The viability of PURLs as PIDs

Hi,

Like many organisations, the Dublin Core Metadata Initiative (DCMI) relies on PURLs as persistent identifiers for the namespaces used in our vocabularies.

The DCMI Metadata Terms constitutes one of the world’s most widely used vocabularies, and the PURL used to identify its namespace - http://purl.org/dc/terms/ - appears millions of metadata records and RDF triples.

Originally developed by OCLC, the PURL system is now hosted and managed by the Internet Archive, and we are grateful to them for taking responsibility for this.

However, for some time DCMI has been experiencing problems with the management of its PURLs, to the extent that questions are now being asked by members of the wider metadata community about the stability and viability of this system. We are also aware of other organisations experiencing similar problems.

If the PURL system were to fail - either catastrophically, or even gradually with organisations unable to update the PURLs they have minted - then the consequences for a huge number of metadata and linked data systems could be disastrous…

I have two questions for this forum:

  1. Does anybody else have concerns about the PURL system? I mean operational concerns, rather than philosophical:wink:
  2. Would anyone be interested in an ad hoc discussion about this at the PIDaPalooza event in Lisbon later this month?

I will also write directly to the PURL system team, to see if they might join such a discussion.

Thanks for reading this far :slight_smile:

Paul

1 Like

Hi Paul,

Yes, I’d be interested in that discussion. Wondering if this post would have been better in the PIDapalooza thread?

FYI - my organisation, the Australian Research Data Commons (formerly ANDS, joined with two others now), issues PURLs for research grants from our two major funders ARC and NHMRC. You can see the grants discovery portal here. The PURLs are used by our research repository community in particular and enable linking of papers and data with grants. We’d like our funders to move to DOIs for grants but for the moment (and for some time now) we have the PURLs instead. Yes, I believe there have been some challenges, for example, getting PURLs updated during the longish handover period from OCLC to IA. Not sure about currently though I’ll try to find out before we meet.

I’ll be at PIDapalooza and it’d be nice to catch up with you in general (you may recall chats we had a lifetime ago about RIOXX :slight_smile:

Best,
Natasha
Australian Research Data Commons

Hi Natasha,
good to hear from you!

I didn’t spot that there was a PIDaPalooza thread - I will put a short message there too.

Good that you think such a discussion will be worthwhile.

See you in Lisbon!

Paul

1 Like

Hi,

I’ll be at PIDapalooza and I’d be interested in joining a discussion on PURLs.

I’m curious to discuss how PURL-focused initiatives would compare to 301Works.org and the work of the URLTeam on short URLs.

Luc

Sadly, I will not be able to join the discussion in Lisbon. I’m however very interested in the outcomes. My institution, ZBW - Leibniz Information Centre for Economics, is using PURLs for access to the folders of our “20th Century Press Archives”. Currently, the content is migrated to another server; the current domain is already outdated.

Since a change of the email address, we have lost access to the PURLs formerly owned by the account, and have no way to change them or add new ones any more.

We have described the issue and asked for help in emails to the only support address we could find (info@archive.org) for four times from October to December 2019, without any response at all.

So yes, we are very concerned about the operability and viability of the purl.org system, and I very much hope that the issue can be discussed with somebody involved from archive.org at PIDapalooza.

Cheers, Joachim

Hi,
Unfortunately I was unable to attend PIDapalooza this year due to a last minute scheduling clash. I’d be very interested to hear any outcomes from this discussion.

cheers

Nick