivers. Now intersect that group with those who are able to successfully predict what a computer can do, and know it. That intersection will be the only group of people safely able to share roads with a computer. But the first group will be the ones who are allowed to do so.
On the other hand, computers _should_ be able to predict what the other computers are gong to do. Particularly with near-range wireless communications, they can simply _ask_ each other what they're seeing, and what they expect to do.
I think the best bet is to have highways built (or lanes declared and partitioned as) automation-only zones, and let people enter the highway under computer control only. Once on the highway, road convoys can self-assemble for speed/fuel efficiency purposes or not, but it's all computer controlled, so there's no drivers either acting unpredictable, or incorrectly predicting other vehicle's movements.
As a reward for reading this far: I did RTFA and I know this isn't actually what the article's about. In fact, I think the article's describing a really good and workable idea which I wholeheartedly support.
It's just something that's been weighing on my mind since having recently taken a motorcycle learner's course, (and I have other relevant background) and some of the comments on this article (who clearly haven't RTFA and think we're talking about a computer-driven convoy) are wildly mistaken about things like how road safety works, how this idea works, and in fact what this idea is.
I would assume they mean "incorrect judgements made during the process of driving in reaction to other drivers and changing road conditions", rather than "any error made in the process of transitioning from zygote to road kill".
Especially since they're addressing the former, not the latter.
Except where a development methodology interferes with the ability to evaluate and feedback into the design, which is usually the case for larger games projects.
Short turnaround time on "is X fun?" is absolutely vital to being able to apply a design philosophy and end up with something that's actually fun on time and on-budget (or at least close enough that whoever's responsible doesn't call time on you before you're done)
Back when intercontinental leased lines were all the rage, it was the case that a nightly financial data transfer from (I believe) a stock exchange trading floor was cheaper and faster done by loading the data onto tape and flying someone by Concorde from the UK to the US, than to transmit the data over the network.
That's both anecdotal and marred by my own recollection of the story, but it supports the "never underestimate the bandwidth of a truck full of backup tapes on the highway" saying.
What I'd like to know is, can any app read any file from an SD card if it knows the path of an existing file? From a previous Slashdot story (a few months back, cannot find the link) I had thought each Android application directory on an SD card was somehow isolated, but for this flaw to work at all that cannot be the case. What is to stop a rogue app from accessing any arbitrary application data at a known path on the SD card?
When you put applications on the SD card, their binary directories are isolated from each other, yes. (Through encrypted loop-mounts, I believe)
But the actual data on the sdcard is completely open to all applications. It's basically a large dumping ground for data.
The issue of this exploit is that you never need to grant anything permission to become vulnerable, whereas a rouge app does need to be given permission to be installed, and (I believe) permission to access data on the SD card.
I say "fork-resistant" not because they've actively blocked being forked, but because when forks do happen, they rarely survive as independent projects, either dying or rolling back into the mainline.
I'd count GCC as one of the latter, because although egcc did become the main-line, we didn't see (that I understand, anyway) a wholesale supplanting of project management and control the way we did with X.org forking from XFree86.
I honestly have no idea whom Apache was forked from, or the circumstances thereof.
I'd like to request substantive support for the parent post.
Games development can have great advantages over non-games development (irrespective of cubicles) but it's posts like the grandparent that can scare people away from an otherwise fulfilling career.
Sure, you probably end up trading in potential salary, but if you find the right studio and right team for you, it's worth the pay cut.
Regards, from someone who took a $15k/year pay cut to join the games industry almost half a decade ago, and is still thrilled to be making video games for a living.
I've been told that the Tier 1 RvR was the most fun of the lot. And the Empire area worked better than the Dwarven area for Tier 1 RvR. So it's probably the best choice of something to make free to play.
On the other hand, I personally quite enjoyed Tier 2 and Tier 3 keep sieges in both attack and defense, although the boss fights at the end tended to drag on a bit. Quite hard to find more than one group an evening that'd do it though, and once the populations got unbalanced, there was often no keeps available to be taken for days at a stretch.
It's 100% reliable for sites in a vulnerable configuration (ie. the sites the attack targets).
They presumably mean this to distinguish from attacks which require some level of luck or coincidence (eg. triggering a race condition) to exploit.
Or for attacks which require a certain state to be achieved but which there's no useful upper bound on when that state might be achieved due to the actions of the attack, eg. a purely passive attack.
So in short, they're saying that the attack is able to drive it's own progress towards its goal, and (I guess) always makes some small amount of progress on every iteration.
From reading the article, I believe what he means is power redundancy.
Specifically, right now, if the burglar cuts power to the home or business, the 48V (or so...) on the line lets the alarm system ring home and tell the company the power's out.
If only the were to put some kind of battery backup option for the ONT...
Oh, and eBay will always have cheaper devices. They generally ship from Hong Kong, offer no warranty or support, and illegally bypass local tariffs and taxes. They usually ship as personal mail with an outrageously low declared value. Legally, you are supposed to declare such purchases but nobody does.
That obviously depends on your local jurisdiction. Where I come from, personal imports below a certain amount legally bypass local tariffs and taxes, and above a certain amount, it seems to be at the discretion of customs as to whether you pay or not.
Be wary of anyone who misdeclares value or nature of goods when shipping to you, that would (probably) interfere with any shipping-related insurance claim you make.
Slashdot Mirror
...a bill that is crazy priced and has a full color Goatse under complaint dept...
I'd think the Goatse guy would be a great choice for a customer complaints person. He's obviously wide open and accepting of criticism, can clearly handle anything you throw at him, and you know he can't help but give a shit, whether he wants to or not...
I think that'd be a terrible idea. The reason Debian works as a distribution base is because it's a complete, centralised, internally consistent and _working_ distribution guided by principles of end-user (admin) choice and unusually strict guidelines for package quality and interoperability, which make it fairly easy to drop new packages into the mix from random sources, mix-and-match packages already in the pool, and pre-decide things for your specific user base.
And you most certainly can build packages that'll work on every Debian-derived system, with a bit of care, and providing you don't need to rely on any packages whose ABIs have been forked downstream or have simply aged and been discarded from the repositories.
Sure, if you build something on Debian/unstable, you have no expectation that it'll work on Debian/stable, but if you build on the oldest distribution you care to support, a litte bit of effort with your dependancies should get you working on nearly-as-old distributions from other distributors, and you'll only be stopped from working on newer distributions when they don't provide a library ABI you need, in which case you need to build a version that works with _that_ ABI and so on.
Except where downstream distributions have created their own rules for shared library package naming, of course.
Ubuntu is to Debian as Stalin is to Lenin?
And check the preview. >_<
Dammit, forgot to log in. >_
Time for ./'ers
to start referring to themselves by the correct handle?
Unless you're calling to arms people who cut dots up with knives and wear extraneous apostrophes?
I don't believe that people and computers will be able to safe share roads. Here's my reasoning.
The major cause of accidents on the roads is surprise. If everyone drives the way everyone else expects then the road is a very safe place to be indeed.
Hence, our road safety largely relies upon drivers being able to successfully predict what other drivers (and technically, landmarks and animals) are going to do.
It's one of the main principles of driving: Don't surprise others, don't be surprised by others. If you don't know what someone is going to do, keep your distance.
You don't (as I've heard people be advised) assume that everyone around you is an idiot just waiting to pull a U-turn and drive head-on into you, because there's no way to drive safely under such a set of assumptions.
It's been statistically shown (i.e. I've been shown graphs) that going 10km/h faster or slower than the surrounding traffic makes you 10 times more likely to be involved in a multi-vehicle collision. Which makes sense if you think about it. If everyone's going in the same speed and the same direction, you cannot physically hit anyone else, your vectors do not intersect.
Now to the computers. The basic reasoning is this: Take your general population. Consider all those who are competent drivers. Now intersect that group with those who are able to successfully predict what a computer can do, and know it. That intersection will be the only group of people safely able to share roads with a computer. But the first group will be the ones who are allowed to do so.
On the other hand, computers _should_ be able to predict what the other computers are gong to do. Particularly with near-range wireless communications, they can simply _ask_ each other what they're seeing, and what they expect to do.
I think the best bet is to have highways built (or lanes declared and partitioned as) automation-only zones, and let people enter the highway under computer control only. Once on the highway, road convoys can self-assemble for speed/fuel efficiency purposes or not, but it's all computer controlled, so there's no drivers either acting unpredictable, or incorrectly predicting other vehicle's movements.
As a reward for reading this far: I did RTFA and I know this isn't actually what the article's about. In fact, I think the article's describing a really good and workable idea which I wholeheartedly support.
It's just something that's been weighing on my mind since having recently taken a motorcycle learner's course, (and I have other relevant background) and some of the comments on this article (who clearly haven't RTFA and think we're talking about a computer-driven convoy) are wildly mistaken about things like how road safety works, how this idea works, and in fact what this idea is.
I would assume they mean "incorrect judgements made during the process of driving in reaction to other drivers and changing road conditions", rather than "any error made in the process of transitioning from zygote to road kill".
Especially since they're addressing the former, not the latter.
Isn't that what "+1 Insightful" is for?
Except where a development methodology interferes with the ability to evaluate and feedback into the design, which is usually the case for larger games projects.
Short turnaround time on "is X fun?" is absolutely vital to being able to apply a design philosophy and end up with something that's actually fun on time and on-budget (or at least close enough that whoever's responsible doesn't call time on you before you're done)
Second, why must Amazon.com report to Joe Lieberman and his staff for any reason?
It doesn't say they spontaneously reported to him. I read that as "someone on my staff asked Amazon.com and they informed him or her that..."
Or just accept that in English, spelling doesn't have to be tied to pronunciation. Also, vowels are overrated.
A variation of the example and the above quote are from Andrew Tanenbaum's book on Computer Networks.
Which I have indeed read, and so it's probably my source. Certainly the quote's source, less certainly the example.
Back when intercontinental leased lines were all the rage, it was the case that a nightly financial data transfer from (I believe) a stock exchange trading floor was cheaper and faster done by loading the data onto tape and flying someone by Concorde from the UK to the US, than to transmit the data over the network.
That's both anecdotal and marred by my own recollection of the story, but it supports the "never underestimate the bandwidth of a truck full of backup tapes on the highway" saying.
What I'd like to know is, can any app read any file from an SD card if it knows the path of an existing file? From a previous Slashdot story (a few months back, cannot find the link) I had thought each Android application directory on an SD card was somehow isolated, but for this flaw to work at all that cannot be the case. What is to stop a rogue app from accessing any arbitrary application data at a known path on the SD card?
When you put applications on the SD card, their binary directories are isolated from each other, yes. (Through encrypted loop-mounts, I believe)
But the actual data on the sdcard is completely open to all applications. It's basically a large dumping ground for data.
The issue of this exploit is that you never need to grant anything permission to become vulnerable, whereas a rouge app does need to be given permission to be installed, and (I believe) permission to access data on the SD card.
In the end freely licensed source code can have no dictators, only obsoleted dickhead.
Unless you count the "Benevolent Dictator For Life" role in some large and fork-resistant open-source projects.
"Benevolent Dictator For Life" has some examples of such.
I say "fork-resistant" not because they've actively blocked being forked, but because when forks do happen, they rarely survive as independent projects, either dying or rolling back into the mainline.
I'd count GCC as one of the latter, because although egcc did become the main-line, we didn't see (that I understand, anyway) a wholesale supplanting of project management and control the way we did with X.org forking from XFree86.
I honestly have no idea whom Apache was forked from, or the circumstances thereof.
Dear Moderators,
I'd like to request substantive support for the parent post.
Games development can have great advantages over non-games development (irrespective of cubicles) but it's posts like the grandparent that can scare people away from an otherwise fulfilling career.
Sure, you probably end up trading in potential salary, but if you find the right studio and right team for you, it's worth the pay cut.
Regards, from someone who took a $15k/year pay cut to join the games industry almost half a decade ago, and is still thrilled to be making video games for a living.
I've been told that the Tier 1 RvR was the most fun of the lot. And the Empire area worked better than the Dwarven area for Tier 1 RvR. So it's probably the best choice of something to make free to play.
On the other hand, I personally quite enjoyed Tier 2 and Tier 3 keep sieges in both attack and defense, although the boss fights at the end tended to drag on a bit. Quite hard to find more than one group an evening that'd do it though, and once the populations got unbalanced, there was often no keeps available to be taken for days at a stretch.
This was pretty much the basis of WinNY. Also worth reading up on Share which was its successor, and appears to be built around the same principles.
It's 100% reliable for sites in a vulnerable configuration (ie. the sites the attack targets).
They presumably mean this to distinguish from attacks which require some level of luck or coincidence (eg. triggering a race condition) to exploit.
Or for attacks which require a certain state to be achieved but which there's no useful upper bound on when that state might be achieved due to the actions of the attack, eg. a purely passive attack.
So in short, they're saying that the attack is able to drive it's own progress towards its goal, and (I guess) always makes some small amount of progress on every iteration.
From reading the article, I believe what he means is power redundancy.
Specifically, right now, if the burglar cuts power to the home or business, the 48V (or so...) on the line lets the alarm system ring home and tell the company the power's out.
If only the were to put some kind of battery backup option for the ONT...
ONT battery backup information
As for having no government... I can't really grasp what that would mean.
The citizens of Australia will let you know, as soon as we work out what it means.
Funnily enough, The Daily WTF has a posting up about why passing your test cases isn't always sufficient.
Role-based Canary
UI design is one of those areas of expertise that is both an "Art" and a "Science" at the same time.
So's the rest of software development, if you're doing it right.
Oh, and eBay will always have cheaper devices. They generally ship from Hong Kong, offer no warranty or support, and illegally bypass local tariffs and taxes. They usually ship as personal mail with an outrageously low declared value. Legally, you are supposed to declare such purchases but nobody does.
That obviously depends on your local jurisdiction. Where I come from, personal imports below a certain amount legally bypass local tariffs and taxes, and above a certain amount, it seems to be at the discretion of customs as to whether you pay or not.
Be wary of anyone who misdeclares value or nature of goods when shipping to you, that would (probably) interfere with any shipping-related insurance claim you make.