OA-RJ project blog

What is the Junction?

The Open Access Repository Junction breaks down into two main parts: a Discovery tool and a Delivery tool. The Junction is the Discovery tool.

The Junction will work in two modes:

1. For a known locus, it will return a list of repositories that may be suitable for deposit,
2. Given an item, it will deduce a set of probably repositories for deposit.

How does it work?

The Junction is, basically, a deduction tool: it uses an internal database of organisations and repositories to create a list based on criteria given to it.

At the most basic level, there is a simple API that a client calls, which returns a list of repositories. This call can be augmented by defining the location of the user, and be restricting the list of returns to a repository type [Institutional, Learning, Data, etc] and/or content type [Articles, Thesis, Data, Learning Objects, etc]

There are also some supporting APIs, which will list all the records of a type and the repositories associated with them (see the “Repository Junction List APIs” blog post), and that provide an AJAX auto-completer return for clients to use (see the “Repository Junction AJAX-able APIs” blog post)

In a more complex role, the Junction will take a deposit object (metadata, metadata & binary file(s), or hopefully even just the binary file(s)), and extract location information from the object, which it will then deduce a list of potential targets.

Documentation

Documentation for the api is at https://oarepojunction.wordpress.com/junction/junction-api/ and the beta release for the API is at http://devel.edina.ac.uk:1201/api

The future

For the life of this project, the Junction will remain fairly simplistic: find the organisation(s) and therefore the appropriate repositories

In the longer term, there are a number of ways that the Junction could improve, for example by:

• taking the [by GPS?] location of the user, and finding all the repositories within a defined radius of that location.
• compiling a list of repositories, and rank them based on various criteria that match
  • …. which would need to include in the returned XML
    • Is associated with the same organisation
    • Is within 2km of your location
    • Is of the repository type you are after
    • Accepts the data type you are depositing
    • Is managed by the same manager
  • …. and if things become desperate
    • Is within 20K, or 50K, or 100K
    • Is in the same country
    • Speaks the same language
    • ….
• text-mining to find all the authors
  • …. and an Authors Disambiguation service to find the appropriate organisation
• text-mining and subject classification to find appropriate subject repositories
• using funding code records to find appropriate funders repositories
• using JULIET to determine the funders policy
• using ROMEO and actively chose to exclude file objects from the records being passed onto repositories, or modify the metadata to comply with embargo conditions