For non-computer situations, good old mechanical pencil and paper, with a good supply of fresh lead and erasers. For the MS-DOS days, good old SideKick by Borland For later MS-DOS days, Edwin (the macros were very helpful) For Windows, Notepad++ For Lots of notes, WikidPad For quick notes on a windows machine I don't own.... Notepad For notes on a linux machine - gedit / WikiPad For notes on RSTS/E - VTedit, or Teco
Doug, there are many non-technical networks in the world which are very complex, have threats against them, yet manage to persist in spite of those threats. For example, consider the world of banking prior to computing. Every branch was subject to attack, but at worst, the financial losses in any theft were limited to those on hand in the vault. There was no way to leverage an activity in one branch against the whole of the banking system.
However, in modern operating systems, there is no practical way to segregate activity of any program to a limited sphere of influence... any line of code can be used as a lever to attack the whole system. There are operating systems which require the user to specify which files and/or folders a process is allowed to use, in a user friendly way.... they are by no means common, nor mainstream... but they do exist, one such example is the Genode project.
This ability to actively and positively limit the scope of changes of any line of code means that complexity doesn't have to equate to insecurity, at least from my perspective. The power grid functions with millions of end points, but circuit breakers keep errant toasters from taking down the grid. The same can be done with computing, and it doesn't have to be user hostile.
The war is not lost, but we have to stop building our fortifications out of crates of C4 before we can turn things around.
Have you looked into Capability based Security Operating Systems such as Genode? (Genode.org) They seem to offer a way for users to decide what to trust, instead of being forced to blindly trust everything every app does.
What do you think about this approach to security?
I bet there'll be a little tiny box in each one, (like the one Nikola Tesla built) which generates enormous useful energy, from no apparent source... meanwhile causing untold pollution in nearby parallel universes.
Hi, I'm a modern day Cassandra... I've been shouting for years about a solution that can actually fix computer security, and render all this "cyberwar" crap obsolete...
Even the Wikipedia page is a mess, but you'll find the solution buried in it... it's called the Principle of least Privilege, and I figure it's 10 more years of hell before people catch on and actually start to fix things.
It is entirely possible to give users a modern GUI interface which transparently and intuitively allows them to decide which resources a program should be allowed to access, which doesn't add any cognitive load, and results in a system which can't be hacked, given a reasonably careful user. Also, by reasonably careful, I mean someone like your parents - someone who understands how cash and credit cards and social engineering work, and hasn't been scammed out of real money. You don't have to be the NSA to secure your PC, but you do have to have an OS that doesn't trust everything. (Good luck finding one!)
Having an admin set up a static set of privileges on each and every program isn't a sustainable approach... what's needed for general purpose use is called the "power box", in which the operating system directly asks the user about which files to open, etc... instead of trusting the application to do it.
Users can generally decide correctly what files to access, etc.. you don't have to have an admin do it.
You've got a lot of hard won experience, I'll give you that... but the problem is a whole new layer, deeper than you're used to thinking about. Imagine if you built a old style fort, moved your troops in, and generally felt secure.... only to find out the bricks it was built out of were actually blocks of C4, and any one of them could send the whole place up in a flash.
If you can imagine that scenario... you know what computer security is really like, no matter how careful you are. Because Windows, Mac-OS, Linux, and pretty much every non-mainframe OS out there runs every line of code with the full privileges of a user account at all times, there's no way for a user to limit the scope of what a program does at run time.
The solution is to use an operating system that is designed from the ground up to simply ask which files the user wishes to operate on, instead of blindly trusting the program to do the right thing. This makes it possible for the user to limit side effects by design, which then makes it possible to have end nodes that are reasonably secure... which makes it possible to have real security.
I still don't see the change to things like Genode happening for at least 10 more years.
No... it's not about money... it's people don't understand the difference between POLA and the way things are done now... until that changes, no amount of money is going to help.
Algebra is easy.. it's all about getting X alone in the corner, so you can find his value. Geometry should go, along with the foreign language requirement.
I'd change the rating system so that there were multiple dimensions of ratings. Humor, Insightfulness, Correctness, and Offensiveness could all be orthogonal to each other.
The introduction of virtual machines was a huge step forward for computer utilization and security. The ability for a single physical server to serve a diverse set of workloads in a secure and efficient manner made mainframes far more versatile and is a cornerstone of cloud computing.
Uranium fueled reactors are the result of a premature optimization... they aren't reactive enough to work with oxides as fuel.. so you end up having to do all sorts of engineering to try to keep it from oxidizing, whilst only a small barrier away from water. It was never a good idea. The hydrogen bubble that almost made 3 mile island even worse is a result of this chemistry at work. Not only that, when Uranium splits, it only yields 90% of the energy immediately, the remaining 10% takes millions of years, which means a reactor producing 1GW of heat at load will still generate 100 Megawatts when you stop the chain reaction... and if you can't cool it, the thing will melt down.
Thorium yields 99% of the energy immediately, which reduces the need for cooling after the fact by a factor of 10... plus in a Thorium reactor, the fuel is a liquid fluoride, which means you just have to divide the critical mass in the event of an emergency, and you're done with it. A few flat wide steel tanks encased in concrete would do the trick, even if dry.
I'd happily live down the street from a Thorium reactor.
So...nothing about a version of windows that doesn't give ambient authority to every line of code that runs... this has a zero percent chance of success.
There were strong warnings in advance of the 9/11 attack, which the Bush/Cheney administration chose to ignore. (Could it be because they needed a "Pearl Harbor" event to catalyze the "New American Century" vision from PNAC?) They then ignored the reality that Iraq had NOTHING TO DO WITH 9/11, and lied their way to get us into Iraq (in accordance with the "New American Century") It didn't work, because the full might of the country couldn't actually be used (being based on a lie, and not having full support) This broke Iraq, and gave rise to ISIS, killing or displacing millions along the way. My heart goes out to the victims in Paris, and the millions of others our nation has displaced, maimed, or killed in the service of our empire.
Consider yourself in a cyber-war... any line of program you run on your computer can be turned against you... why do you trust any of it with your full authority?
Because you don't have a choice, your OS doesn't give you one. Read up on the principle of least privilege, and the ambient authority model we currently use.
There's no way to specify "run this task with this type of access only to this set of stuff" in Linux... which means you're giving your authority to everything you execute. Until this gets fixed... any flaw in any of the code you run can be used against you.
If you could specify authority and filter it, in a similar manner to unix pipes, you'd be able to build a database that can only take local connections, then build a read-only connection to it, then build a web page that could only connect to that and respond to requests... and finally the web server to take requests from the web and query the page.... and an outside hacker would have to pick through each layer on his way to the database... even if the code was only 99% effective, that's a 99.9999% effective block with very minimal effort.
This type of stuff doesn't have to be user-unfriendly, in fact if implemented correctly it would be fairly transparent to them.
Actually there are 100 choices..
Trust the code
Don't use it
Write your own
Use an OS that sandboxes it by default, and only gives it the capabilities you supply.
If I upgrade to this kernel, how many of those lines of code do I actually have to trust not to give away everything on my machine to a black hat? Hopefully less than 100k lines of code are actually critical.
The grid voltage isn't the problem, it's the DC currents which would be induced on the power lines... transformers can't handle DC bias, it causes the core to saturate, which then causes the inductance to fall, primary currents to rise, and then failure from all the extra current heating the windings.
Slashdot Mirror
For non-computer situations, good old mechanical pencil and paper, with a good supply of fresh lead and erasers.
For the MS-DOS days, good old SideKick by Borland
For later MS-DOS days, Edwin (the macros were very helpful)
For Windows, Notepad++
For Lots of notes, WikidPad
For quick notes on a windows machine I don't own.... Notepad
For notes on a linux machine - gedit / WikiPad
For notes on RSTS/E - VTedit, or Teco
Doug, there are many non-technical networks in the world which are very complex, have threats against them, yet manage to persist in spite of those threats. For example, consider the world of banking prior to computing. Every branch was subject to attack, but at worst, the financial losses in any theft were limited to those on hand in the vault. There was no way to leverage an activity in one branch against the whole of the banking system.
However, in modern operating systems, there is no practical way to segregate activity of any program to a limited sphere of influence... any line of code can be used as a lever to attack the whole system. There are operating systems which require the user to specify which files and/or folders a process is allowed to use, in a user friendly way.... they are by no means common, nor mainstream... but they do exist, one such example is the Genode project.
This ability to actively and positively limit the scope of changes of any line of code means that complexity doesn't have to equate to insecurity, at least from my perspective. The power grid functions with millions of end points, but circuit breakers keep errant toasters from taking down the grid. The same can be done with computing, and it doesn't have to be user hostile.
The war is not lost, but we have to stop building our fortifications out of crates of C4 before we can turn things around.
Have you looked into Capability based Security Operating Systems such as Genode? (Genode.org) They seem to offer a way for users to decide what to trust, instead of being forced to blindly trust everything every app does.
What do you think about this approach to security?
I bet there'll be a little tiny box in each one, (like the one Nikola Tesla built) which generates enormous useful energy, from no apparent source... meanwhile causing untold pollution in nearby parallel universes.
Hi, I'm a modern day Cassandra... I've been shouting for years about a solution that can actually fix computer security, and render all this "cyberwar" crap obsolete...
Even the Wikipedia page is a mess, but you'll find the solution buried in it... it's called the Principle of least Privilege, and I figure it's 10 more years of hell before people catch on and actually start to fix things.
It is entirely possible to give users a modern GUI interface which transparently and intuitively allows them to decide which resources a program should be allowed to access, which doesn't add any cognitive load, and results in a system which can't be hacked, given a reasonably careful user. Also, by reasonably careful, I mean someone like your parents - someone who understands how cash and credit cards and social engineering work, and hasn't been scammed out of real money. You don't have to be the NSA to secure your PC, but you do have to have an OS that doesn't trust everything. (Good luck finding one!)
What's wrong with police cars being marked as police cars? Why put Google drivers at risk for no good reason?
The difference is that doing this would put Google maps drivers in danger.
Just like when the CIA sent spies disguised as vaccine workers, and set back the effort to eliminae smallpox worldwide.
Having an admin set up a static set of privileges on each and every program isn't a sustainable approach... what's needed for general purpose use is called the "power box", in which the operating system directly asks the user about which files to open, etc... instead of trusting the application to do it.
Users can generally decide correctly what files to access, etc.. you don't have to have an admin do it.
You've got a lot of hard won experience, I'll give you that... but the problem is a whole new layer, deeper than you're used to thinking about. Imagine if you built a old style fort, moved your troops in, and generally felt secure.... only to find out the bricks it was built out of were actually blocks of C4, and any one of them could send the whole place up in a flash.
If you can imagine that scenario... you know what computer security is really like, no matter how careful you are. Because Windows, Mac-OS, Linux, and pretty much every non-mainframe OS out there runs every line of code with the full privileges of a user account at all times, there's no way for a user to limit the scope of what a program does at run time.
The solution is to use an operating system that is designed from the ground up to simply ask which files the user wishes to operate on, instead of blindly trusting the program to do the right thing. This makes it possible for the user to limit side effects by design, which then makes it possible to have end nodes that are reasonably secure... which makes it possible to have real security.
I still don't see the change to things like Genode happening for at least 10 more years.
No... it's not about money... it's people don't understand the difference between POLA and the way things are done now... until that changes, no amount of money is going to help.
The solution to this problem is known, but nobody seems to know about it...
https://en.wikipedia.org/wiki/...
Why is Microsoft pushing Windows 10 so bloody hard? What financial incentive is there for this?
Algebra is easy.. it's all about getting X alone in the corner, so you can find his value. Geometry should go, along with the foreign language requirement.
I'd change the rating system so that there were multiple dimensions of ratings. Humor, Insightfulness, Correctness, and Offensiveness could all be orthogonal to each other.
The introduction of virtual machines was a huge step forward for computer utilization and security. The ability for a single physical server to serve a diverse set of workloads in a secure and efficient manner made mainframes far more versatile and is a cornerstone of cloud computing.
Uranium fueled reactors are the result of a premature optimization... they aren't reactive enough to work with oxides as fuel.. so you end up having to do all sorts of engineering to try to keep it from oxidizing, whilst only a small barrier away from water. It was never a good idea. The hydrogen bubble that almost made 3 mile island even worse is a result of this chemistry at work. Not only that, when Uranium splits, it only yields 90% of the energy immediately, the remaining 10% takes millions of years, which means a reactor producing 1GW of heat at load will still generate 100 Megawatts when you stop the chain reaction... and if you can't cool it, the thing will melt down.
Thorium yields 99% of the energy immediately, which reduces the need for cooling after the fact by a factor of 10... plus in a Thorium reactor, the fuel is a liquid fluoride, which means you just have to divide the critical mass in the event of an emergency, and you're done with it. A few flat wide steel tanks encased in concrete would do the trick, even if dry.
I'd happily live down the street from a Thorium reactor.
But... what if the secret people with secret lab stuff already did it, and it is red? or maybe anti-oxygen has a red hue for some odd reason?
Have you ever asked yourself what color anti-mercury would be?
So...nothing about a version of windows that doesn't give ambient authority to every line of code that runs... this has a zero percent chance of success.
There were strong warnings in advance of the 9/11 attack, which the Bush/Cheney administration chose to ignore. (Could it be because they needed a "Pearl Harbor" event to catalyze the "New American Century" vision from PNAC?)
They then ignored the reality that Iraq had NOTHING TO DO WITH 9/11, and lied their way to get us into Iraq (in accordance with the "New American Century")
It didn't work, because the full might of the country couldn't actually be used (being based on a lie, and not having full support)
This broke Iraq, and gave rise to ISIS, killing or displacing millions along the way.
My heart goes out to the victims in Paris, and the millions of others our nation has displaced, maimed, or killed in the service of our empire.
Consider yourself in a cyber-war... any line of program you run on your computer can be turned against you... why do you trust any of it with your full authority?
Because you don't have a choice, your OS doesn't give you one. Read up on the principle of least privilege, and the ambient authority model we currently use.
There's no way to specify "run this task with this type of access only to this set of stuff" in Linux... which means you're giving your authority to everything you execute. Until this gets fixed... any flaw in any of the code you run can be used against you.
If you could specify authority and filter it, in a similar manner to unix pipes, you'd be able to build a database that can only take local connections, then build a read-only connection to it, then build a web page that could only connect to that and respond to requests... and finally the web server to take requests from the web and query the page.... and an outside hacker would have to pick through each layer on his way to the database... even if the code was only 99% effective, that's a 99.9999% effective block with very minimal effort.
This type of stuff doesn't have to be user-unfriendly, in fact if implemented correctly it would be fairly transparent to them.
Actually there are 100 choices..
Trust the code
Don't use it
Write your own
Use an OS that sandboxes it by default, and only gives it the capabilities you supply.
If I upgrade to this kernel, how many of those lines of code do I actually have to trust not to give away everything on my machine to a black hat? Hopefully less than 100k lines of code are actually critical.
The grid voltage isn't the problem, it's the DC currents which would be induced on the power lines... transformers can't handle DC bias, it causes the core to saturate, which then causes the inductance to fall, primary currents to rise, and then failure from all the extra current heating the windings.