- From: Franconi Enrico <franconi@inf.unibz.it>
- Date: Mon, 25 Mar 2024 16:12:54 +0000
- To: Gregory Williams <greg@evilfunhouse.com>
- CC: "Lassila, Ora" <ora@amazon.com>, RDF-star Working Group <public-rdf-star-wg@w3.org>
- Message-ID: <9C5B06D6-C907-4B5F-8132-8CCDF269B438@inf.unibz.it>
Greg, you are right. I overlooked the case with literals. My example should still work if the database involves IRIs. WEDDING-CERTIFICATE +-----+---------+----------+------+ | ID | name | location | date | +-----+---------+----------+------+ | :b1 | :enrico | :rome | 1962 | | :b2 | :mary | :rome | 1962 | | :b3 | :enrico | :rome | 1980 | | :b4 | :mary | :rome | 1980 | +-----+---------+----------+------+ Primary Key: ID Alternate Key: name, location, date << :b1 | :enrico :married-in :rome >> :date 1962 . << :b1 | :enrico :married-on 1962 >> :location :rome . << :b1 | :enrico :married-in :rome >> :location :rome . << :b1 | :enrico :married-on 1962 >> :date 1962 . Cheers --e. On 25 Mar 2024, at 16:28, Gregory Williams <greg@evilfunhouse.com> wrote: On Mar 25, 2024, at 4:54 AM, Franconi Enrico <franconi@inf.unibz.it> wrote: As I wrote, :enrico identifies the name of a person, and it does not identify a person. How can this be squared with the assertion :enrico :born-in :rome ? A name is not born in a city. Generated graph-2: << :b1 | :enrico :born-on 1962 >> :location :rome . << :b2 | :mary :born-on 1962 >> :location :rome . This doesn’t look right to me. A birth certificate doesn’t have a location, the *birth* has a location. But the ID here seems to identify the birth certificate (as it is used as the primary key for "BIRTH-CERTIFICATE”?). And there’s an asymmetry in that :born-on is a property of the name, while :location is a property of the reification (in this case, I guess that’s really mean to be a property of the brith certificate?) I am not convinced by this example at all. The modeling seems incomprehensible to me. Beyond the modeling issues, though, I think this is a fundamental problem: Two departments decide to expose this data as LOD, but in different ways. Why should we expect sensible data to result if two different departments use the same *universal* identifiers to model data in different ways? Of course this will lead to bad or confusing data. Thanks, Greg
Received on Monday, 25 March 2024 16:13:07 UTC