VOSpace vs WebDAV

[Norman Gray]

Greetings, all.

Is this question of persistent identifiers going to be discussed in Madrid next week?  I've cc-ed @Alberto, in case this is on the agenda.

On 2014 May 15, at 13:20, Markus Demleitner <msdemlei at ari.uni-heidelberg.de> wrote:

> On Thu, May 15, 2014 at 11:16:49AM +0000, Paul Harrison wrote:
>> ubiquity of this protocol and the ivo: and vos: just do not have
>> the client support - if we could reinvent/reinterpret ivo: and vos:
>> URNs via redirects off a central http://ivoa.net/ service URL  so
>> that you could use curl to get resources, I think that it would be
>> a very good thing?
> 
> I don't know about vos: yet (though my feeling is that if we can do
> what we do with VOSpace using WebDAV+conventions, I'd find that
> nifty), but for ivo: that central redirector (that would in principle
> be trivial) won't work as there's not nearly a 1:1 relation between
> ivo: and https?:  

A possible way out of the conundrum is ARK.

The Wikipedia page at <https://en.wikipedia.org/wiki/Archival_Resource_Key> is good, but fails, I think, to adequately communicate the simplicity of the idea at the heart of the approach.

An ARK is something like

    ark:/12345/my-identifier-for-a-thing

The /12345 is a centrally-registered identifier for the organisation whose identifier is listed.

The clever thing is that this is _not_ yet another URI scheme.  It's intended to be dereferenced by being appended to a 'Name Mapping Authority Host' (NMAH), such as:

    http://ivoa.net/ark-service/ark:/12345/my-identifier-for-a-thing

This seems pretty obvious, and in retrospect it is, like all the best ideas.  But...

The point is (1) The NMAH is arbitrarily replaceable, so that

    http://ivoa.net/ark-service/ark:/12345/my-identifier-for-a-thing
    http://esa.int/ivoa-inherited-ark-service/ark:/12345/my-identifier-for-a-thing

are deemed to be equivalent.  In theory, these could also be transferred to some future HTTP replacement.

Also (2) the ark:/.... can be curated or stored separately from the lookup service (NMAH) it's appended to.

Also (3) There's a mechanism for indicating immediately how persistent this object is, or is intended/expected to be, though that's of less significance for our context in that we perhaps hope that things will be indefinitely retained.

Though I think it would be a good idea, in principle, to use 'ark:' identifiers for the IVOA, this sort of trick could be repurposed for an 'ivo:' resolution service.  That way, the IVOA gets the best of both worlds.

All the best,

Norman


-- 
Norman Gray  :  http://nxg.me.uk
SUPA School of Physics and Astronomy, University of Glasgow, UK