JSON has keys that point to values. BibJSON is JSON subject to a loose schema which provides
- conventions about meanings of keys and the structure of their values
- ability to refer to namespaces, for compatibility with RDF and Linked Data
When data is imported to a BibServer, the display is configurable to work on whatever particular keys one desires (at the moment, this is only possible by the person running the instance, but soon will be available as a user config). So these keys and values could be anything that a data provider wants them to be. At this level, the requirement on the BibJSON spec is just to provide clarity to data providers about how to embed their data in JSON for various desired BibServer displays, e.g. to
- create facets for a particular key
- ensure that particular fields are searched over
- transmit internal and external links so they are properly displayed by the BibServer
- ensure that particular text is dispayed in particular parts of a generic BibServer display (e.g. author/title/howpublished and description/abstract elements)
For public sharing of bibliographic data, and for collaborative editing of a bibliographic collection by a community, it is desirable to have shared conventions to structure information according to some agreed community schema. This facilitates a number of processes, especially
- automated deduplication and merging of records,
- aggregation of collections,
- maintenance of community-wide tables of e.g. people, journals, subjects, entries from which which are then available for insertion into any biblio dataset managed by anyone in the community.
There are already numerous standards, specifications, vocabularies, conventions, for sharing data; we are not short of options, nor are we short of complexity. But we lack a consensus standard for sharing of open bibliographic records. It is challenging to tailor a general solution to the various and slightly different use cases that different groups have. But that is what we are attempting to provide with BibJSON.
Sharing of information via the world wide web has been technically possible for decades now. Many efforts at open sharing of bibliographic information have been made, but all to date have been limited in scope, and the various attempts have not been easily aggregatable. We aim to change that, by providing a simple solution that is readily taiored to meet the needs of specific data providers and audiences.
A useful method of sharing data via the web must be simple, useful, and targeted to the needs of a community. Ideally then, it should exist as a convergence of shared and beneficial conventions for sharing the data we care about – bibliographic metadata.
The aim of BibServer is to enable quick and easy presentation of bibliographic metadata via the world wide web; because of this, we use JSON documents, as they are intrinsic to web infrastructure. If is already possible to export these records from a BibServer as BibJSON-ish JSON, which is good enough for me – or for other individuals. But if the goal of our community is discovery and dissemination of information, then we should go one step further and present data that people can easily use in their own systems.
So, here is the challenge for BibJSON – can we converge on a community consensus for representing bibliographic metadata in a simple JSON format? Can we identify where keys from different vocabularies mean the same thing, can we agree on some basic default keys, can we support the use of namespaced keys, and can we make sure people can do useful things with the BibJSON?