A thought that keeps coming to me is that computers are very good at doing database stuff, but Linux does not have a database layer, in the same way that it has a graphics layer.
We end up with every application implementing its own database, or having a dependency on one of the database products such as sqlite, mysql, or postgres. If there was a database provided as a standard part of the stack then apps could rely on that rather than us ending up with multiple databases installed.
Part of the problem is that relational databases are fine for specific applications, but are not much use for the sort of searching that a user wants to do. A user will often want to search across the data from several applications, and it is a bit much for them to have to know what tables and columns each application has used and try to do some sort of join query across them.
A database that emphasizes the relationships between data objects is a much better fit for a user looking for stuff. Technologies like RDF and OWL might have the beginnings of the answer.
We end up with every application implementing its own database, or having a dependency on one of the database products such as sqlite, mysql, or postgres. If there was a database provided as a standard part of the stack then apps could rely on that rather than us ending up with multiple databases installed.
Part of the problem is that relational databases are fine for specific applications, but are not much use for the sort of searching that a user wants to do. A user will often want to search across the data from several applications, and it is a bit much for them to have to know what tables and columns each application has used and try to do some sort of join query across them.
A database that emphasizes the relationships between data objects is a much better fit for a user looking for stuff. Technologies like RDF and OWL might have the beginnings of the answer.