[openbiblio-dev] BibJSON and multiple languages

Wed Mar 7 17:35:58 UTC 2012

Etienne Posthumus <etienne.posthumus at okfn.org> wrote:

> An alternative would be to allow either simple strings OR objects.
> (similar to what JSON-LD does in
> http://json-ld.org/spec/latest/json-ld-syntax/#string-internationalization
> as mentioned by Mark previously)
>
> This would push the complexity to the software in the parsing and
> displaying, but allow the end-user to still use as simple as possible
> bibjson.

I strongly support this. I do not think that all values should
be made objects. At an earlier stage  Peter+Mark+I agreed we should support 
duck-typing in bibjson, e.g. a title could be a simple
string, or expanded as an object if needed, and any parser/display unit should
be able to handle that. However, this has not happened: e.g. if the
present BibSoup BibServer is offered BibJSON with the value of a journal field
as a simple string and not an object, it fails to parse the record. I think we 
should commit to supporting simple key-value pairs to the greatest extent 
possible. Only when when this gets stretched should we require users to provide object values as needed.  Another example I would support is
allowing users to provide an author string that is not a list, plus a convention (e.g. bibtex convention, or splitting on ";") for parsing the string.
We do support this for parsing from bibtex, but not currently for parsing from BibJSON.

A related issue is that of a canonical form of BibJSON.
I think having such a form is a good goal. It should simplify deduplication
or records.
So I think we should be supporting a general canonical form with object values, and accept that typically we will be ingesting or processing simpler
forms with mostly just key-value pairs and some indications or conventions
how to map these to objects.

--Jim