status of 1.3

Hi,

Just a short note regarding the codebase of pyblio-1.3. It is regularely
developped, but in a very specific direction (that fits my current needs
as a matter of fact :-)). However, it might soon reach the point where
it will be of general interest.

What does it do as of today:

  - it can manage (ie add, remove, search, sort) structured records.
    These records are composed of attributes that have distinctive
    properties (persons, urls, identifiers, enumerated values, free form
    text,...) The way a record is composed is not defined in the
    framework itself, but lives in a user-provided "schema".

    Each attribute can be qualified by additional informations, so that
    we can have a record that looks like:

      Record ...:
        
	 Title : ....
	 
	 Author: ....
	    Role: Editor
	    
	 Author: ....

	 URL: .....
	    Description: ....

  - these records can be stored in a native XML format, or in a bsddb
    backend.

  - There are already some classes that can handle importing and
    exporting to other "standard" formats, like BibTeX, EndNote (XML),
    RIS, ISI, XML MARC. 

    Important remark: theses importers and exporters are generally not
    concerned about the actual semantic of the data they manage: they
    provide a convenience to build the actual filters that are
    customized for a given schema, your schema, and can process all the
    special cases you have to adapt to with a minimal amount of extra
    code.

  - a simple formatting language makes it easy to specify a
    bibliographic output style.


Design choices:

  - works as a framework: it is quite hard to embbed the current
    pyblio-1.2 code base in some other code. The 1.3 version is so far
    only a framework, on which actual applications will be built (but
    are not, today)

  - not intended as a BibTeX editor: even if the framework should
    support most of the features of bibtex, the common pattern of use
    should be to import a bibtex file in a native database, and only
    export it again when it is needed.

  - extensible, flexible: there is no attempt to fit to a specific
    standard (BibTeX, MARC, MODS, Dublin Core,...) but only to give
    enough power to make it easy to adapt to the one _you_ need, in a
    pragmatic approach. This flexibility intentionally uses python as
    its substrate rather than some XML or XSL format, as I think it is
    really more efficient for the user's brain.


What's coming next:

  - for the data model: there are only a few things that I plan to add
    in the near future, but with interesting impacts, namely links
    between records (giving birth to a non-flat structure of records),
    and references in attributes (allowing authority records, or simply
    shortcuts like BibTeX's @string command)

  - for the ease of management, a system to identify more explictely the
    different schemas, and the importers that support them

  - code to support querying external databases as easily as importing
    file formats. 

I guess I should not have called that a short note... :-)

Frédéric

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFC1tVaFjQHpltE9KURAjN+AKClSg9NRqeLJIRuqIbmqLQwhpStoACgmdE3
aORAENclv5IvjRDPkhiRp5U=
=y3OT
-----END PGP SIGNATURE-----