mctrivia said...
I am not suggesting removing 000000 from the file in that way. instead if the 16k cluster contains 000000 delete it and start over this will keep 0000 the same odds
Think it through.· It won't keep it the same odds.
same odds of 0000 just not same odds of 000000. unless you are considering 0000 inside 0000.
statistically I am only deleting 1 in 128, 16k clusers removing clusters with %00 0000 0000 0000 0000 0000 or %11 1111 1111 1111 1111 1111
very few. greater then 21 still possible on cluster edges
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
well when I write the code I will leave both options available. true random or true random with retry if long string of bits
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
Remember that the Germans 'killed' randomness with the way they used their Enigma machines.
The operator used his codebook to find that day's setting(which wheels go where, starting positions, and patch cables that needed to be moved... ). Then, somewhere in the beginning of the message, he sent NEW SETTINGS for the wheels, and coded the rest of the message with that code.
Now, they did this in the hope that the 'common' part of that days messages would be too short for anyone to ever break the code, and also that this 'trick' would confuse anyone even trying to crack it.
(Remember, the operator manual was a classified document, and it was only by chance that the French got hold of it, and passed it onto the Polish, who again passed it on to the British. In other words, security by obscurity)
Their next mistake was that the operator had to send the new settings TWICE. In succession...
(That's German thoroughness. No danger of the receiver missing the important settings... )
So, they ended up with a lot of messages with what looked like random symbols, but which had a specific pattern to the plaintext, at or near a specific position.
Mucking about with trying to eliminate 'long runs' of similar output or input symbols will be just as disastrous. (You will force specific patterns into the coded output and sooner or later, someone will notice)
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Don't visit my new website...
i don't think mathematically the removal of 1 in a million odds event gives any basis to crack it. Even i tell a cracker that the key is true random but you will never see more then 21 0s in a row it would not help. they could assume from that information that any plain text they saw was indeed not what it looked like unless very short. but knowing something is not 1 thing and could be any of 1 billion things does not help. knowing not 1 thing but could be 1 of 10 things that would be a help.
Now if I was doing as some suggested not allowing patterns to happen then this would severely weaken it because you would remove many options from the list.
But for those that do not want to remove even 1 option I will make sure any code I write has that as an option. Just got home now so after work tomorrow I will be starting coding on randomness experiments.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
mctrivia said...
same odds of 0000 just not same odds of 000000. unless you are considering 0000 inside 0000.
statistically I am only deleting 1 in 128, 16k clusers removing clusters with %00 0000 0000 0000 0000 0000 or %11 1111 1111 1111 1111 1111
very few. greater then 21 still possible on cluster edges
I give up.
Randomness and variation are _extremely_ difficult concepts for people, I think largely because there's survival value in being mistaken about randomness and using that wrongness to drive behavior that is at times wildly overcautious and at other times wildly overoptimistic. I understand that fortune tellers are having no problem at all with the slow economy, and people remain impressed by seers who only "predict" events that have already happened.
We're also overly devoted to the notion that effectiveness requires intentionality, so much so that apparently millions of people are convinced that anything nonintentional is completely random.
I have no idea how we'd manage it, I'd sure like to see much more effective general mathematics education on randomness and variation.
Post Edited (sylvie369) : 4/14/2009 10:05:57 AM GMT
Comments
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
· -- Carl, nn5i@arrl.net
statistically I am only deleting 1 in 128, 16k clusers removing clusters with %00 0000 0000 0000 0000 0000 or %11 1111 1111 1111 1111 1111
very few. greater then 21 still possible on cluster edges
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
· -- Carl, nn5i@arrl.net
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
The operator used his codebook to find that day's setting(which wheels go where, starting positions, and patch cables that needed to be moved... ). Then, somewhere in the beginning of the message, he sent NEW SETTINGS for the wheels, and coded the rest of the message with that code.
Now, they did this in the hope that the 'common' part of that days messages would be too short for anyone to ever break the code, and also that this 'trick' would confuse anyone even trying to crack it.
(Remember, the operator manual was a classified document, and it was only by chance that the French got hold of it, and passed it onto the Polish, who again passed it on to the British. In other words, security by obscurity)
Their next mistake was that the operator had to send the new settings TWICE. In succession...
(That's German thoroughness. No danger of the receiver missing the important settings... )
So, they ended up with a lot of messages with what looked like random symbols, but which had a specific pattern to the plaintext, at or near a specific position.
Mucking about with trying to eliminate 'long runs' of similar output or input symbols will be just as disastrous. (You will force specific patterns into the coded output and sooner or later, someone will notice)
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Don't visit my new website...
Now if I was doing as some suggested not allowing patterns to happen then this would severely weaken it because you would remove many options from the list.
But for those that do not want to remove even 1 option I will make sure any code I write has that as an option. Just got home now so after work tomorrow I will be starting coding on randomness experiments.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
Post Edited (mctrivia) : 4/14/2009 8:01:19 AM GMT
We're also overly devoted to the notion that effectiveness requires intentionality, so much so that apparently millions of people are convinced that anything nonintentional is completely random.
I have no idea how we'd manage it, I'd sure like to see much more effective general mathematics education on randomness and variation.
Post Edited (sylvie369) : 4/14/2009 10:05:57 AM GMT