In a previous post I express my worries about the fact that the importance of semantics is sometimes neglected. In my opinion this creates messy Linked Data that is useless when it comes to interoperability.
What do I mean by messy Linked Data..? (some detailed examples here!)
There are syntax errors and semantic errors. A syntax error happens for example when you forget to end a Turtle statement with a period. Linked Data tools such as ontology editors and triple stores will prevent you from making these kind of mistakes. However if you choose to write RDF with a text editor and publish it as a file on a web server (which is perfectly fine) , you are on your own.
A semantic error is a mistake that you make while working with the vocabulary that describes the data. When Linked Data is semantically incorrect, it is not fit to be combined with other Linked Data. SPARQL queries do not work, data is put in the wrong places. It is like merging two spreadsheets and not taking care of the column description..
It is very easy to create a semantic error, and the bad news is that it is kind of hard to detect them. Why is this? Because the tools were not designed to be restrictive! This might seem a little strange at first but because of AAA and the fact that we are working in an open world, the tools allow you maximum freedom to create. And they assume that what you create is intended and correct.
The lesson here is that you have to be aware of this, and really, really understand the tools and the vocabularies that you are working with. Yes, this is hard work, but challenging and therefor fun! But if you don’t take the time and effort, you are creating Linked Data that does not make sense to the world, only to you…another silo..