Is it advisable to establish a DOI number for rare objects that won't change hands?

If a library has a collection of rare, non-book items, is it advisable to establish a permanent digital record for these?

If the items are unlikely to change hands, does that influence the decision?

jonsca

• digitization

Comments

• dsalo: Are these items in fact digital? If not, then a DOI makes no sense.
• jonsca: @dsalo That's what I thought, too, but from doi.org "DOI is an acronym for "digital object identifier", meaning a "digital identifier of an object". A DOI name is an identifier (not a location) of an entity on digital networks. It provides a system for persistent and actionable identification and interoperable exchange of managed information on digital networks. **A DOI name can be assigned to any entity** — physical, digital or abstract — primarily forring with an interested user community or managing as intellectual property."
• jonsca: Maybe it's a misreading on my part
• dsalo: No, you're right -- but I still don't see what assigning a DOI to an analog object achieves. What are you thinking it will help you (or anyone else) do?
• jonsca: @dsalo I am by no means steeped in this field at all, so I could be way off of the mark, I defer to your expertise
• jonsca: @dsalo For tracking and permanence of the record
• ksclarke: I could see the use if you are using it as an identifier within a system that tracks your objects (digital and physical). Having a consistent identifier scheme across digital and physical objects might be a benefit.
• Chris Adams: I wanted to second @ksclarke's comment: it's incredibly useful to have a single answer for how to identify things, no matter what they happen to be: programmers love to have reliable primary keys and most programs reflect that.

Answer by Ed Summers

Personally, I think it's more practical to focus effort on managing your URL namespaces so that your URLs can persistently identify your digital resources. Tim Berners-Lee discusses the importance of this in his Cool URIs Don't Change. The advantage here is that people bookmark URLs, and will create links to your content with them. It is in the best interests of your institution that these URLs are stable, and can persist across content management systems.

Fortunately there are lots of software options in the CMS and web framework space that let you manage your URL namespaces. HTTP and DNS software is widely deployed, and has lots of eyes on it, whereas DOI and the Handle System which it relies on are niche technologies that have only a handful of implementations.

My advice is to delay implementing anything like DOIs or Handles and get your institution to instead have a plan for how it manages its URL namespaces.

Comments

Answer by ksclarke

Another option might be to use ARKs (Archival Resource Keys).

The first page of that link explicitly says they are for physical objects, too, and cites other advantages (e.g., you can mint your own ARKs, once you've been assigned a Name Assigning Authority Number (NAAN), for free (whereas DOIs cost)).

I do tend to think of DOIs as being for digital objects (despite the note in the comments above that they can be used for any entity) and I also tend to think of DOIs as being used for things which are published (the DOI becomes a part of the citation). A physical object that is owned and not transferred might not be considered a published item.

The California Digital Library has a brief ARK vs. DOI comparison (comparing DOIs that are minted through DataCite).

And, use of an ARK doesn't have to run counter to the advice Ed gave above. You can plan for patterned URLs that incorporate the archival resource key, like UNT has done (for example):

http://digital.library.unt.edu/ark:/67531/metapth61018/

Comments

• Ed Summers: ARKs are a nice halfstep between DOIs/Handles and managing your URL namespace. My only beef w/ them is that the pattern for putting an ARK in a URL uses the colon, which is a reserved character that needs to be URL encoded. While this isn't a major problem since most browsers will do the encoding for you, it can lead to subtle problems sometimes when working with them programatically. At least they're not OpenURLs :-)
• ksclarke: True, if you are using patterns that extend beyond the ARK sometimes you have to URL encode the ARK so that it stands alone/together within the URL. For example, an image server that uses the IIIF Image API (draft) might have URLs that look like: `http://images.org/service/ark%3A%2F38305%2Fg4t153vf%2Fis%2F1/80,15,60,75/full/0/native.jpg`

Answer by Jenn Riley

I can add a bit of rationale for why formal, web-style unique identifiers of whatever scheme might be useful for analog materials, especially rare ones. It boils down to citation and linking.

Scholars and enthusiasts use these materials presumably, and they create new works (written or otherwise) on the basis of them. In an increasingly web-integrated and machine-readable world, the traditional citation could be enhanced through the use of these unique identifiers. They would point directly and unambiguously to your resource that was cited or otherwise used in the new work, and provide a machine-readable mechanism for pulling together various different uses of the same material. Initiatives like the Open Annotation Collaboration (http://www.openannotation.org/) would strongly benefit from such an approach when used as a part of digital annotations or other secondary works, even if your material that's being cited or used is not (yet?) digital.

That's not to say this is easy - identifier management is indeed quite hard. I agree with Ed that a coherent namespace policy is the key, no matter which particular identifier scheme is implemented. Know clearly what it is you're identifying (oops, there's the FRBR rabbit hole) and know how you're going to assign and maintain those identifiers. You'd be way ahead of the curve in doing this, but it's a direction I see the community going in strongly for the next while.

Comments

Answer by Michael Hopwood

I'm looking into this question and similar questions in the museum space right now.

Here are some of the best clues I've come up with so far:

http://museumid.net/documentation

http://www.cidoc-crm.org/URIs_and_Linked_Open_Data.html

http://ispiders.blogspot.co.uk/2007/10/buying-selling-doisand-same-for.html

http://ispiders.blogspot.co.uk/2007/10/biodiversity-informatics-needs-business.html

http://www.crossref.org/CrossTech/2011/04/content_negotiation_for_crossr.html

I wouldn't call Handle and DOI exactly niche technologies. Yes, there are a limited number of implementations - but look how big those implementations are, and what kind of (meta)data they are being used to exchange.

Rare books are an edge case between library and museum best practice. There's actually a data model for that: FRBRoo - see e.g. http://www.cidoc-crm.org/frbr_papers.html

Although as you rightly note, rare books are (hopefully) unlikely to change hands too much, there are other reasons you might want to register them separately and maintain the data somewhere centrally - principally to do with mobility of the (meta)data rather than the actual source item - and of course, how about persistence of any digitised versions there may be?

It would be very interesting to see what can be done with linked data and DOI in future now that this has been kick-started somewhat.

Comments