RE: Duplication of provisional URI namespace tokens in 2717/8-bis

I suggest that the Internet world is better off to the extent that URI
schemes fall into  well-demarcated classes:

Wild-type:
  - No formal approval, recognition, agreement, consensus... 
  - People will do what they will for experimental or closed system use.
  - no protection against token-duplication, no official sanction,
unlikely 
    application take-up, no defensible claim of precedence

Provisional:  (assuming unique tokens)
  - Low barrier, lightweight entry into globally-registered schemes
  - Deployers benefit from the controlled namespace and protection 
    from duplication of tokens
  - schemes are easily discovered in a single registry
  - migration path to higher technical review
  - no guarantee of long term maintenance, so application take-up is 
    unlikely, but communities can do specialized applications without 
    fear of name collisions
  - availability of this path reduces pressure on IESG to approve 
    new schemes, while affording a reliable environment to support
    innovation 
  - deployment failure or lack of takeup results in minimal or no mess 
    to clean up 

Permanent: 

  - Full technical review assures compliance with global Internet 
    philosophy and standards
  - Imprimatur of IESG makes take up by Web applications more likely
  - Stability of time-proven schemes of demonstrated utility a prime 
    criterion

Received on Wednesday, 12 January 2005 15:43:05 UTC