# [jifty-devel] Re: Data partitioning (ie multiple databases)

Robert Barta rho at bigpond.net.au
Mon Jan 8 16:36:42 EST 2007

On Sun, Jan 07, 2007 at 05:40:09PM +0000, Edmund von der Burg wrote:
> Has anyone had any thoughts regarding using jifty with several
> datasources.

Edmund,

<delurking>

The Jifty, RoR, ... class of web dev frameworks are all assuming that
underneath they are dealing with data in a relational model.

I here am developing a data integration framework based on Topic Maps
(TM), one of the SW technologies. Accordingly, any application
experiences the data

(a) as a semantic network, organized along the TM paradigm. How the
information is actually stored (relational database, XML files,
OO database, ...) is just a matter of implementation (= speed vs.
flexibility).

If _several_ data sources come into play, then all are experienced
as map and can be merged. "Merging" of maps is an intrinsic
operation in the TM universe to maintain the 'identity' of things.

or, alternatively, with one additional layer

(b) as autovivified objects with classes gleaned from the ontological
information behind the sources.

What I would like to see at some stage is a TM::Jifty where a we
application is resolving backend queries against such a TM engine.

</delurking>

Not sure how this can harmonize with your ideas.

\rho