You make each program/script self sufficent and then debug it seperately.
So you don't test the units together? Sounds silly to me.
I was once a part of a project that designed a system to be lots of processes communicating over pipes. Each process could have been in any language and free to do whatever it wanted internally - they were to be black boxes. The problem was that a bug in one would cause it to crash at certain times, causing a broken pipe error for the other processes, that then broke their pipes until everything dies and there was nothing more than core! It was darn hard to debug
I'm not opposed to multi language things, just putting everything in different processes is brain dead... at least use bindings and link the stuff together into one pile that can be debugged reasonably. If it can't be debugged reasonably you're doing something wrong.
This sounds like an absolute nightmare to debug! Having lots of languages doing different bits executing as individual processes probably means that there is no posibility of using a debugging tool to generate a clear stack back trace showing functions and arguments. In fact, this type of system probably requires many different debuggers all running the various parts at the same time, something that is going to be difficult to manage, prducing diagnostics that are hard to visualise/comprehend and using a hell of a lot of memory.
I wouldn't like a heisen-bug to get into *that* system.
I have some pirated mp3's on my computer, but they are of bands whos cd's I would NEVER purchase. Generally, if I like even two songs off of the same CD, I go out and buy it.. and most other people out there are similar in nature.
You are entirely correct. I do the same. In fact, the process I used to use to decide whether I was going to buy an album was something like this
Hear a good single on the radio/in club/wherever...
Goto Amazon and find the album.
Amazon samples are crap. Usually only have the first 30 seconds of each song (which is therefore often intro and not real song) and often don't have samples for all (if any of the songs).
Goto Napster, start pulling some of the songs.
Listen to them and evaluate them.
Often listen a few times before deciding if good or not (some songs are real growers, while some are the other way).
Optionaly buy it.
I think Amazon should carry the whole tracks instead of 30 second samples & the RIAA should goto hell.
All the in car nav systems I have seen/used always pop up a box warning the driver not to use the system while driving. I guess this allows the manufacturer to disclaim liability problems that might happen for accidents with claims like "I was looking at my nav system when I hit x..."
So on the one hand we are discouraged to use the device then driving around, and to only look it at to get directions, but now the device is going to be advertising junk - trying to get our attention?
Seems like a liable case waiting to happen, unless it only displays spam when the vehicle is detected to be stationary (which would make the spam low volume->not spam).
You are not the average user. You know about IPsec etc....
The average user probably does not know about IPsec, encryption in 802.11 makes some sense for these people - even if the solution offered is imperfect.
Then encrypt it at both ends at the transport layer (ever heard of https??). Oh, and to do the encryption on servers makes it a bit harder to scale them - since most news sites
Fine if you own the servers and can get them to use https... but if you don't?
Please, someone enlighten me as to why, exactly, 802.11 in itself has to be secure!
It doesn't have to be secure (think IPsec), but I'm sure that everyone can see major benfits of making a technology that openly broadcasts data more secure.
Paper based photos degrade rapidly in time, oxidising in the air and being quite fragile.
Surely digital photos are better *if* you take the time to copy them to modern storage and make ample backups. If you can do this, the images will be preserved in pixel quality.
Oh, and the fact that you need a lot of space to store paper based documents, and digital can sit in a little hard disk.
I think that it is more about the love and attention you want to spend keeping images or documents of any sort. With digital you probably have a better chance at preserving more for longer...
Bastards! What, do they think they invented the Internet or something?
It seems odd that someone would invent one system for themself, and a completely different one for all others... although this mess probably evolved, as mess does.
So why not just stick to.co.uk companies
Lots of.co.uk companies think it is cool to be a.com! Depending on what you are looking for, you can be limited by adding.co.uk as a constraint - esp if bargain hunting.
It is true that lots of US companies will ship, as will companies from all around the world, but there isn't a term you can easily put into Google to differentiate such companies. Putting in site:.co.uk should do a good job, but doesn't for reasons given.
You're taking outdated geographical notions and superimposing them on the Internet, which goes beyond illogical to simply unreasonable
I'm not sure I agree with the whole Internet as a different place concept. It is very firmly rooted in the real world. Companies, be them.com's or whatever will have some registered trading address that ties them to some place on Earth. Some may have a number of addresses, in which case they can take a number of country codes.
In fact, if you think about it, a limited company is a legal entity, so it must be tied to laws of at least one country.
Slashdot or The Onion is
Slashdot is.us because of it's US biased content. The Onion is.us too, since at calls itself 'America's Finest News Source' (go there and look at the title bar of your browser yourself).
Oh, and I'd keep.org for global organisations such as Amnesty International.
It seems to me that lots of people want to get good games for Linux so that they can drop Windows and end dual booting. I know that I would like that.
However, why don't we just go out an buy a console? That would sort the dual boot problem for good and give a purpose built system on which to run games, cutting the need for Windows...
Country code suffixes such as ".fr" for France have been sources of national pride
worldwide, but in the United States it is the forgotten stepchild compared with ".com."
It's always annoyed me how the world seems to use country codes for it's TLD's, and then the US has some other TLDs that is just uses.
For example, when shopping online I want to know if a company will ship to the UK. If it is a.co.uk company I can be sure it will. If it is a.com, it might or might not.
Essentially it seems logical for organisations to just register the TLDs for the countries in which they operate/are registered, and for the.com TLD to be scrapped (Although this would never happen).
Oh, I'd scrap.edu too..ac.us would be a fine replacement.
Re:Little content, little meaning...
on
MS DOS: A Eulogy
·
· Score: 1
Still, I'd be willing to argue that the removal of legacy DOS functionality isn't always a good thing. You break functionality with code that used to run on previous MS Operating systems
I'd have to disagree. It is legacy that makes Wintel boxes as bad as they are. In the case of windows, parts of the API's are broken and have to be that way to remain compatible with older programs (a problem for WINE as they have to implement the bugs!). Leaving these bugs in is bad for developers and also leaves the API's a mess. In the case of chip makers, a lot of them would probably love to ditch compatability with older chips to make something fast and lean...
curl | csh? Danger will Robinson, danger!!!
on
GNU-Darwin Goes Beta
·
· Score: 3, Insightful
Erm. Isn't this a bit of a dangerous install strategy? e.g. sourceforge get hacked again and http://gnu-darwin.sourceforge.net/one_stop points to a script that starts with 'rm -rf/'. Not so fun now...
Wouldn't it be better to use something that does a bit of public key crypto and verifies that you are really downloading something signed by a darwin guy or sourceforge? At least using https would help to stop a man in the middle attack...
some sort of subscription system where you can pay a fee to disable them honestly.
I'm sure that there will be a spate of software that dynamically removes the adds as slashdot passes through a proxy server, if the existing pile of add-blockers don't work already.
From the changelog "changed... to avoid n^2 behavior..."
Why the hell is n^2 in there in the first place? And in more than one place according to the changelog. What sort of idiots have they got writing this...
I thought that everyone knew n^2 was a Bad Thing and to avoid it and bubble sorts like the plague.
It's not as if it is any effort to use an already written better data structure/sort algorithm that is in some library of which there are hundreds!
In fact, there are many really really good libraries around - I thought that half the point of open source was that you can find excellent code and reuse it under the GPL or whatever.
I really can't see these things ever being useful in a normal office context.
Sure, a handheld tricorder-like device would be really good for military situations, labs or post offices (sit one over the line), even-more-so if it could be extended to other pathogens.
However, it seems like another non-product who's market rides on peoples current paranoia. Unfortunately in this case, the terror will hopefully be over way before this is even a prototype!
Never played with magnetic fluid???
on
Magnetic Fluids
·
· Score: 3, Informative
Never played with magnetic fluid though.
Michael can't have been to a science museum of late. I can remeber seeing small tanks of magnetic fluid that allow you to wave magnets around near them to see what happes in museums a couple of years ago.
Oh, and Wired magazine had a lovely picture of magnetic fluid in a beautiful state that was to be shown at Siggraph (See wired for article check here for video)
That's a cool idea, but how powerful of a processor/how much memory would be needed to encrypt pictures on a small device like that?
Well, the CPU in a camera does a pretty good and fast job of JPEG compression, so it's likely to be able to do public key encryption pretty well too. What would probably actually happen in the camera is that a message key would be chosen for an image. The image is then encrypted with a bulk cypher using the message key (bulk cyphers are generally faster). The message key is then encrypted with the public key and added to the message. What might be slow is picking a secure random key for the bulk cypher...
Of course, key length is a factor here too, and that will affect speed. I doubt it would be that slow, but if it was, you could always implement in hardware (The AES candidates were tested for hardware implementation suitablilty, so should be easy).
Solution: copy the program and disassemble the key out of it?
So security now relies upon the physical dongle. I like it - steal that and you have broken the law in quite a conventional manner, that should be easy to prosecute for etc...
I guess you could always use an iButton for the dongle - that would be more secure than most dongles (and they break if you open them up to look for data:)
I think it's just a matter of (little) time before hordes of enterprising geeks figure out how to get the pics out and reuse it without paying the fee
If they use public key cryptography they can make this impossible...
The camera just has the shops public key embedded and encrypts the pictures as it encodes/saves them. Without the private key held at the shop, even the best geek is scuppered.
Also assume that the public key held in the camera can't be swapped changed and is in ROM
The whole reputation system is flawed because untrustworthy people are allowed to give out good feedback.
Err, you are lacking a base case without this!
(the trusted people have to come from somewhere!!!)
Maybe a system where it is hard to gain trust and easy to lose is better. Kinda like TCP windows. Trust is gained point at a time, but lost by half. That way, one person with a complaint can really blow the whistle on a bad guy.
This would put the cost of scamming up quite a bit assuming each ripped off person provides the negative feedback.
A nice fiber optic wire has a lot more bandwidth than some gold or copper. And this really does eliminate "all these problems" (except latency).
Now I'm not sure here, so correct me if I'm wrong, but isn't latency a really *important* factor for the CPU? If the latency goes up then the pipeline in the CPU is going to have the get longer (the pipeline is effectively special memory/registers on the chip, no?) to ensure that instructions and operands are available for execution well in advance; that's going to add more logic and cost to a CPU and also make the cost of a stall a lot lot bigger.
How is it possible, on first glance, that a serial protocol, sending a single bit at a time, is faster than a parallel one, sending bytes at a time? Simple! It sends a bit much more often
I'm sure I've seen this discussed before and that a number of problems exist with an optical bus in a non-optical system.
Firstly, the length of the bus on a motherboard is so short that there are few real gains over a copper/gold track, and those gains that are made are outweighed by the encoders/decoders that do the photonelectron conversions.
Also, it would probably put the cost of add-in cards up since the row of gold contacts has to be replaced with something far more sophistocated.
Also, one of the problems with existing bandwidth to the memory is not only the speed, but also the bus width. Unfortunately a wider bus gives more bandwidth (assumming that data lines are added, and not address), but also means more pins on the chip, which costs more.
In a pure optical system, it maybe possible to eliminate all these problems, but I'm not convinced from what I have read that it is a solution for todays computers...
Re:Evolution proceeds towards what works...
on
Autonomic Computing
·
· Score: 2, Interesting
My high school biology teacher must've said a thousand times, "Evolution proceeds towards what works, not towards what is best."
Erm, sort of true. Evolution actually works to do what is 'best' in terms of the fitness function... i.e. seeks to maximise or minimise the result of some metric. If you pick your fitness function correctly, you can make the system optimise towards any required goal.
Just make sure you don't have any bugs - because GA's and GP's will find and exploit bugs that give higher fitness metrics faster than the programmer.
Slashdot Mirror
You make each program/script self sufficent and then debug it seperately.
So you don't test the units together? Sounds silly to me.
I was once a part of a project that designed a system to be lots of processes communicating over pipes. Each process could have been in any language and free to do whatever it wanted internally - they were to be black boxes. The problem was that a bug in one would cause it to crash at certain times, causing a broken pipe error for the other processes, that then broke their pipes until everything dies and there was nothing more than core! It was darn hard to debug
I'm not opposed to multi language things, just putting everything in different processes is brain dead... at least use bindings and link the stuff together into one pile that can be debugged reasonably. If it can't be debugged reasonably you're doing something wrong.
Ahh, Haskell - there's elegance for you :)
This sounds like an absolute nightmare to debug! Having lots of languages doing different bits executing as individual processes probably means that there is no posibility of using a debugging tool to generate a clear stack back trace showing functions and arguments. In fact, this type of system probably requires many different debuggers all running the various parts at the same time, something that is going to be difficult to manage, prducing diagnostics that are hard to visualise/comprehend and using a hell of a lot of memory.
I wouldn't like a heisen-bug to get into *that* system.
I have some pirated mp3's on my computer, but they are of bands whos cd's I would NEVER purchase. Generally, if I like even two songs off of the same CD, I go out and buy it.. and most other people out there are similar in nature.
You are entirely correct. I do the same. In fact, the process I used to use to decide whether I was going to buy an album was something like this
I think Amazon should carry the whole tracks instead of 30 second samples & the RIAA should goto hell.
All the in car nav systems I have seen/used always pop up a box warning the driver not to use the system while driving. I guess this allows the manufacturer to disclaim liability problems that might happen for accidents with claims like "I was looking at my nav system when I hit x..."
So on the one hand we are discouraged to use the device then driving around, and to only look it at to get directions, but now the device is going to be advertising junk - trying to get our attention?
Seems like a liable case waiting to happen, unless it only displays spam when the vehicle is detected to be stationary (which would make the spam low volume->not spam).
You are not the average user. You know about IPsec etc....
The average user probably does not know about IPsec, encryption in 802.11 makes some sense for these people - even if the solution offered is imperfect.
Then encrypt it at both ends at the transport layer (ever heard of https??). Oh, and to do the encryption on servers makes it a bit harder to scale them - since most news sites
Fine if you own the servers and can get them to use https... but if you don't?
Please, someone enlighten me as to why, exactly, 802.11 in itself has to be secure!
It doesn't have to be secure (think IPsec), but I'm sure that everyone can see major benfits of making a technology that openly broadcasts data more secure.
Paper based photos degrade rapidly in time, oxidising in the air and being quite fragile.
Surely digital photos are better *if* you take the time to copy them to modern storage and make ample backups. If you can do this, the images will be preserved in pixel quality.
Oh, and the fact that you need a lot of space to store paper based documents, and digital can sit in a little hard disk.
I think that it is more about the love and attention you want to spend keeping images or documents of any sort. With digital you probably have a better chance at preserving more for longer...
Bastards! What, do they think they invented the Internet or something?
It seems odd that someone would invent one system for themself, and a completely different one for all others... although this mess probably evolved, as mess does.
So why not just stick to .co.uk companies
Lots of .co.uk companies think it is cool to be a .com! Depending on what you are looking for, you can be limited by adding .co.uk as a constraint - esp if bargain hunting.
It is true that lots of US companies will ship, as will companies from all around the world, but there isn't a term you can easily put into Google to differentiate such companies. Putting in site:.co.uk should do a good job, but doesn't for reasons given.
You're taking outdated geographical notions and superimposing them on the Internet, which goes beyond illogical to simply unreasonable
I'm not sure I agree with the whole Internet as a different place concept. It is very firmly rooted in the real world. Companies, be them .com's or whatever will have some registered trading address that ties them to some place on Earth. Some may have a number of addresses, in which case they can take a number of country codes.
In fact, if you think about it, a limited company is a legal entity, so it must be tied to laws of at least one country.
Slashdot or The Onion is
Slashdot isOh, and I'd keep .org for global organisations such as Amnesty International.
It seems to me that lots of people want to get good games for Linux so that they can drop Windows and end dual booting. I know that I would like that.
However, why don't we just go out an buy a console? That would sort the dual boot problem for good and give a purpose built system on which to run games, cutting the need for Windows...
Country code suffixes such as ".fr" for France have been sources of national pride worldwide, but in the United States it is the forgotten stepchild compared with ".com."
It's always annoyed me how the world seems to use country codes for it's TLD's, and then the US has some other TLDs that is just uses.
For example, when shopping online I want to know if a company will ship to the UK. If it is a .co.uk company I can be sure it will. If it is a .com, it might or might not.
Essentially it seems logical for organisations to just register the TLDs for the countries in which they operate/are registered, and for the .com TLD to be scrapped (Although this would never happen).
Oh, I'd scrap .edu too. .ac.us would be a fine replacement.
Still, I'd be willing to argue that the removal of legacy DOS functionality isn't always a good thing. You break functionality with code that used to run on previous MS Operating systems
I'd have to disagree. It is legacy that makes Wintel boxes as bad as they are. In the case of windows, parts of the API's are broken and have to be that way to remain compatible with older programs (a problem for WINE as they have to implement the bugs!). Leaving these bugs in is bad for developers and also leaves the API's a mess. In the case of chip makers, a lot of them would probably love to ditch compatability with older chips to make something fast and lean...
curl http://gnu-darwin.sourceforge.net/one_stop | csh
Erm. Isn't this a bit of a dangerous install strategy? e.g. sourceforge get hacked again and http://gnu-darwin.sourceforge.net/one_stop points to a script that starts with 'rm -rf /'. Not so fun now...
Wouldn't it be better to use something that does a bit of public key crypto and verifies that you are really downloading something signed by a darwin guy or sourceforge? At least using https would help to stop a man in the middle attack...
Let MicroSoft AND XP/Passport users learn the hard way.
I hope your morgage advisor, bank manager, car sales man, plumber, electician and doctor all have the same attitude as you.
One of the benefits of society is that the stronger, more intelligent, gifted or whatever, can help the weaker...
some sort of subscription system where you can pay a fee to disable them honestly.
I'm sure that there will be a spate of software that dynamically removes the adds as slashdot passes through a proxy server, if the existing pile of add-blockers don't work already.
You can't fight the geeks!
From the changelog "changed ... to avoid n^2 behavior..."
Why the hell is n^2 in there in the first place? And in more than one place according to the changelog. What sort of idiots have they got writing this...
I thought that everyone knew n^2 was a Bad Thing and to avoid it and bubble sorts like the plague.
It's not as if it is any effort to use an already written better data structure/sort algorithm that is in some library of which there are hundreds!
In fact, there are many really really good libraries around - I thought that half the point of open source was that you can find excellent code and reuse it under the GPL or whatever.
n^2? No-neeed!
I really can't see these things ever being useful in a normal office context.
Sure, a handheld tricorder-like device would be really good for military situations, labs or post offices (sit one over the line), even-more-so if it could be extended to other pathogens.
However, it seems like another non-product who's market rides on peoples current paranoia. Unfortunately in this case, the terror will hopefully be over way before this is even a prototype!
Never played with magnetic fluid though.
Michael can't have been to a science museum of late. I can remeber seeing small tanks of magnetic fluid that allow you to wave magnets around near them to see what happes in museums a couple of years ago.
Oh, and Wired magazine had a lovely picture of magnetic fluid in a beautiful state that was to be shown at Siggraph (See wired for article check here for video)
That's a cool idea, but how powerful of a processor/how much memory would be needed to encrypt pictures on a small device like that?
Well, the CPU in a camera does a pretty good and fast job of JPEG compression, so it's likely to be able to do public key encryption pretty well too. What would probably actually happen in the camera is that a message key would be chosen for an image. The image is then encrypted with a bulk cypher using the message key (bulk cyphers are generally faster). The message key is then encrypted with the public key and added to the message. What might be slow is picking a secure random key for the bulk cypher...
Of course, key length is a factor here too, and that will affect speed. I doubt it would be that slow, but if it was, you could always implement in hardware (The AES candidates were tested for hardware implementation suitablilty, so should be easy).
Solution: copy the program and disassemble the key out of it?
So security now relies upon the physical dongle. I like it - steal that and you have broken the law in quite a conventional manner, that should be easy to prosecute for etc...
I guess you could always use an iButton for the dongle - that would be more secure than most dongles (and they break if you open them up to look for data :)
I think it's just a matter of (little) time before hordes of enterprising geeks figure out how to get the pics out and reuse it without paying the fee
If they use public key cryptography they can make this impossible...
The camera just has the shops public key embedded and encrypts the pictures as it encodes/saves them. Without the private key held at the shop, even the best geek is scuppered.
Also assume that the public key held in the camera can't be swapped changed and is in ROM
The whole reputation system is flawed because untrustworthy people are allowed to give out good feedback.
Err, you are lacking a base case without this! (the trusted people have to come from somewhere!!!)
Maybe a system where it is hard to gain trust and easy to lose is better. Kinda like TCP windows. Trust is gained point at a time, but lost by half. That way, one person with a complaint can really blow the whistle on a bad guy.
This would put the cost of scamming up quite a bit assuming each ripped off person provides the negative feedback.
Has anyone mentioned that security through obscurity is a dangerous thing that does not work?
A nice fiber optic wire has a lot more bandwidth than some gold or copper. And this really does eliminate "all these problems" (except latency).
Now I'm not sure here, so correct me if I'm wrong, but isn't latency a really *important* factor for the CPU? If the latency goes up then the pipeline in the CPU is going to have the get longer (the pipeline is effectively special memory/registers on the chip, no?) to ensure that instructions and operands are available for execution well in advance; that's going to add more logic and cost to a CPU and also make the cost of a stall a lot lot bigger.
How is it possible, on first glance, that a serial protocol, sending a single bit at a time, is faster than a parallel one, sending bytes at a time? Simple! It sends a bit much more often
Or FDM...
I'm sure I've seen this discussed before and that a number of problems exist with an optical bus in a non-optical system.
Firstly, the length of the bus on a motherboard is so short that there are few real gains over a copper/gold track, and those gains that are made are outweighed by the encoders/decoders that do the photonelectron conversions.
Also, it would probably put the cost of add-in cards up since the row of gold contacts has to be replaced with something far more sophistocated.
Also, one of the problems with existing bandwidth to the memory is not only the speed, but also the bus width. Unfortunately a wider bus gives more bandwidth (assumming that data lines are added, and not address), but also means more pins on the chip, which costs more.
In a pure optical system, it maybe possible to eliminate all these problems, but I'm not convinced from what I have read that it is a solution for todays computers...
My high school biology teacher must've said a thousand times, "Evolution proceeds towards what works, not towards what is best."
Erm, sort of true. Evolution actually works to do what is 'best' in terms of the fitness function... i.e. seeks to maximise or minimise the result of some metric. If you pick your fitness function correctly, you can make the system optimise towards any required goal.
Just make sure you don't have any bugs - because GA's and GP's will find and exploit bugs that give higher fitness metrics faster than the programmer.