CMS module for Phprojekt

CMS module for Phprojekt
Tasklist

FS#40 - RDF deep support for better data search and presentation.

Attached to Project: CMS module for Phprojekt
Opened by Mario A. Valdez-Ramirez (mvaldez) - Thursday, 08 July 2004, 12:49 GMT-6
Last edited by Mario A. Valdez-Ramirez (mvaldez) - Thursday, 08 July 2004, 13:09 GMT-6
Task Type New feature
Category Backend / Core
Status Assigned
Assigned To Mario A. Valdez-Ramirez (mvaldez)
Operating System All
Severity Medium
Priority Normal
Reported Version any
Due in Version Undecided
Due Date Undecided
Percent Complete 0%
Votes 0
Private No

Details

This is derived from an answer sent to Bellave Jayaram after a suggestion to support RDF.

Currently, the CMS module uses a very rigid hierarchical tree to store the documents. There are really very few ways for the document authors to add metadata to the document (they can add abstracts, keywords, etc., but nothing flexible).

Hierarchies are easy to understand and users (the ones I've seen on some tests) mostly build documents repositories using some kind of categories list. So far so good for small/closed repositories.

However, problems arose when repositories become big, when the original category list is no longer easy to understand (or some user understand it but others don't), and when they want to search (search is a nightmare with big repositories, as most data returned by the search function is useless for the user). Also, data is presented to the user by date and group access, many users find that the information listed at the main page is mostly useless for them.

For the last year I've been thinking on what to do to:
a) Let the users define what metadata to add to a document.
b) Present the document tree in a flexible way (user defined).
c) Provide more meaningful search returns.
d) Present first relevant documents according to the user's interests.

So, I'm still thinking about how to do all that. Will RDF help me with that? I still don't know enough about it.

I think I need to change several low-level functions before going that way. And I'm still struggling with some more essential problems I hope to solve in the next versions. So, I'm still in planning stage for those extended features... maybe RDF is the answer.
This task depends upon

Comment by Mario A. Valdez-Ramirez (mvaldez) - Thursday, 08 July 2004, 12:50 GMT-6
Pending more study about ontologies, RDF, the semantic web, etc, etc, etc.
Comment by Mario A. Valdez-Ramirez (mvaldez) - Thursday, 08 July 2004, 12:51 GMT-6
Forgot to enter 2 resources given by Bellave Jayaram as examples to get an idea:

http://www.xmlportfolio.com/xmleurope2003/

http://www.callistocms.net


Comment by Mario A. Valdez-Ramirez (mvaldez) - Thursday, 08 July 2004, 13:09 GMT-6
For completeness, I'm including the answer from Bellave Jayaram (partial extract):


> Hierarchies are easy to understand and users (the ones I've seen
> on some tests) mostly build documents repositories using some kind of
> categories list. So far so good for small/closed repositories.

Basically, these are also called taxonomies and can be represented using RDF or OWL.

> However, problems arose when repositories become big, when the
> original category list is no longer easy to understand (or some user
> understand it but others don't), and when they want to search (search
> is a nightmare with big repositories, as most data returned by the
> search function is useless for the user). Also, data is presented to
> the user by date and group access, many users find that the
> information listed at the main page is mostly useless for them.

For this, I am thinking there needs to be "role based" portlets that can customize data per the person's role - for eg., 3rd year medical student with special interest in Pathology, Biochemistry etc. By portlets, I mean a facility to present aggregation of data from different categories.


> For the last year I've been thinking on what to do to:
> a) Let the users define what metadata to add to a document.
> b) Present the document tree in a flexible way (user defined).
> c) Provide more meaningful search returns.
> d) Present first relevant documents according to the user's
> interests.

Yes, RDF will help you to do all of the above - initially one can think of doing a, b and c and d is something that can be added on later once the framework is in place. This is an area of interest to me and I would be willing to help out on how to proceed. Since all this involves XML a lot and I am still coming up to speed on RDF, OWL etc. it might be slow going right now but as more knowledge becomes available about these subjects, the above goals can be realized. I am currently using protege (http://protege.stanford.edu) to learn about all these.

There are some commercial products besides the open source stuff and it takes time to look at each one of them and decide which parts make sense in your given context. However, increasingly I see RDF being used as the backbone for most of these. Once you start understanding RDF and write some applications, it becomes possible to use a cookie cutter approach to develop newer apps based on the same techniques.

Loading...