Understanding IVORNs

[John Swinbank]

Hello,

On 15 Aug 2014, at 8:55 , Mike Fitzpatrick <fitz at noao.edu> wrote:
> On Thu, Aug 14, 2014 at 6:35 PM, John Swinbank <j.swinbank at uva.nl> wrote:
> 
>> Based on the above, I have a two part question:
>> 
>> 1. Does the “local_ID” following the “pound sign" in the example given in the VOEvent standard constitute part of the identifier or not?
> 
> My interpretation is "no", i.e. the intent is that the ivorn up to the '#' is a resource identifier that can be resolved in the Registry to some repository or event stream that when presented with the local_ID (via some TBD access protocol) will return the corresponding event packet identified by the 'local_ID' fragment of the URI.
> 
> Now, it also depends on what you mean by 'identifier' in your question:  If you mean is local_ID part of the identifier for the packet then my answer is yes, if you mean is it part of the identifier for the service/resource then no. 

[…]

Thanks; agreed.

>> 2. If so, do the same limitations in terms of reserved characters apply to the local_ID as to the rest of the IVORN?
> 
> According to the URI doc (RFC 2396:  http://www.ietf.org/rfc/rfc2396.txt), they do (see Sec 4 on fragments which says syntax follows reserved chars from Sec 2).

If the ‘#' terminates the the identifier, then we could argue that following that we revert to the generic RFC 2396 restrictions. If the fragment following the ‘#’ is part of the identifier, then I agree that the more stringent requirements continue to apply.

Taking a specific example, what do we think of

  ivo://nasa.gsfc.gcn/SWIFT#Point_Dir_2014-08-15T18:35:00.00_16857763-673

?

A “:” is a reserved character which should not appear in either part (authority identifier or resource key) of an IVOA identifier; here, it appears only in the fragment following the #. Is this ok?

Thanks,

John