Questions from Christopher Reay in two sections, creditCommonsProtocol, and existing implementation
this is a living issue description for Christopher Reay
- Comments on any of the things raised here are very welcome
- clear answers will be integrated into the description
questions about the protocol
general section
- what would it mean to add a ledger to a parent
- it seems that there are many situations where this is possible, however it seems implied that there is always defined the original root ledger.
stuff I got from looking at the database specification
- what is a
super annuated transaction
? - what does the word
quant
in the database mean?
questions about the implementation
- current UUID appears to be implemented by Mat in source here, which implies it is copying the UUIDv4 protocol (anonymous, no IP information embedded)
- v4 UUID's provide no guarantee of uniqueness, and must be checked.
- UUIDs are designed for machines operating on the internet. "unique" UUID guarantees are based on MAC and IP address analysis, which is not anonymous.
- the UUID specification is not suitable here
- obviousle (lol) I have my own format which is more or less impervious to this problem, which adds a nanosecond timestamp to a UUID like string. We can discuss this later
refactoring suggestions
- primary transactions could be in their own table with dependent transactions in their own table. It is only the primary transactions that change state. Also I believe that yesterday Mat was mentioning that there was a
Edited by Victor Vorski