Re: Ontology on PostgreSQL - is there something?

From: "Mike Rylander" <mrylander(at)gmail(dot)com>
To: "Rodrigo Hjort" <rodrigo(dot)hjort(at)gmail(dot)com>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Ontology on PostgreSQL - is there something?
Date: 2006-11-19 01:37:44
Message-ID: b918cf3d0611181737y20ca8919sb0fb382eb30482ee@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 11/18/06, Rodrigo Hjort <rodrigo(dot)hjort(at)gmail(dot)com> wrote:
> Dear PG-hackers,
>
> Based on the paper below [1], I ask: is there anyone working on, or already
> tried to such native implementation on PostgreSQL? I didn't find anything
> related on pgFoundry. There is also a presentation [2] related to the paper.
>
> "By Souripriya Das, Eugene Inseok Chong, George Eadon, Jagannathan
> Srinivasan, Proceedings of the 30th VLDB Conference, Toronto, Canada.
> Ontologies are increasingly being used to build applications that utilize
> domain-specific knowledge. This paper addresses the problem of supporting
> ontology-based semantic matching in RDBMS. The approach enables users to
> reference ontology data directly from SQL using the semantic match
> operators, thereby opening up possibilities of combining with other
> operations such as joins as well as making the ontology-driven applications
> easy to develop and efficient. In contrast, other approaches use RDBMS only
> for storage of ontologies and querying of ontology data is typically done
> via APIs. This paper presents the ontology-related functionality including
> inferencing, discusses how it is implemented on top of Oracle RDBMS, and
> illustrates the usage with several database applications."
>
> [1]
> http://www.cs.toronto.edu/vldb04/protected/eProceedings/contents/pdf/IND3P1.PDF
> [2] http://www-scf.usc.edu/~csci586/ppt-2005/dguo.pps
>
> Besides, what are your opinions on the subject?
>
> Is it worthful to have such native implementation on PG, even as a modest
> contrib module?
> Based on your development know-how on the backend internals, how difficult
> is it to develop this (3 operators + 1 index type)?

For simple situations of BT/NT ontology relationships, check out the
ltree contrib module*. It allows building and indexing of directed
graphs and other network-like structures, and seems ideal for both
path storage and search.

[thinks some more]

If applied as a column per axis, ltree may be enough ... interesting
stuff in any case.

* http://www.sai.msu.su/~megera/postgres/gist/ltree/

>
> --
> Best regards,
>
> Rodrigo Hjort
> http://icewall.org/~hjort
> CELEPAR - Cia de Informática do Paraná - Brasil
> http://www.pr.gov.br
>
>

--
Mike Rylander
mrylander(at)gmail(dot)com
GPLS -- PINES Development
Database Developer
http://open-ils.org

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Neil Conway 2006-11-19 02:44:55 Transaction start in pg_stat_activity
Previous Message Russell Smith 2006-11-19 01:18:29 Re: Nasty VACUUM/bgwriter/segmentation bug