Well, you can place nearby particles in such a way (technical difficulties aside), but what do you do about further ones? A near-zero-speed particle is almost everywhere, so how can you stop an electron over Jupiter from "existing inside" your hole?
Well, for a computer with finite space you could analyse all possible states - but then again to do so you would need a comuter with a larger space, so how do you know that one is correct?
Three outcomes are possible: - Proven incorrect - Proven correct - Unprovable using available system. You just treat "unprovable" as "proven incorrect" for mission-critical pieces. Sanity checks for running out of -its own- space on the proving system are quite simple. Yes, you need a large system, MUCH larger to that. With the example given, you either inline all the functions and prove the program standard way, for all possible n, or if you can't, you take it is possible that "unknown" xn_plus_yn_equals_zn_satisfiable() and mpz_add_ui(); are incorrect (i.e. never return, or always return false) and thus prove the program incorrect. (program correct for certain values, and incorrect for other, IS incorrect as a whole.)
Actually, most of holes on Earth are full on air. Even void isn't quite empty. If you have a couple of atoms forming a particle, the space between them isn't quite empty either - they partially overlap, the uncertainity principle says they "partially are" there. The idea is about making small holes with REAL void - no particles, no photons, no "with a little probability, there" electrons, just total null. Not quite easy. I, for one, can't quite imagine how are they going to stop neutrinos from entering that space...
I can already imagine this...
on
Google TrustRank
·
· Score: 3, Informative
So, links from pages of bad reputation give your page bad reputation? I can see this already....
This page contains very objectionable content. If you are easily offended, don't enter. Blah, blah, blah. Blah, blah, blah.
Except you're running the UNSTABLE release. What if I wanted STABLE? That's important for many, but Woody contains so ancient versions of packages that it's mostly useless. Sid works 99.9% of the time because it contains software with newly fixed bugs. Woody survives 90% cases maybe - old bugs still present, new hardware, new protocols, new requirements of the real world just make it obsolete. We don't need another "unstable" release. But what about finally releasing an up-to-date STABLE one where reliablity counts but the system can use -some- modern technology, and not only what has been obsoleted by much more reliable versions ages ago?
There's a small problem with that... Since the "stable" was released, a lot of apps included in it have been released in newer "stable" versions with numerous bugfixes - so instead of getting pre-alpha quality package from "Stable" (which was included because it was the only thing that was there at that time) I force upgrade to "1.0 stable" version from the "unstable" tree, because it's a year old and proven to be stable, as opposed to 5 years old, pre-alpha. Recently installed Dosemu from the stable tree, for some important work that could be done only with certain ancient DOS application. It kept crashing. Unstable Dosemu worked like a charm - bug fixed. Because of this ages-long release cycle, there's NO REASON TO USE STABLE at all. Because it still has bugs that have been fixed years ago in other releases, making it LESS stable than most newer "unstable" distros.
It'll need significantly more computational power than that. After all a program unintentionally entering an infinite loop is a bug. And since the Halting Problem is uncomputable - no computer (well, not without a *major* breakthrough) can determine whether an arbitrary program will have such a bug.
Why? Not at all. The process checks ALL branchings and ALL possible combinations of states of the program (that's why it's so computationally intensive), and once entering endless loop, the program will keep changing its state in a closed cycle - Pretty simple autocorelation analysis of the time-state function of given branch will reveal it's an endless loop and terminate analysis of the branch. Actually, float/double variables, conditionals inside long loops etc are worse. And of course, multitasking, underlying OS etc will apply to (and kill) the solution if you take timing into consideration. If all you want is the result value in some finite time (not "realtime"), then you may neglect the the timing issues, just guarantee enough resources. Say, calculating stresses of a construction node in some CAD/FEM, as opposed to fuel mixing doesn't have to finish in 0.02ms, but just in such a time that the designer doesn't fall asleep - but it must be just as correct.
So... I write a database that sorts the search using BubbleSort. Only. Nothing else. There's a competing database where I can use arbitrary plugin for sorting, be this quicksort, bubblesort or bogosort. There are many. Most people use the fastest ones, but sometimes they use some odd sorting methods and replace the default quicksort plugin with their own. So I start the benchmark, my database vs the other one. - set up to run on bubble sort. Whoa, my database sorts data faster than the other one! I won! My database is faster!
Actually there ARE techniques of "proving there are no bugs". A program can be mathematically proven to be correct and error-free.
As usually, there's a hook. Proving correctness of anything more complicated than 2-3 nested loops and a handful of conditional statements would require more computational power that exists in the whole world.
Not quite useless - 20-line routine about mixing fuel in a jet engine is something worth proving, and these things are subjected to this technique. But 3 megabytes of an antivirus - sorry...
Let me wake you up. Car manufacturers fight really hard to stop this from getting more of media attention, but modern cars are known to have SERIOUS software bugs. Just google car software bug or similar for stories and references - running 100MPH down a motorway and have the engine switched off, everything shut down (and even the steering wheel blocked), or having the central lock imprison you in the car, so you can't get out, or having random pieces of equipment (wipers, windows, chair adjustment) to start at random... These are real stories. Cars aren't what they used to be...
But if it's not... Killed by antivirus software...? I'd like to see this number confronted with the number of people killed by computer viruses (like crashing medical equipment or blocking some emergency services)
AFAIR, JPEG doesn't support transparency at all. On the other hand, PNG instead of plain 1-bit "transparency" known from GIF supports "alpha channel" which is "level of opacity", allowing for nicely antialiased borders of bullets on arbitrary backgrounds, effects of fading background/text etc. For a neat example of PNG alpha channel in action, choose Help>About>Credits in Firefox - the fading of the text has been done with a simple PNG image with alpha channel gradient overlaying the scrolling text.
No, "Write Once Read Many." If it's not removable, and only 300G, it's dead. But I don't think they would be THAT dumb to invest in a soooo dead technology so I strongly believe there is -something- to this. Like "removable"...
I'd run the test anyway. (I remember buying a "wonderful" Kingston die for a server, lifetime warranty and all, horribly expensive, and finding it faulty) And it's not like I spend 45 minutes in front of the PC, I just start the program and go watch TV or something. I have to install the die, no matter if it's brand or not. So, for now no extra costs.
The costs start only when the die appears to be faulty. At worst 1 in 20 case. Then I need to return it to the store, and repeat the procedure. So divide your "expected" cost by about 20 (slightly more because the next one can be faulty too, slightly less because you have the test setup ready to go again) and reconsider again. You're lucky you live in a rich country. For me, one 512M die is salary of about 3 days of work, so, thanks, I'll take the risk.
Some PC manufacturers get away with selling PCs without MS Windows preinstalled.
If the product is clearly marked as untested, it's my responsiblity to test it (and replace if it's faulty). I pay less, I take more work on myself. It's my choice. Or I risk using a faulty device, but that's my choice again.
Of course this cuts into market share of the "quality brand manufacturers", and they aren't happy about that, but that's a perfectly honest competition.
I definitely prefer to go to shop, get the die, plug it in, run a test program for a few hours and have it replaced if I find any errors, than to pay some 80% extra for a sticker saying that some malaysian kid did it for me.
1) New Technology. Think speed/volume ratio not quite far below 1x CD-ROM. Where will it be in 5-10 years, if the drive enters mainstream? 2) If that's WORM, 300G of fixed drive is useless. 300G of replaceable medium is great. Think situation from early days of CD-ROM again. 3) If you need to move bulk amounts of data, fast, 20M/s is slow. If you want to USE the data even not directly, like watching a movie, just processing it with the machine, like searching database or decrypting data on the fly, 20M/s is quite a lot and requires very decent CPU power.
Well, 300GB removable medium (CD-R alike), with blank disks for less than corresponding volume in CDs (or even a bit more, convenience factor), I'd say promising.
For classic corporate licenses, the priorities of license convenience are like this: 1. The corporation. It's all about the profit to the corporation, isn't it? 2. The customer. Sure we want to attract the customer to buy the product, and the license is partially responsible... 3. The developer. Actually, pretty much the license only takes all the rights away from the developer... They are compensated for that - in form of salary.
For Open Source/Free Software, the licenses must be written in exactly opposite order. 1. The developer. If there's nobody to write the software, nobody will be able to use it. 2. The user. It's for users, and the idea is user is free to do what they want with the software, so let's give it. 3. The corporations. Actually, protecting the program against abuse by corporations.
Of course, SUN as a corporation, wanted the first kind of license. GPL, as the second kind, obviously didn't meet their expectations. What SUN forgot though, is that the goals their license meets are straight opposite to goals -expected- by Open Source developers. A corporate license in service of a corporation, taking some rights away from the customer and many rights away from the developer, while the developer isn't compensated for that in any way?
The difference is that a GNU developers still "own" code they wrote. They retain all rights to the code. SUN's license says, "You will be assimilated. All your code are belong to us."
Recently I saw an ad in a pro magazine (about vision systems). It was about a PC-like device that can interface several types of industrial cameras, has gigabit ethernet etc... It comes bundled with Windows XP Embedded Edition.
The slogan on the bottom of the ad read: "Spend less time integrating your components and more time developing the application."
- After disk crash, rebuild -exactly- the same partition table to have your new hdb1 fit the backup of the old one. - Back up all the empty space... - at least mount/dev/hdb1 -o ro,remount if you want to avoid SERIOUS inconsistencies (in many cases it means downtime of certain services) - except trouble with recovering to a different drive. Whether you back up hdb1 or hdb, the problems are different but present in both cases.
dd is very crude and sucks at backups. Good only if you want to "try something" and then recover the system to the same drive, same partition. Not for preventing disk crash, surface damage etc. A good backup software should be able to survive read errors to recover whatever possible from already damaged disk - dd dies on read errors.
Imagine a Best Buy employee buying racks of CDs using his employee discount, and taking them home and selling them out of his car for a little mark-up. That'd get you fired or your discount privledge suspended right quick.
I imagine. My boss telling the employees he's sorry but they are short on cash and the salary will be delayed, but whoever wants, they can get some of the product they make (for production price) and go, sell it on their own if they can't wait 2 days more.
If they offer a discount program that allows for 3rd party profit that causes loss to the company, that means the program is flawed and they should change it. A reasonable program would still bring them profit, lower than normal sales but definitely on the "plus side". Sell stuff in the shop, while paying normal salaries to the employees and having cost of these salaries added to the price of the wares, simultaneously let your employees sell the same stuff in their spare time on any conditions they desire, letting them buy it for its shelf price minus their salary cost and some (cost of maintenance of the shop, shelf space etc). The profit is the same, no matter if you sell it for full price and then return a part of the price to the buyer in form of salary, or if you just cut on the price right then.
Of course if your "discount programme" causes loss, but you intend i.e. to take customers away from the competition, it's called dumping and it's forbidden by law. A common method of fighting with the dumpers is to come and call their bluff, buy all the products at the price they offer and then sell it yourself for a right price. Of course they won't be happy about that, but serves them right!
Slashdot Mirror
Well, you can place nearby particles in such a way (technical difficulties aside), but what do you do about further ones? A near-zero-speed particle is almost everywhere, so how can you stop an electron over Jupiter from "existing inside" your hole?
But there is, for other, much more stable (newer) distributions of Linux, that contain newer software with less bugs -total- (not just -security-).
Well, for a computer with finite space you could analyse all possible states - but then again to do so you would need a comuter with a larger space, so how do you know that one is correct?
Three outcomes are possible:
- Proven incorrect
- Proven correct
- Unprovable using available system.
You just treat "unprovable" as "proven incorrect" for mission-critical pieces. Sanity checks for running out of -its own- space on the proving system are quite simple. Yes, you need a large system, MUCH larger to that.
With the example given, you either inline all the functions and prove the program standard way, for all possible n, or if you can't, you take it is possible that "unknown" xn_plus_yn_equals_zn_satisfiable() and mpz_add_ui();
are incorrect (i.e. never return, or always return false) and thus prove the program incorrect. (program correct for certain values, and incorrect for other, IS incorrect as a whole.)
Actually, most of holes on Earth are full on air. Even void isn't quite empty. If you have a couple of atoms forming a particle, the space between them isn't quite empty either - they partially overlap, the uncertainity principle says they "partially are" there. The idea is about making small holes with REAL void - no particles, no photons, no "with a little probability, there" electrons, just total null. Not quite easy. I, for one, can't quite imagine how are they going to stop neutrinos from entering that space...
So, links from pages of bad reputation give your page bad reputation?
I can see this already....
This page contains very objectionable content.
If you are easily offended, don't enter.
Blah, blah, blah.
Blah, blah, blah.
Do you agree to these conditions?
Yes No
Except you're running the UNSTABLE release.
What if I wanted STABLE? That's important for many, but Woody contains so ancient versions of packages that it's mostly useless. Sid works 99.9% of the time because it contains software with newly fixed bugs. Woody survives 90% cases maybe - old bugs still present, new hardware, new protocols, new requirements of the real world just make it obsolete.
We don't need another "unstable" release. But what about finally releasing an up-to-date STABLE one where reliablity counts but the system can use -some- modern technology, and not only what has been obsoleted by much more reliable versions ages ago?
There's a small problem with that...
Since the "stable" was released, a lot of apps included in it have been released in newer "stable" versions with numerous bugfixes - so instead of getting pre-alpha quality package from "Stable" (which was included because it was the only thing that was there at that time) I force upgrade to "1.0 stable" version from the "unstable" tree, because it's a year old and proven to be stable, as opposed to 5 years old, pre-alpha. Recently installed Dosemu from the stable tree, for some important work that could be done only with certain ancient DOS application. It kept crashing. Unstable Dosemu worked like a charm - bug fixed.
Because of this ages-long release cycle, there's NO REASON TO USE STABLE at all. Because it still has bugs that have been fixed years ago in other releases, making it LESS stable than most newer "unstable" distros.
It'll need significantly more computational power than that. After all a program unintentionally entering an infinite loop is a bug. And since the Halting Problem is uncomputable - no computer (well, not without a *major* breakthrough) can determine whether an arbitrary program will have such a bug.
Why? Not at all.
The process checks ALL branchings and ALL possible combinations of states of the program (that's why it's so computationally intensive), and once entering endless loop, the program will keep changing its state in a closed cycle - Pretty simple autocorelation analysis of the time-state function of given branch will reveal it's an endless loop and terminate analysis of the branch.
Actually, float/double variables, conditionals inside long loops etc are worse. And of course, multitasking, underlying OS etc will apply to (and kill) the solution if you take timing into consideration. If all you want is the result value in some finite time (not "realtime"), then you may neglect the the timing issues, just guarantee enough resources. Say, calculating stresses of a construction node in some CAD/FEM, as opposed to fuel mixing doesn't have to finish in 0.02ms, but just in such a time that the designer doesn't fall asleep - but it must be just as correct.
So...
I write a database that sorts the search using BubbleSort. Only. Nothing else.
There's a competing database where I can use arbitrary plugin for sorting, be this quicksort, bubblesort or bogosort. There are many. Most people use the fastest ones, but sometimes they use some odd sorting methods and replace the default quicksort plugin with their own.
So I start the benchmark, my database vs the other one. - set up to run on bubble sort.
Whoa, my database sorts data faster than the other one! I won! My database is faster!
Actually there ARE techniques of "proving there are no bugs". A program can be mathematically proven to be correct and error-free.
As usually, there's a hook. Proving correctness of anything more complicated than 2-3 nested loops and a handful of conditional statements would require more computational power that exists in the whole world.
Not quite useless - 20-line routine about mixing fuel in a jet engine is something worth proving, and these things are subjected to this technique. But 3 megabytes of an antivirus - sorry...
Let me wake you up.
Car manufacturers fight really hard to stop this from getting more of media attention, but modern cars are known to have SERIOUS software bugs. Just google car software bug or similar for stories and references - running 100MPH down a motorway and have the engine switched off, everything shut down (and even the steering wheel blocked), or having the central lock imprison you in the car, so you can't get out, or having random pieces of equipment (wipers, windows, chair adjustment) to start at random... These are real stories. Cars aren't what they used to be...
But if it's not... Killed by antivirus software...?
I'd like to see this number confronted with the number of people killed by computer viruses (like crashing medical equipment or blocking some emergency services)
AFAIR, JPEG doesn't support transparency at all.
On the other hand, PNG instead of plain 1-bit "transparency" known from GIF supports "alpha channel" which is "level of opacity", allowing for nicely antialiased borders of bullets on arbitrary backgrounds, effects of fading background/text etc.
For a neat example of PNG alpha channel in action, choose Help>About>Credits in Firefox - the fading of the text has been done with a simple PNG image with alpha channel gradient overlaying the scrolling text.
Duct Tape for an Engineer is like the Dark Side for a Jedi Knight!
No, "Write Once Read Many."
If it's not removable, and only 300G, it's dead. But I don't think they would be THAT dumb to invest in a soooo dead technology so I strongly believe there is -something- to this. Like "removable"...
I'd run the test anyway. (I remember buying a "wonderful" Kingston die for a server, lifetime warranty and all, horribly expensive, and finding it faulty) And it's not like I spend 45 minutes in front of the PC, I just start the program and go watch TV or something. I have to install the die, no matter if it's brand or not. So, for now no extra costs.
The costs start only when the die appears to be faulty. At worst 1 in 20 case. Then I need to return it to the store, and repeat the procedure. So divide your "expected" cost by about 20 (slightly more because the next one can be faulty too, slightly less because you have the test setup ready to go again) and reconsider again. You're lucky you live in a rich country. For me, one 512M die is salary of about 3 days of work, so, thanks, I'll take the risk.
Some PC manufacturers get away with selling PCs without MS Windows preinstalled.
If the product is clearly marked as untested, it's my responsiblity to test it (and replace if it's faulty). I pay less, I take more work on myself. It's my choice. Or I risk using a faulty device, but that's my choice again.
Of course this cuts into market share of the "quality brand manufacturers", and they aren't happy about that, but that's a perfectly honest competition.
I definitely prefer to go to shop, get the die, plug it in, run a test program for a few hours and have it replaced if I find any errors, than to pay some 80% extra for a sticker saying that some malaysian kid did it for me.
1) New Technology. Think speed/volume ratio not quite far below 1x CD-ROM. Where will it be in 5-10 years, if the drive enters mainstream?
2) If that's WORM, 300G of fixed drive is useless. 300G of replaceable medium is great. Think situation from early days of CD-ROM again.
3) If you need to move bulk amounts of data, fast, 20M/s is slow. If you want to USE the data even not directly, like watching a movie, just processing it with the machine, like searching database or decrypting data on the fly, 20M/s is quite a lot and requires very decent CPU power.
Well, 300GB removable medium (CD-R alike), with blank disks for less than corresponding volume in CDs (or even a bit more, convenience factor), I'd say promising.
For classic corporate licenses, the priorities of license convenience are like this:
1. The corporation. It's all about the profit to the corporation, isn't it?
2. The customer. Sure we want to attract the customer to buy the product, and the license is partially responsible...
3. The developer. Actually, pretty much the license only takes all the rights away from the developer... They are compensated for that - in form of salary.
For Open Source/Free Software, the licenses must be written in exactly opposite order.
1. The developer. If there's nobody to write the software, nobody will be able to use it.
2. The user. It's for users, and the idea is user is free to do what they want with the software, so let's give it.
3. The corporations. Actually, protecting the program against abuse by corporations.
Of course, SUN as a corporation, wanted the first kind of license. GPL, as the second kind, obviously didn't meet their expectations. What SUN forgot though, is that the goals their license meets are straight opposite to goals -expected- by Open Source developers. A corporate license in service of a corporation, taking some rights away from the customer and many rights away from the developer, while the developer isn't compensated for that in any way?
The difference is that a GNU developers still "own" code they wrote. They retain all rights to the code. SUN's license says, "You will be assimilated. All your code are belong to us."
Recently I saw an ad in a pro magazine (about vision systems). It was about a PC-like device that can interface several types of industrial cameras, has gigabit ethernet etc... It comes bundled with Windows XP Embedded Edition.
The slogan on the bottom of the ad read:
"Spend less time integrating your components and more time developing the application."
How accurate...
- After disk crash, rebuild -exactly- the same partition table to have your new hdb1 fit the backup of the old one. /dev/hdb1 -o ro,remount if you want to avoid SERIOUS inconsistencies (in many cases it means downtime of certain services)
- Back up all the empty space...
- at least mount
- except trouble with recovering to a different drive. Whether you back up hdb1 or hdb, the problems are different but present in both cases.
dd is very crude and sucks at backups. Good only if you want to "try something" and then recover the system to the same drive, same partition. Not for preventing disk crash, surface damage etc. A good backup software should be able to survive read errors to recover whatever possible from already damaged disk - dd dies on read errors.
I'd rather go with LVM/RAID1.
Or just like for James Bond movies, use
plot.continuity.hole=sizeof( DRIVABLE_THROUGH(TRUCK) );
Imagine a Best Buy employee buying racks of CDs using his employee discount, and taking them home and selling them out of his car for a little mark-up. That'd get you fired or your discount privledge suspended right quick.
I imagine. My boss telling the employees he's sorry but they are short on cash and the salary will be delayed, but whoever wants, they can get some of the product they make (for production price) and go, sell it on their own if they can't wait 2 days more.
If they offer a discount program that allows for 3rd party profit that causes loss to the company, that means the program is flawed and they should change it. A reasonable program would still bring them profit, lower than normal sales but definitely on the "plus side". Sell stuff in the shop, while paying normal salaries to the employees and having cost of these salaries added to the price of the wares, simultaneously let your employees sell the same stuff in their spare time on any conditions they desire, letting them buy it for its shelf price minus their salary cost and some (cost of maintenance of the shop, shelf space etc). The profit is the same, no matter if you sell it for full price and then return a part of the price to the buyer in form of salary, or if you just cut on the price right then.
Of course if your "discount programme" causes loss, but you intend i.e. to take customers away from the competition, it's called dumping and it's forbidden by law. A common method of fighting with the dumpers is to come and call their bluff, buy all the products at the price they offer and then sell it yourself for a right price. Of course they won't be happy about that, but serves them right!