I really enjoyed Steve Yegge’s long post last week about the shortcomings of Google’s architecture. Google provides great services that I use every day but building systems as Amazon does of composable web services (AWS) to build more complex products and services seems like a better approach.
I have been experimenting with Semantic Web (SW) technologies since reading Tim Berners-Lee, James Hendler, and Ora Lassila’s 2001 Scientific American article. I have not often had customer interest in using Semantic Web technologies and I think that I am starting to understand why people miss the value-add:
Just as AWS provides composable web services SW helps information providers to provide structured and semantically meaningful data to customers and users who decide what information to fetch, as they need it. These consumers of SW data sources must have a much higher skill set to build automated systems compared to a user of the web who manually navigates around the web to find information that they need.
So I think that the issue becomes how can to make it relatively easy for system designers and software engineers to fetch and consume information. The easy answer is to point them to a good book on SPARQL and RDF data sources. A better answer is probably to provide examples using common programming languages, the “best” libraires for making SPARQL queries, and small sample applications tailored to the types of data that an information provider provides and what type of inferencing makes sense to discover implicit data that is not explicitly in the provider’s data store.
I would describe the SW as building and using composable data sources that are defined in terms of ontology’s that make it possible to merge data from different sources and to discover implicit data through inference/reasoning.