- From: Juan Sequeda <juanfederico@gmail.com>
- Date: Sun, 18 Jul 2010 11:01:37 -0500
- To: public-rdb2rdf-wg@w3.org
- Message-ID: <AANLkTimEYurKTQ_pEnwZrCWybLnPpU19pNb76sYrZB0C@mail.gmail.com>
Hi Everybody, Hopefully everybody can read the following: http://tech.puredanger.com/2010/07/16/mapping-relational-data-to-the-semantic-web <http://tech.puredanger.com/2010/07/16/mapping-relational-data-to-the-semantic-web>I would like to point out the following from the post: Due to these deficiencies, we have created our own mapping language. It shares some of the ideas present in D2RQ, however it departs significantly from D2RQ in the details. One major difference is that the Revelytix mapping language depends on importing the database schema and building a source ontology that represents the tables, columns, keys etc from that source. The mapping ontology then references that source ontology. One major benefit of this strategy is that many of the ontology references to columns and keys are done by pointing to a resource in the database ontology rather than indicating that reference in text. This reduces the chance of errors in the mapping, makes the mapping process easier in an editor, and most importantly, exposes actual linkages from database to domain. We can then write SPARQL queries against the mapping ontology that provide answers about data lineage. Maybe we should invite them to give a presentation? Juan Sequeda +1-575-SEQ-UEDA www.juansequeda.com
Received on Sunday, 18 July 2010 16:02:10 UTC