OGSA-Naming-WG Meeting at GGF 14

Minutes by Mark Morgan (University of Virginia)

* Status
	- Working group approved week and a half ago by steering comittee
	- RNS was taken out for time being until it gets through public comment period (realistically about 3 months from now)
		-- Still a few concerns about file system'y things in there
	- WSNR -- WS Name Resolution moving forward rapidly
		-- Still a little discussion on "AbstractNames"
* Review of what Naming is for and the basics
* 0 or 1 Resolver, or do we want 0 or more Resolvers
* Should we allow more then 1 abstract name within an EPR?
* Trying to get consensus
	- Should AbstractNames be part of the spec?
		-- Well, couldn't we have two specs...one for abstract names, one for resolution
		-- Why split the specs?
		-- The sole objection here is, does including AbstractName delay resilient resolution.
		-- We will give ourselves so many weeks to figure this out.  Then we can split them if things are looking like they aren't going to happen in time.
	- We need to look close at WS-Addressing
	- What about putting a "type" attribute onto the name:AbstractName element.
	- Three things could be in the abstract name element
		- A string and only a string
		- A string with a given, well defined format
		- xsd:any
	- What about having a URI with a well known scheme (Andrew's scheme) that is uri:andrew:and some base 64
	- Keep it single with a structureless string, OR a richer abstract name on which additional structure could be layered....
	- Votes
		Can we live with IRI's as our abstract names
	- We have consensus that it should be a URI/IRI thing.
	- Should we allow more then one abstract name in an EPR
		- We can keep this as long as we keep the equivalence check simple

On soon to come telecons we will talk about whether or not to have multiple abstract names and who binds the abstract name to the EPR.