Standards and Pseudo Standards

I participated in some of our company's training and consulting activities this week. One of the interesting issues that came up was the question of standards and how well they are supported by tools. Many participants elaborated on the pain of using SQL. SQL is a database query and interchange language that has been around many many years. However, all vendors have their own dialects and refuse to export their databases into an SQL dialect that other tools would understand without choking.

A similar case came up today when a user attempted to convert UML files into OWL, using TopBraid Composer. Due to changes in the UML specification and its reference implementation for Eclipse, his UML files became slightly incompatible and failed to load. We are still in the process of figuring out how to adjust MagicDraw to create the format the Eclipse needs and vice versa. I find it just incredible how slow the UML standard comes along, given that the OMG has spent so many years on it already.

Enter the Semantic Web: For OWL you have a fine selection of tools that can really use the same input files without choking. Unless in perhaps extreme cases, you can nowadays seamlessly move back and forth between SWOOP, Protege, TopBraid or whatever else you need. This is probably because no large tool vendors are trying to hijack the standard (like in the SQL and UML worlds), but likely also a result of rigorous and efficient standardization processes at the W3C.