Friday, December 30, 2016

Library Discovery and the Open Access challenge - Take 2

Earlier this year, over at medium , I blogged about the Library Discovery and the Open Access challenge and asked librarians to consider how library discovery should react to the increasing pool of free material due to the inevitable rise of open access.

At the limit when nearly everything is freely available it is possible to consider whether library will have a place in the discovery business. After all, if all researchers have access to the same bulk of journal articles, does it really make sense for each institutional library to provide a separate discovery solution? Even today, many researchers prefer using Google Scholar and other non-institutional discovery solutions that operate at web scale and some (mostly students) grudgingly use our discovery systems to restrict discovery to things they have immediate access to.

This of course is the library discovery will be dead scenario when (almost) everything is free  and not everyone agrees. Some argue, that libraries can add value by providing superior and customized personalized discovery experiences because we know our users better (e.g what courses they taking/teaching, their demographics etc). Then there are plans to leverage linked data etc but I know regretfully little of that.

But the day when open access is dominant is still not here. We live in the world where there is a mix of toll based access and rising but uneven free access, Scihub notwithstanding. I opined that for now "if we really want to stay in the discovery business we need to be able to efficiently and effectively cover the increasing pool of open access resources".

So how does you ensure the library discovery system includes as much discovery of free open access articles as possible?

The idea of a open content discovery matrix by Pascal Calarco, Christine Stohn and John Dove comes to mind.


For most academic libraries who subscribe to commercial discovery indexes (Worldcat Discovery, EBSCO Discovery Service, Summon and Primo, with the later two having merged indexes), there isn't much libraries can do beyond hoping that discovery vendors include such content in their index.

Well I recently came across services like  1science's oaFindr that claims to have a high quality 20 million database of open access papers that perhaps could help? There's also a oafinder+ product that can identify green and gold OA articles for your institution only.




Even if  you can find open access metadata for content that is available for indexing, delivery issues still might occur in index based discovery services as link resolvers are infamously bad at linking to hybrid journals and practically ignore Green Open access articles.


Oadoi.org 

A alternative approach to such "pull" approaches is a push approach. The new Oadoi.org service  (and an earlier service DOAI) is one of the more interesting things to emerge from this year's open access week and it can used together with discovery services.

The idea is simple. One of the challenges of discovery of open access journals in particular Green open access articles archived at subject repositories and institutional repositories is that in general there is no systematic easy way to find them.

 With the Oadoi.org service, you can feed the service a doi and it will attempt to locate a free version of the paper, and this includes both articles made free via the Green or Gold roads.

Here's an example say you land on this article page,  Grandchild care, intergenerational transfers, and grandparents’ labor supply on Springer and you have no access.



Quick as a flash, you grab the DOI 10.1007/s11150-013-9221-x and look it up like this

http://oadoi.org/10.1007/s11150-013-9221-x . And you get autoredirected to the preprint full text on our institutional repository.




Looks like magic! How does it work? The oadoi service uses a variety of means to try to detect if a open access version of an article is available (see below), but it looks to me that the main source for detecting articles on institutional repository in particular is via the aggregator BASE, so make sure your institutional repository is indexed in BASE.

https://oadoi.org/about


My own limited testing with Oadoi was initially pretty disappointing as it failed to find most of the articles hosted on my Institutional repository (hosted on Bepress digital commons). It's possible that the way our institutional repository exposes the doi was not correctly picked up by BASE, but this seems to have been resolved somewhat. More testing required.


Why Oadoi.org?

Savvy readers of this blog might already be screaming, why bother? Just use Google Scholar or plugins like Google Scholar button or Lazy Scholar buttons (which use Google Scholar in background) and all your problems are solved.

It's true that Google Scholar is pretty much unbeatable for finding free articles but the value in OADOI is that it offers a API.

Already many have been quick to use it to provide all kinds of services. For enable Zotero uses it as a lookup engine, librarians have created widgets etc.

But it's greatest value lies in the fact that it can be embedded into discovery services and link resolvers.

Here's work done on SFX doi service and alma libraries like Lincoln University have not been slow to include it either.


These are fairly basic uses of oaidoi and enable users to help direct users to open access content. Still such implementations are usually a "last resort, try it if it works " kind of deal currently and there is no guarantee clicking on the link will work. If you are Exlibris customer on Primo do consider supporting this the feature request "Add oadoi.org as an option in uresolver" which proposes "oadoi.org displays as an option if the API's value of is_free_to_read is true".


To DOI or not to DOI?

A lot of the problems about discovery and delivery of open access content lies in the fact that there are different variants of the same content.

In the old days it was pretty straight forward the only thing that we tracked and access was the article that appears in the journal.

Today, we make accessible a wide variety of content (data, blog posts, conference papers, working papers) and even worse different versions of the same content at different stages of the research lifecycle (preprint/postprint/final published version).

This leads to a great challenge for discovery.

It doesn't help there is a terminology muddle (despite NISO's best efforts at standardising terminology on Journal article Version names and license and access indicators), with people using terms like preprint/postprint/final published versions while others use author submitted manuscripts, author accepted manuscripts and version fo record.

But I think even beyond that, the question I always wonder is , how do we identity/address each version and these days it means assigning dois. The final version of record will have a DOI of course but what about the rest?

As such, I've always been confused about the practice of assigning dois to non peer reviewed papers. For example, should one assign dois to preprints? post-prints? working papers? Should they be a different doi from the final published version? It doesn't help that when you upload items to ResearchGate it offers to create a doi.

I could be wrong, but up to recently I don't think there was a clear guide. But in recent months there seems to be two developments that seemingly clarify this.

First crossref announced they are allowing members to register preprints. The intention here seems to be that the doi of the final version of record is to be a different doi, though there are ways to crosslink both papers, There's even a way to show a relationship between preprints and the later versions as explained in the crossref webinar.

The oadoi service mentioned earlier seems to be pushing in the other direction , encouraging postprints listed in repositories to be added using the same doi as the final version of record to make the postprint findable (but does it mean the preprint isn't since it will have a different doi?). This allows you to find the postprint using the oadoi service as both postprint and final version of record as the same doi.

I'm not quite sure if this is a good idea, while studies show most postprints are not that different then the final published version, it does seem to be a good idea to be able to track the two versions differently. Still mulling over this.


Conclusion

This will probably be my last post for 2016. This year I was particularly inspired towards the end of the year with many ideas but didn't have the time to craft them so expect a flood in the coming year.

I also would like to thank all my loyal readers for following this blog and reading my long winded posts. Next year this blog will be celebrating it's 8th anniversary and my 10th anniversary in the library industry and I might do something special.

Till then, stay happy and healthy and have a great new year's day!




blog comments powered by Disqus

Share this!

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
Related Posts Plugin for WordPress, Blogger...