|
|
|
which is one of the documents forming the bases of SQL-ish (RDQL/SquishQL/TAP) query languages - it clearly explains what a Data Description Language (DDL) like SQL allows to abstract over triple-stores. It explains also a bit the ideas of semi-structured database vs. traditional RDBMS; it tries to explain why RDBMS are not enough for semi(un)-structured data sets like on the Web, where the schema of the information is not under control of the application and if you use 3rd normal form tables in SQL the result is *very* inefficient and expensive. It mentions the problem of tight coupling of application (e.g. JSP code) with the SQL schema it self, resulting in a lot of maintenance problems and "hard-coding" into the (e.g. Java) code the "logic" of the relation model you actually put on tables. This is very true, and is of the issues that 3rd normal form tries to address (flexibility, changing, parametric/dynamic database). But then you have that kind of orphans problems and dangling references (i.e. removes in object-related databases are "dangerous").
It also tries to go a step further to about interfacing non-RDF system to SQL-ish query (i.e. query legacy data as RDF).
Some info about JDBC (Inkling implementation) and Postgres backend (not "abusing" like we are doing); ODBC things are also mentioned.
Here is a long list of available GIS related open-source software.
[....
"To do that requires a change both in our user interface and also a change in our model for how information is stored on the computer."
But Allchin bristled at the comparison. "Google's a very nice system, but compared to my vision, it's pathetic," he said.
The workings of file systems may be arcane to the average user, but Microsoft plans to use the technology as a selling point. It would emphasize how the system can make businesses more productive by making it easier for workers to find and share documents.
....]
see also XP successor Longhorn goes SQL, P2P
worth reading