There is a benefit to throwing a die a 100 times, test the sequence, and if no evidence of bias is found, use that same sequence for the bitcoin address generation. What benefit? Well, I have personally had a situation where I found out on the first day that there was an evidence of bias. But on another day with the same die but a different surface, there was no evidence of bias. So it can be useful to just use the same sequence for both testing and, if the test is good, real-world use. Hence this tutorial.
I have tested these steps on Ubuntu 14.04 LTS only, but I suppose they will work on other Ubuntu/Linux systems as well.
Here are the steps:
- On a virtual machine (or a real machine if you want) that is allowed to connect to the internet, do a clean install of the same system that is on your offline computer. In my case, that is Ubuntu 14.04 LTS.
- On this fresh online system, download this Python script by Alan Reiner (the originator of Bitcoin Armory).
- On the Terminal, run python apt-get-offline.py r-base.
- Use a USB key to get the resulting download_r-base folder onto your offline computer.
- On your offline computer, use sudo dpkg -i *.deb in that folder to install R on your offline computer.
- You can now use R on your offline computer by simply running R on the Terminal.
See also my other tutorials:
Tutorial: Using R to statistically test a die for bias
Tutorial: Compiling Armory and getting it onto an offline computer
Tutorial: Creating a bulletproof 255-bit entropy Armory wallet