OK what happens:

There are schemas which define  the structure of bibliographic records.

There are also what are called RIP files which seems to be config files
containing info about schemas:

e.g. 

    [org.pybliographer/bibtex/0.1]
    
    path: <path-to-schema-file>
    
    readers:
    	Pyblio.Parsers.Semantic.BibTeX.Reader
    
    writers:
    	Pyblio.Parsers.Semantic.BibTeX.Writer
    
    citationkeys:
    	Pyblio.Cite.Keys.AuthorYear

All available RIPs (system-wide and the ones in ~/.pyblio) get parsed at the
start when you do:

     Regitry.parse_default()

After the parsing there a bunch of schemas available in the Registry and you
can do:

Registry.getSchema(<schema-name>)

The actual schema itself seems to be specified in the .sip files found in
Pyblio/RIP/. For example there is bibtex.sip and crossref.sip etc. The exact
requirements for the format of this file aren't entirely clear but the existing
examples are a good place to start.

## What are schemas used for ##

Schemas are then used to define the format of the local pybibliographer
database we use.

Pyblio.Store

# factory for creating dbs with different persistence backends (only file at
# present)

Database
    LocalDatabase
    ExternalDatabase

Record
    Attribute (name, type, value)