[Zope-dev] Re: [Zope] Multi tier partitioning meaningful in some Zope apps ?

Jerry.Spicklemire@IFLYATA.COM Jerry.Spicklemire@IFLYATA.COM
Fri, 10 Dec 1999 18:26:09 -0500


Greetings Zope Gurus,

I am immensely intrigued by the possibility of leveraging the innate Object
/ Relational features in PostgreSQL as an alternate storage method for Zope
Objects, especially Code Objects. In PostgreSQL, new data types may be
defined, with inheritence relationships such that base properties may be
derived from parent data types. The classic University Postgres example is
to start with goemetric primitives (node, cell, etc.), and build up to
multi-dimensional object data types. One might then expand to a map data
type, and from here  define the boundaries of nations, super-imposed on the
globe. Think about defining an element data type, then an amino acid
molecule, which could be used to describe a gene, and then a chromosome. A
machine part data type, and then a space shuttle. You get the idea. 

Why not Zope Object data types, for DTML Methods, DTML Documents, External
Methods, etc. If there's a pressing need to further partition content and
code, this might be one very clear way to achieve that end. Just pondering
the range of new data types for Zope makes me dizzy! Imagine reverse-sorting
a PostgreSQL Table on the Origination Date / Time Property of DTML Methods
to step back to the version right before you broke an application . . .

Later,
Jerry S.