blueshift 0.0.1

Quickly create large amounts of unique serial numbers.

blueshift

Quickly create large amounts of unique serial numbers.
Forked from JosephTLyons/andromeda

Creating Unique Serial Numbers

Documentation has not been completed at this time.

How it Works

1. Check to make sure that it is possible to create the amount of unique licenses requested based on the the serial number length and character set.
   For example, if the user requests 1,000,000 unique licenses, but only asks for serial numbers of length 4 and wants only numbers as characters, it will not be possible, as 10^4 < 1,000,000. We must have more possible combinations than the number of licenses requested in order to ensure all licenses will be unique.
2. Make a characterList that is comprised of all the symbols the user wishes to use in their serial numbers (numbers, lowercase letters, uppercase letters, and symbols).
3. Shuffle this list and push a copy of it into another list, called listOfCharacterLists. This step is executed as many times as the length of the serial number requested.
4. Create a new list of numbers called indexList. Each cell in this list holds an index that corresponds to a characterList in the listOfCharacterLists.
5. A serial number is generated by using the indexList to provide an index to each of the corresponding internal characterLists in listOfCharacterLists. The listOfCharacterLists contains as many internal characterLists as the length of our serial number. This can be seen sort of as a speedometer that starts with 0000 miles. As we roll through the miles in the rightmost column, it resets back to zero and the next value to the left increments by 1. This is exactly how the indexList and listOfCharacterLists work to obtain serial numbers.

In order to avoid printing serial numbers that look like this:

PWZdySrxntWyBxYYZAlT
PWZdySrxntWyBxYYZAlB
PWZdySrxntWyBxYYZAlZ
PWZdySrxntWyBxYYZAlJ
...

we need to pull serial numbers evenly over the entire spectrum of possible combinations. We can do this by taking the total possible combinations based on the user settings and dividing it by the number of licenses needed. This will give us the value that we need to be apply to the indexList before printing the next serial number to the file. Using the example from earlier:

Serial number amount: 1000
Serial number length: 20
Enter 'y' to use numbers: y
Enter 'y' to use uppercase letters: y
Enter 'y' to use lowercase letters: y
Enter 'y' to use symbols: n

Our licenses are of length 20 and use 62 types of symbols (10 numbers + 26 lowercase letters + 26 uppercase letters). We have 62^20 possible unique serial number combinations. Because we want 1000 licenses, the spacing between serial numbers to be printed to the file is (62^20)/1000 = 704423425546998022968330264616370. If we set n = 704423425546998022968330264616370, then the function increaseIndexVectorBy() efficiently applies this number to the indexList in roughly log(n) with base 62 operations. Using this method, we are able to print licenses that are equally spaced apart in the total combination space, as shown in the first sample output earlier.

We can even print all of the values of the indexListr to the terminal to see the patterns better by uncommenting the function printIndexList():

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 03 52 20 20 51 36 12 55 34 13 55 03 29 16 22 50 36 43 40
00 07 42 40 41 41 10 25 49 06 27 48 06 58 32 45 39 11 25 18
00 11 32 61 00 30 46 38 42 40 41 41 10 25 49 06 27 48 06 58
00 15 23 19 21 20 20 51 36 12 55 34 13 55 03 29 16 22 50 36
00 19 13 39 42 09 57 02 29 47 07 27 17 22 19 52 04 59 32 14
00 23 03 60 00 61 31 15 23 19 21 20 20 51 36 12 55 34 13 54
00 26 56 18 21 51 05 28 16 53 35 13 24 18 52 35 44 08 57 32
00 30 46 38 42 40 41 41 10 25 49 06 27 48 06 58 32 45 39 10
00 34 36 59 01 30 15 54 03 60 00 61 31 15 23 19 21 20 20 50
00 38 27 17 22 19 52 04 59 32 14 54 34 44 39 42 09 57 02 28
00 42 17 37 43 09 26 17 53 04 28 47 38 11 56 02 60 31 46 06
00 46 07 58 01 61 00 30 46 38 42 40 41 41 10 25 49 06 27 46
00 49 60 16 22 50 36 43 40 10 56 33 45 08 26 48 37 43 09 24
00 53 50 36 43 40 10 56 33 45 08 26 48 37 43 09 26 17 53 02
00 57 40 57 02 29 47 07 27 17 22 19 52 04 59 32 14 54 34 42
...

Pitfalls to be Aware of

Andromeda does not care if you choose settings that result in a very low pool of license combinations. You should be aware of this. If you run the application with the following options:

Serial number amount: 1000
Serial number length: 4
Enter 'y' to use numbers: y
Enter 'y' to use uppercase letters: n
Enter 'y' to use lowercase letters: n
Enter 'y' to use symbols: n

the output will be:

9444
9474
9494
9464
9484
9434
9424
9404
9414
9454
9244
9274
9294
9264
9284
9234
...

Notice that the licenses are fairly similar. Also, note that it would be fairly easy to guess a serial number. The probability that a random guess would be an actual serial number is 1000/(10^4) = 0.1. It is up to the user to understand this and adjust the settings to increase the complexity of the output and decrease the chances of guessing a license number. Using the example from earlier with 1000 serial numbers of length 20 using all symbols, the probability that a random guess would be an actual serial number is 1000/(62^20) = 1.4196007e-33.