Re: [xri] Issue 1: Clarifying * Semantics

[Fen Labalme <fen@idcommons.org>]

Loren West wrote:
> Fen wrote:
> 
>> Note that :12:34 would still be a legal persistent identifier, it
>> just would not imply a separation (or delegation) between two
>> parts.  In other words, it is similar to the identifier :12.34 
>>(using the new semantics for dot as a regular character).
> 
> *******
> You raise an important point here - that the proposed change is not
> backwardly compatible with 1.0 XRI syntax.  In my opinion, changes that
> aren't backwardly compatible should go through a more rigorous process
> for approval.  
> 
> Does anyone know if there's precedent for this on other Oasis committees?

1) I would think there must be such a precedent, as, for example, changing dot 
to star is non-backwards compatible, and I believe that has been agreed to 
(though not yet voted on) already.  Also, I believe some on the changes made 
to SAML were not backward compatible, but I don't have a reference at hand.

2) I wonder if, in Epok's current application and usage, something like the 
string :12:34 is actually using delegation, versus being taken together as an 
opaque string?

2a) If delegation is being used, is this in the authority-path (before the 
first slash) or the local-path part?  My understanding is that if it is in the 
local-path segment, it is up to the local resolver to decide how to interpret 
the string, so you would be free to continue to consider colon as both 
decoration and separator in the local-path.

Just trying to make things easier for you, as simplifying the semantics 
certainly makes things easier for me - especially while explaining the 
meanings of the various symbols to early adopters.

Fen