Dear all,
In a recent conversation, a colleague (and a friend ) asked about permalinks for identifying OER. I never thought about this in-depth: I just considered having a DOI ( Digital Object Identifier) assigned and did not even look for alternatives; my bad. So I started looking into it a little but havenāt found much so far.
What I found so far is:
there are multiple registration agencies that assign DOI (you can find a list here) and these agencies manage communities in different areas: entertainment, standards, the built environment, natural history collections, scholarly communications and research data.
My colleague mentioned also ARK (Archival Resource Key, persistent identifier for information objects of any type) so I started looking into it. Then, I landed on this glossary that ādescribes some commonly seen object identifiers.ā I honestly never encountered most of them!
Long story short, the question remains: permalinks for identifying OER: is there one that is better than others?
What are the pros and cons of different solutions for different kinds of OERs (textbooks with an ISBN, exams, software with SWHIDā¦)?
If anyone knows more and is willing to share, Iād be happy to learn!
Thanks in advance
My experience with registry-based addressing systems such as DOI is that the registrations frequently lapse, ie., the document becomes unavailable, the document is delisted, or an access barrier is created. For these reasons my preference is to employ a distributed content-based addressing system, such as described here: https://www.downes.ca/presentation/515 This wonāt solve your needs in the short term, since no such widely used system exists, but this is the direction I believe OER addressing should and will move toward in the future.
This āseems likeā something that should have a solution, but gazing down the glossary you shared, @paola does make the head swim. One would have thought the magic of blockchain could have handled this (Alan ducks).
Iād agree with @Downes that central registries (as there are several listed in the DOI family) are potentially problematic and as @edtechdev points out seem to need affiliation with an organization involved withe registry service.
Also, there seems to be a consideration of what kind of open resource we are talking about. Books and research papers/articles that have single points of origin (original publisher) work well with registration number style references, especially as the identifier usually travels with them (included in the source). But if we are talking about other formats - websites, media, docs, identifiers are typically not āattachedā.
Note that the National Teaching Repository (based in the UK but they do take submissions from beyond) do generate DOIs for anyone for all types of items submitted to their site, though my observation there is that the DOI is attached to the entry in the repository, not the item itself.
Yesterday, i just noticed that a local openly licensed early reading project called Vula Bula was no longer available. This was the first graded reading programme in African languages where progression from level to level is based on the phonics of each language. But because of governance issue, the project went into liquidation. And a wealth of important early reading resources in vernacular languages have been lost. Permanant identifiers are an extremely important issue for the OER community to address,
from what I can see after a quick search, we see rather datacite-oriented practices for version scalability, and the use of other permaliens (ARK, isbn for textbooks) and specific PIDs (e.g. SWHID) for software.
I havenāt looked in detail at national strategies except Australia, the Netherlands and Canada but RDA has an interest group on the issue. I donāt see anything on OER!
A question that we at LibreTexts have wrestled with is are DOIs and ISBNs appropriate for OER? Our conclusion is no. For ISBNs according to ISBN international you need another isbn for
āA revised edition of an existing book
A new language edition of an existing book
A change of format (paperback/hardback)
Different formats of digital publications (e.g. pdf, html) also need separate ISBNs.
A change of title
A change of publisher
A change in usage constraints including open copyrights, text to speech functions and any changes in the DRMā
DOIs are pretty similar. Both were designed for publish once formats, not something like OER which undergoes constant (hopefully improving) changes and can be found on multiple websites (you would need a new ISBN for each)
As far as I understand, through Zenodo, you can also have a DOI for the concept, which refers to all versions at once: FAQ versioning. It is ājustā another DOI, but you can keep a DOI for each version you upload AND also have one āConcept DOIā, as they call it, which refers to all versions together. When reusing and adapting, you can use the DOI of the specific version you are referring to, but if you want to refer to the evolution of a resource, you can use the Concept DOI.
I know that itās been a few weeks since you posted this request.
My comments are similar to Downes initial comment earlier regarding distributed content-based systems, but adds some color regarding how multiple systems could be combined to solve a number of thorny issues in the OER field todayā¦
I would suggest instead of a link, we should be seeking to build graphs of OER content, not links to single content objects. Then within each graph, you could have your specific commit ID that relates to the desired version / state of this evolving content. Fundamentally, this is how GIT repositories work in the software development world.
My logic here is that because content evolves, it is important not just to identify the object, but also the objects relationships (including various licensing requirements for which it is a derivative). Software developers have been working through these challenges now for some time, not just for evolving code (as in the GIT example above), but also to create the āBill of Materialā (BOM) for dependencies and license verification (linked graphs). Although this conversation is early for OER, there is some interesting work and technology that is starting to evolve in this regard in the general content / Web3 space. You can take a look at Fluree as one such example:
Itās not obvious on the site without some digging, but the idea being proposed regarding a āCircular Economy of Dataā is powered by Graph Database technology. Although I appreciate the simplicity that a single DOI provides, Iām not sure that having a single massive repository / naming convention / id for all digital objects will make sense for much longer given the growth in content volume and evolving needs to track and verify upstream licensing rights.
Two EU initiatives related to the Data Governance Act (DGA) may pave the way as to how this could work from an accessibility perspecitve:
I am sincerely grateful to each one of you for your insights and the resources you shared. I need some time to dig into them and know more, happy to explore the topic further.
) asked about permalinks for identifying OER. I never thought about this in-depth: I just considered having a DOI ( Digital Object Identifier) assigned and did not even look for alternatives; my bad. So I started looking into it a little but havenāt found much so far.