Quote:
Originally Posted by Farling
When I first started writing the Import tool I tried examining the UUID pattern that RW was using, and I couldn't marry up the UUIDs in the RW export files to any particular RFC-based UUID pattern.
My hope was that I could use the UUID pattern that included the date/time of creation of the UUID in order to create what would be a guaranteed UUID for the generated import tool. But without a guarantee that RW uses standard UUID patterns I wasn't able to guarantee that any self-generated UUIDs wouldn't conflict in the future with a UUID created internally by RW.
|
RFC 1422 v4 doesn't use a date/time stamp at all (it uses a pseudo random number instead) and v5 hashes the date/time so the only identifiable thing would be the version info embedded in the UUID, the leading 5 in group 3.
Anyway the whole point of UUID's is that the domain of possible values is so large it is unlikely that two UUID's would ever be identical. The issue became that with widespread use across the net the date time stamp versions relied on no two UUID's generated at the same millisecond (IIRC, whatever the resolution of the time used is) which became increasingly possible. The newer versions should prevent that.