IEEE-standard decimal64 occupies the same number of bytes as int64, and the maximum monetary base of Freicoin has already been adjusted to fit within the precision of the decimal64 type, allowing exact representation of all existing values. Internally, arithmetic would have to be done exactly using arbitrary-precision rational arithmetic and time-value adjustments truncated at 16 digits.
Multiple transaction outputs could be used if greater than 16 decimal digits of precision are required, otherwise the exact remainder would be added to the transaction fee (and if the miner chooses not to claim it, it's destroyed).
This would be a candidate for a BIP proposal and submission upstream to the Bitcoin Satoshi client.
IEEE-standard decimal64 occupies the same number of bytes as int64, and the maximum monetary base of Freicoin has already been adjusted to fit within the precision of the decimal64 type, allowing exact representation of all existing values. Internally, arithmetic would have to be done exactly using arbitrary-precision rational arithmetic and time-value adjustments truncated at 16 digits.
Multiple transaction outputs could be used if greater than 16 decimal digits of precision are required, otherwise the exact remainder would be added to the transaction fee (and if the miner chooses not to claim it, it's destroyed).
This would be a candidate for a BIP proposal and submission upstream to the Bitcoin Satoshi client.
Forked from the discussion surrounding GH #67.