OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

xri message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: RE: [xri] versioning of authirity section [Issue 5]

More specifically, I have several open questions and points w/r/t versioning naming authorities:

1) What implication does this have for equivalence rules? (not specific to versions-in-the-naming-authority) That is, are two URIs with different version tags (or one with a version tag and one without) *ever* equivalent? URI schemes should specify equivalence rules - generally the default is "syntactic character-by-character equivalence" (usually with some ignorance of case for domain names)

2) What implication for the resolutoin process (RDDDS for URIs). I see two cases:

a) Version parts are interpreted "opaquely" -- that any version string (including the ':', ',', etc leading characters) are essentially unparsed by a resolver. In this case, i think DDDS-style resolution isn't really affected.  

b) Version parts need to be "canonicalized" or otherwise parsed to make sure they are resolved correctly (from a semantic point of view). 

It seems to me that #2 is the more likely situation since how does a resolver understand that a version of the naming authority 'as of January 20th' is the same one as that which is in place today? DDDS (and RDDDS) do resolution based on pattern matching - it seems to me that versions in the naming authority parts would totally eliminate DDDS as a viable resolution mechanism. I don't think versioning breaks the HTTP resolution, but it may break some caching semantics (because things which are not syntactically equal may in fact be semantically equal - need to think this through). 

3) I tend to think the distinction between XRI URIs (versioned namespace authorities) and XRI URNs (not versioned, since I think this breaks the semantics of "once and forever assigned..") makes sense. 

Maybe the answer to #2 is that versioning is important in XRI URIs and that DDDS simply isn't an option there? 


> -----Original Message-----
> From: Lindelsee, Mike [mailto:mlindels@visa.com]
> Sent: Monday, April 21, 2003 10:38 AM
> Subject: RE: [xri] versioning of authirity section
> I agree completely that versioning the naming authority could 
> be useful in a number of contexts, but I'm still trying to 
> understand what exactly that would mean when it comes to 
> resolution.  For instance, how would versioned naming 
> authorities work in a DDDS-style resolution?  
> Mike
> > -----Original Message-----
> > From: Peter C Davis [mailto:peter.davis@neustar.biz]
> > Sent: Monday, April 21, 2003 8:38 AM
> > Subject: [xri] versioning of authirity section
> > 
> > 
> > (In case you haven't noticed, I am finally catching up on xri 
> > activities :-)
> > 
> > line 159:
> > [ISSUE: It would be nice if naming authority parts were NOT 
> > versioned to 
> > make resolution simpler. Then, only the  local part  segments have 
> > versioning info in them, and interpretation of this versioning 
> > information would be local to the client and ultimate  leaf  
> > directory. ]
> > 
> > However, for URI-type XRI's, it MAY prove usefull to allow 
> > versioning of 
> > the authority part:
> > 
> > xri://some.domain.int[;2001-03-04T20:15:40Z]/foo/bar
> > 
> > esp. since URIs are re-assignable. This allows for "This was the 
> > authority at time T".
> > 
> > --- peterd
> > 
> > 

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]