I recently wrote in my personal blog about how, since joining TopQuadrant, I've grown to appreciate how well SPARQL can serve as a rules language. SPARQL Inferencing Notation, or SPIN, lets you associate rules and constraints expressed in SPARQL with classes of triples. While you don't need to use TopQuadrant products to take advantage of SPIN, they sure make it much easier, especially if you want to use those rules as part of an application. I just finished writing a tutorial (pdf) on how to implement SPIN rules and constraints with your models using TopBraid Composer, and except for one optional detail of the tutorial, it all works with the free edition, so it's available for anyone with a Mac or Windows machine to try.
Using a small collection of data about service contracts and materials purchases, the tutorial walks you through the creation of:
your own functions, written in SPARQL and returning values of whatever type you like
inferencing: the generation of new triples based on other data (in the case of the tutorial, the generation of ISO 8601 yyyy-mm-dd format invoiceDate values from "mm/dd/yy" date values stored in the original data)
constructors: the automatic generation of a postingDate value when a new MaterialsPurchase or ServiceContract instance is created
constraints: setting up the system to alert the user to unpaid materials purchases that are more than 90 days old or unpaid service contracts that are more than 60 days old. Instead of using a lot of redundant code to achieve these two different but similar goals, the tutorial shows how to define a reusable template with a SPARQL query and pass parameters to it (in this case, the numbers 60 or 90) when using the template.
I hope the tutorial demonstrates the potential connections between SPIN technology and real-world business issues to its readers, as well as the ease of implementing it all with TopBraid Composer.
Posts
