display | more...

It's weird that no one has chosen to write anything here; I'd have thought this would be a popular node. Or, perhaps people have written here and everything has been deleted. Either way, I'm game as I know a little about early work in relational databases.

Let's get the boring stuff out of the way quickly. A relational database, typically known as an RDBMS is a way of storing data in tables and relating data in one table to data in other tables by following the relational database model. The de facto standard language for querying and updating the data is SQL, a language based on a very simple relational algebra.

I hate to say when this happened since I'll immediately age myself so I'll make you work just a little. In the summer when Elvis Presley died I was fortunate enough to have a job as an intern, working for IBM in their office in Peterlee. I had the great fortune to be working on the Interactive Planning System (IPS) - the world's first commercially available relational database system.

The wording in the previous paragraph was very carefully chosen. Firstly, IPS wasn't the world's first RDBMS, it was based on a predecessor: the Peterlee Relational Test Vehicle (PRTV). Even PRTV wasn't the first but that's a different issue. IPS wasn't the most successful early RDBMS. IBM really put their weight and marketing power behind the work of the San Jose office and they delivered System R.

I don't remember if we used an actual algebra or a language that was to become SQL. But all the relational databases are based on the simple algebra that Ed Codd described. A side note; don't, for a moment, think that relational algebras are only good for describing databases! There's a whole, deeply technical article just waiting to be written about that subject.

One last idea that was ahead of its time was an extension of IPS known as IPSX. A collection of subroutines had been written to manage geographic data: points, lines, and areas. These subroutines, known as the Spacial Processing Routines (SPR), had been integrated with IPS to for IPSX. That meant that we had the ability to perform relational operations such as joins on geographic data. The most complex relation being Area A does not occupy an area not occupied by Area B. A relation made more complex when an area could have a hole in it!

It was maybe 20 years later when I realized, as the world was catching up, just how far ahead of its time IPSX really was!

Log in or register to write something here or to contact authors.