New! Unbreakable Security Locker Algorithm Demo

[Rodwell]

Hi guys,

I have designed a security locker page with a android pattern like GUI, except it doesn't use pattern style but rather number pad format :(
The design has a nice looking interface with animations and with this I think that's a good start to making persistent locker/security apps like Vault.

The new input code( "7character" ) is ciphered with Caesar Cipher I've rewrite in JavaScript and is run through an encryption algorithm ( "RodoAlgorithm" ) I've made which have several functions to encrypt the code, the final code generated is a "128character" security code.
include:

• Patternize - Returns a string of the pattern form of the given code.
  
• Ramble - Uses an array of alphanumeric letters to replace identity code letter with new ones.
• etc...

I haven't finished the code yet because of other projects I'm doing so I will attach a demo spk file later( internet slow ) and see if you guys like it.

;) im out!

[Rodwell]

Take a look at the spk file here.
Tell me what you think and any ideas of making a pattern out of it.

[Netpower8]

the pattern design can be improved

[Rodwell]

yes and we can make it a "draw pattern" locker :)

@netpower8 -you come up with an improved pattern or other ideas we can put on it.

How is the encryption system?

[Netpower8]

Dont have time to check the encryption yet. But just a glance i saw you have a key map of sort. And base on this its hard coded. I had my own idea (and created it already) but not so good. But the key map can be generated you just need to know the seed key and it generate the key map. Having it hard coded might not be a good idea

[Netpower8]

There is also a sample pincode code somewhere in the forum you might want to check it out also. This could be a start for a draw pattern program code

[luke x3]

It's not too bad on the crypto side, you may need to check the ramblecode function is working properly, it looks like it isn't using the letters table
Net is right, encrypt/decrypt should have a variable key, and it needs to be quite big these days as computers can check hundreds in no time if they get the chance

For passwords it's common to use a good one-way encryption instead, so there's no worry of decrypting the passwords, but these can be tricky to get right

[Rodwell]

@Netpower8

I wouldn't say its hard coded but actually sequential steps(algorithm) for handling string arrays meaning that each step can be replaced with a completely different step as long the returned value is a string array.

I have seed key as a global variable, I used a default key for the demo.

Thanks for checking the code out ;)

[Rodwell]

@luke x3

oh yes I figured :( the ramblecode function works with the letters table, I want it shift letters by pointing out the identical letters and shift them with new letters but as #netpower8 said the code was becoming hard coded so I kept it simple and not too complicated. You can replace ramblecode with a better string handler.

I used a default password key(13) for the demo which gives out a < 128 character long string, I can make it big to generate 512 longer password and key should probably be a 6-8 numbers long.

Yes, the decryption is not an issue, a matching password is by matching the encrypted entered passwords with a stored encrypted password.

[Netpower8]

I havr not check your code but there should be another table to prevent some guessing. Example. User will try to guess the replacement code by giving a duplicate letter. Like paaaassssword. By doing this. Try look for same replacement code a and s.

[Netpower8]

here's an idea... it was done a long time ago during my turbo pascal programming days... and the code is lost. but you can replicate the idea

the output ramble code should have a signature code which computes the ramble code. the signature code differs since its dependent on the ramble code

if the ramble code has been altered then the signature code will not match the ramble code therefore the ramble string should not be accepted.
if the signature code has been altered the signature code generate with the ramble code wont match either therefore the ramble string should not be accepted.

either way. alteration on the string wont match the each other therefore the ramble string will not be accepted...

this method was used to for an executable code in windows where a signature code checks the executable if it was altered (or infected by virus) and refuse to continue if it is.

[luke x3]

Having a little play around today, thought this was kind of interesting

function crypt_lib(){
  this.hex = {};
  //-------------//-------------
  // cryptlib.hex //
  //-------------//-------------
  this.hex_init = function(){
    alert("--cryptlib.hex_init");
    var vhex = this.hex; // <---------- vhex
    var hexinit = false;
    //-------------
    vhex.num2bit = function(x, n){
      //...
      }
    //...
    }
  //...
  }

var cl = new crypt_lib();
cl.hex.num2bit(x,y);    

[Rodwell]

:-0 idea sounds interesting, I will try to make something like that or any function that implements the idea of using a signature code as a fingerprint.

I haven't been working on the encryption code since my last post (not long) but rather I am working on the pattern locker. Not long I didn't have a clue on where to start but
I think I might have broke the code, found a way to make a pattern using the "processing library" its actually cool.

I'm almost there, just need to figure out how to join the pattern lines.

anywhere if you have time, maybe you can finish the encryption code whilst I finish the pattern then we can see how cool the program will look :)

[Rodwell]

I've finally done it, the Android pattern locker I was working on is finally finished ;)

The interface, the functionality is exactly as the default android pattern locker bet you guys can't see the difference.

Will attach the SPK!

[Rodwell]

I have created a new topic on the pattern lock and you can download the demo app which includes all two lock systems.
here:
https://groups.google.com/forum/#!topic/androidscript/dyfHKddI6lQ