TRB-I Addressing Scheme Proposal

Since discussion of the Ideal Bitcoin progressed some in the Republican forum today, and since I can't say I'm a great fan of the current addressing scheme, here's a proposal for an alternative implementation :

A TRB-I address under this scheme would be composed as follows :

1. To a one byte versionⁱ is added a 512 byte public key.
2. The 513 byte numeral from the previous step is put through MPFHFⁱⁱ, obtaining a numeral of an unspecified length.
3. The numeral of unspecified length from the previous step is put through keccakⁱⁱⁱ, resulting in a 256 bit output^iv.
4. A one byte checksum is calculated as the successive xor of the 32 bytes from the previous step and added at the end, resulting in a 33 byte output.
5. The 33 byte output from the previous step is put through base64^v. The result is a TRB-I address, I guess you can prepend B to it or whatever.

Comments welcome.

1. Version 0 is 4096 bit RSA (as discussed). Other versions may be implemented, up to a total of 255 alternatives. If the implementation uses less than 4096 bits for its public key, it is required to also provide a padding scheme as part of its specification. [↩]
2. MP's Fabulous Hash Function. [↩]
3. Specifically no NIST "standards" are being contemplated or seriously considered. All those who seriously believe USG crypto may or should be taken seriously are rank imbeciles who can't, nor should, be taken seriously. [↩]
4. The keccak implementation will be configured to actually offer 256 security bits ; in any case the number of rounds used will not be less than 64. [↩]
5. And let the idiots who run terminals with bad glyphs buy better hardware / run better software. [↩]