This session starting with Dr Sarah Bridle, a physicist from UCL. Sarah saying because of arXiv in her subject area the library could happily cancel all relevant journal titles from her point of view, since they only serve to ‘badge’ the papers, and she doesn’t need the subscription for this to exist.
However, clearly some issues with the functions journal server (peer-review) and the need for a business model, or process, which enables these functions.
Number of things lead to the idea of ‘overlay journal’ – very little copy editing seems to happen in published version, and often leads to confusion (e.g. with page numbers) between arXiv version and published version. So, they started talking to the library about the idea of running an ‘overlay journal’
Now Dr Panayiota Polydoratou from UCL library relating work that they undertook in partnership with other institutions (funded by JISC) to look at the issues – i.e. the RIOJA project.
The aims of RIOJA were:
- Build the RIOJA toolkit
- APIs etc.
- Sustainability
- estimate running costs for arXiv overlay journal
- …
[Panayiota going very very fast – can’t get all of this – I guess it will be on the RIOJA website somewhere…)
Started by surveying 4000+ researchers (got 683 responses), and interviewing editorial boards and publishers. In general the latter two categories interested in looking at new models (which isn’t so suprising I guess – the current model, like music etc. is clearly not going to work well in the internet age)
Interestingly when researchers were asked about what they thought ought to be prioritised in terms of payment, paying referees was very low on the agenda.
More work needs to be done on exploring sustainability issues, business model and potential implementations.
Not very clear exactly what RIOJA has done except the survey – but now Antony Lewis from the Institute of Astronomy, Cambridge is going to talk about the APIs and show demos so perhaps that will make it clearer…
OK – so technical objectives were:
- Develop open API for communication between repositories and journals
- Develop software for hosting overlaid journals using the API
- Demonstrate journal s/w using API implemented on arXiv.org repository
- Develop version of ePrints repository s/w to make complete open source package for any subject area
The RIOJA APIs support:
- Paper metadata communication
- Author authentication
- Integrated submission
- Publication status
I’m worried there has been no mention of SWORD or OAI-PMH yet – which would seem to cover some of these areas?
There were specific issues:
- Paper version tracking – only a specific version of the paper on the repository should be ‘published’
- Science specific issues e.g. handling equations
- Simplifying workflows, dealing with ‘continuous publication’
Journal software used was ‘Open Journal Systems’. The submission was done ‘via repository ID’ – presumably meaning this is what was used as the identifier (surely would have been better to use a DOI or other independent identifier?)
Antony now showing some screenshots etc. Submitter asked to assign keywords as part of ‘submission’ process.
Conclusions:
- Software and API infrastructure now mostly in place
- academics who want to run journals covering costs themselves, they can do it using this
- Make any number of journals based on any nyumber of repositories in any subject areas
- Aim to have suite of Open Source s/w for easily setting up repositories
- Some work still to do
- Support for metada with equaicty latex display, referee report publciation, options for ways of handling copyediting
- Then just need some good editors and a small amount of money
Test site: http://arxivjournal.org
Source code and information: http://arxivjournal.org/rioja
API specification: http://cosmologist.info/xml/APIs.html
A number of questions:
- Are we not rushing ahead here where we haven’t sorted some basic problems around citation (e.g. things cited in arXiv before publication)
- Mention of ‘Storelink’ as a project looking at some of these areas
- I asked about overlap with SWORD and OAI-PMH and the Rioja APIs – not convinced by the answer to be honest – but my main concern is lack of overlap here.
- Some defense of ‘publishers’ and the roles they play from the floor: legal issues, communication, marketing, specialist expertise etc. Bottom line – don’t think you can do high quality publication without publishers, and if you get rid of publishers from the system, you would have to reinvent them in some form.
In response to the questions:
* arXiv identifiers are what scientists use to refer to papers. The main driver of this project is to provide something scientists want and can use easily. (also I think arXiv does not publish DOIs for non-published work). And yes, the idea is to rush ahead and actually do something! Note that an overlay journal solves all problems to do with pre-publication citation, since their is a unique mapping between overlay journal IDs and repository IDs: citation agregators can just treat them as synonyms for most scientific purposes.
* There are various reasons we didn’t use OAI-PMH: 1. we require immediate API response for interactive submissions, rather than allowing for deferred poling implemented in OAI for harvesting; 2. we need access to metadata before it is available publicly on OAI (to allow for integrated submission); 3. We need detailed paper version information; 4. We need a number of other custom functions. As far as possible we have tried to keep the API data objects similar to popular protocols, e.g. internal abstract formatting is basically the same as CrossRef (+latex).
* From my point of view data linking issues (like Storelink) are related to the repository end, and are fairly independent of the quality stamping process.
* SWORD is aimed at automating deposits of papers in repositories, which is logically the reverse of what we are trying to do (though indeed there may be a useful link if the overlay journal does copyediting and wants to automatically update the version on the repository; i.e. I think it is complementary API that could be used rather than overlapping with RIOJA).
* Legal issues, dissemination, etc are the problem of the repository in the overlay model, since that is where all the content resides. For arXiv overlays at least marketing is redundant since everyone reads papers on arXiv.
But if I’ve misunderstood something important about the various other projects please do let me know!
Talk slides should be up on the RIOJA site this evening.
Antony