[Zope-PTK] Topics support in PTK

Kent Polk kent@goathill.org
17 Feb 2001 17:26:29 GMT


On 12 Feb 2001 23:30:01 -0600, Tres Seaver wrote:
>
>Topics will be "addable content", like other content objects,
>with their own metadata.  They will, in effect, function as
>"canned catalog queries", and will offer a variety of
>presentations, as time and resources allow:
>
> * A folderish view (matching objects will seem to be "in" the
>   topic);
...
> * Perhaps tree-like views for nested topics;

This is essentially what we did with the PDMS which DC and SwRI
created about 4 years ago, except that we used sql to store the
document metadata, including description, keywords, etc. as well
as workflow states. A folderish heirarchy was held in the zodb for
object management, etc. and later on, we added sql queries which
built tree views of a query, the default was to arrange all documents
in a heirarchy specified by the query. Basically similar to the
old Principia Tabula Product, which was a very powerful tool (why
did that disappear? Seems like the guts of Tabula would make an
excellent start on a tool for data mining...)

One thing that I wanted to do, but never got around to was to
combine the tree layout with the query results such that the
documents matching the the query were simply marked and the ones
that didn't were not marked. This would make it easier to browse
related documents (by heirarchy) that didn't quite match the query.
One might way that it would be better to simply improve the query,
but in practice, I found that few people were really able to do
this efficiently. I set up some hard-coded examples and they went
over very well. Like querying a library catalog and having all the
books that match be illuminated for you in the stacks.

Anyway, IMO, this will be a very important addition to the CMS.