Re: doi equivalent?

The two suggestions so far are PURLs and OpenURLs.

I'm pretty sure openURLs are out. As I understand it these basically
summarise the metadata associated with a publication and depend on one
metadata item already being a unique id (eg an ISBN), which is just what
I don't have. I'm also not sure about the interaction I'd have with link
resolvers which don't recognize my site as a valid data source.

PURLs sound more promising, but I still don't think they're quite right.
They assume a permanent location associated with a particular
publication, whereas if the ebooks I generate are free I don't care what
their location is (in fact it would be better for me if others host
them), though I do care if two copies in different locations are the
same (and so have the same uniqueid). I think I want two things: a
unique id independent of location (a URN), which would be stored in the
epub metadata, and a PURL which would give a constant reference to my
particular copy of the epub, for use by library link resolvers.

And I still haven't found a suitable URN type. ISBNs, NBNs and ISSNs are
out. So are DOIs. OIDs (http://www.oid-info.com) sound possible, but I
can't work out how to find a registering authority. So far of the ones
I've looked at UUIDs seem the best as not needing a central registration
authority, just ironic they're a Microsoft-derived standard.

How does CKAN id its entries? Or does it just reference external urls
and rely on the external hosts to manage their own URNs?

Graham

Peter Suber wrote: