growth – Handbook work as supply of entropy
Engaged on an opensource air-gapped crypto vault, wherein I’ve the common technique of counting on OS to supply cryptographically safe random supply and use that to generate the mnemonic phrase.
Additionally want to present a handbook possibility of utilizing cube rolls or coin flips to supply your individual entropy and never needing to belief the software program to do it for you. Exploring how trustless it may be.
As, I perceive there are some bias related to cash & dices as properly. They could are likely to favor one or some outcomes greater than different. I’ve learn lots about offering your individual entropy supply and the way that can often be insecure when you by accident introduce some bias. However contemplating, I’m keen to go so far as flipping/rolling 256 instances to generate a 256 bit binary quantity, does that remove the bias?
If sure, then is it doable to make the most of much less variety of flips/rolls to generate much less entropy which is properly verified to be safe sufficient for the aim of mnemonic era? Want to discover this feature as its merely quite a lot of monotonous handbook work for the extraordinarily cautious consumer & if its doable to cut back the burden with out sacrificing safety.
Once more, I’ve learn many statements by fellow individuals on web however I’ve no means to confirm anyone’s declare therefore reverting to this neighborhood to chime in.
A number of the strategies I examine:
Utilizing a number of dices to successfully roll x dices in each try and do it quicker. Unsure how the order of studying dices matter right here. Noticed somebody mentioning they need to be learn in the very same order.
Method of blending coin flip together with 4 dices to make use of the offered mapping to derive the phrases. I do not wish to belief this. I’d very very like to reach at a random index for every phrase after which learn the worth at that index from the wordlist.
Utilizing Ionman’s BIP39 software, the way it exhibits the offered entropy to be sufficient, it certainly required placing in 256 inputs (0/1) random values in there earlier than the safety validation message (mnemonic appears to be like safer than it truly is) goes away. In order per that software, something lower than 256 flips/rolls is not safe sufficient. True?
Tips on how to interpret cube rolls: Then there may be additionally the query of deciphering the cube rolls. Do I take advantage of odd even numbers as 0 & 1 respectively or ought to I take advantage of half the vary of doable outcomes (1-3) as 0 or 1 and the opposite half vary as the opposite binary bit?