Unique context-based identifier

UCID is a structure for creating ID strings, which contains a context hint (like "user" or "session"), has a timestamp to minimize conflicts and has a good readability.

Structure

context: string, lower case, only letters [a-z], min 2, max 8 characters timestamp: Unix timestamp in milliseconds, 11 hex digits, 44 bits random part: Random generated number, 5 hex digits, 20 bits

Decisions

How should the hex part be represented

The question was, should the hex part of the ID split, and when yes, how often.

user_0123456789abcdef
or
user_01234567_89abcdef
or
user_0123_4567_89ab_cdef

To make it easier for the developer to see the individual parts, 4 character splits were choosen.

Should the hex part be split?

"_" or ":" as separator The question was, what separator would be better to use.

user_0123_4567_89ab_cdef
or
user:0123:4567:89ab:cdef

It has been decided to use "_" as separator. The reason for this were:

• Systems often use combinations of letters, numbers and underscores as possible ids
  • We don't want to have struggles with collision, event though the risk of colliding with a system is low
• The individual parts are more easily readable