HCLSIG BioRDF Subgroup/Tasks/URI Best Practices/Recommendations/Purls
From W3C Wiki
[[/../../Recommendations|URI Note main page]]
What about purl.org?
Are http://purl.org/ URIs the answer to HCLS's URI stability wishes? Should we recommend - endorse - the use of purl.org?
People are using it. That's fine. But no one is issuing anything that smells of an official endorsement for HCLS as visibly as the proposed URI note would.
If we end up discouraging ephemeral URIs ([[/../AttitudeTowardMigration]]), then we are morally obligated to provide a reasonable and concrete alternative, and this is the only one we know about.
The purl infrastructure is going to be worked on over the next year. Maybe it will get better and some of the worries listed below will go away.
If we accept that we can deal with ephemeral URIs using the resolution ontology, then this question becomes much less critical.
- Purls are not very spontaneous - publishing requires an extra unfamiliar step
- yes, it's very easy, but barrier to entry is perceived to be high
- will/should HCLS community trust purl.org? (who is this OCLC, anyhow?)
- sometimes purl.org is down
- purl.org doesn't do 303s or 307s
- weak security
- are not necessarily maintained - they may break and become useless, with no way to obtain authorization to repair them
This last point is the deepest. The maintenance of a purl is not only in the author's interest, it's in the community's interest, since it will be using the purl as a vocabulary term. Relying on the author to maintain a purl is not robust - there could even be a conflict of interest. Yes, the owner of a purl can do the right thing and set up additional maintainers or some kind of escrow, but the right thing is currently not the easy thing, and any situation like that is asking for trouble.