As alluded to in the results of the recent survey on MIRIAM URIs
some respondents mentioned the use of resolvable URLs instead of URNs for
MIRIAM identifiers. This topic has been raised in several discussions, on
various medias, and centres on the the suspicion of an incompatibility of
MIRIAM URIs with the linked data agenda of the W3C.
With a growing misunderstanding between data identification and data
location has come a large endorsement of PURLs as URIs directly in
metadata. This may jeopardize or hamper the adoption of MIRIAM identifiers,
and therefore one of the goals of MIRIAM itself, i.e the unambiguous and
perenial identification of concepts used in knowledge representation. We
think this is an issue that must be tackled fairly urgently.
Consequently, we are considering developing a (persistent) URL version of
MIRIAM URIs. It is important to note that:
* These URLs will not be a replacement of the URNs, but an alternative
version of the identifiers. People will be able to use both. MIRIAM
Resources will offer similar services with both identifier types, and will
allow their interconversion. The structure in domains, subdomains and
data-set identifiers will be conserved, as well as the strings currently
used for those.
* These URLs will not be a return to the initial unstructured URLs of 2005.
They will follow the structured approach we used for the current URNs.
* These URLs will not redirect directly to the data-set distributed by a
specific data resource but to a document providing a choice of location(s)
where the data-set is stored or described.
Such URLs should facilitate the adoption of MIRIAM identifiers by people
concerned by possible compatibility problems between the current
identifiers with linked data and semantic web initiatives. It will offer a
solution for people who want to use PURLs, but are at loss when choosing a
PURL for cross-reference when several resources are distributing the same data.
With this message, we invite you to provide input on this idea. For example:
Is this a good plan? How should the URLs be structured? Should they use a
root 'purl.org', 'miriam.org', 'miriam.purl.org', 'purl.miriam.org', or
something completely different? What should be the target of the URL
redirection? What should we put in the document targeted? All the possible
resources? A resource chosen based on a given criterion (geolocalisation of
the caller, reliability, etc.)?
Thank-you all for your support and help.
Nicolas LE NOVERE, Computational Systems Neurobiology, EMBL-EBI, WTGC,
Hinxton CB101SD UK, Mob:+447833147074, Tel:+441223494521 Fax:468,
Skype:n.lenovere, AIM:nlenovere, twitter:@lenovere