I was told the other day by an ontologist acquaintance of mine that they only use Protege or Top Braid Composer for their modeling work, rather than writing things out by hand. It’s a not uncommon reaction, and it’s one that I’ve seen used by many data modelers. I’m also not complaining about either of those tools — they are very good for ensuring the integrity of models once you have things worked out.
However, the the thing that I think so many people miss is not that I believe that writing Turtle files by hand is a superior way to build a schematic structure . Instead, it’s that when modeling in general, jumping to the tools in your suite (whether UML based or OWL based or XSD based) should only be done after you’ve sat down with pencil and paper (or the text file equivalent) and worked out exemplars by hand.
What’s an exemplar? The term, as you might expect, is a recognized variant of example, and indicates a specific fragment of code in the language in question that illustrates a particular scenario. It is, in other words, a code use case.
Exemplers in general should be where the vast majority of modeling is done, because each of the exemplers in question is what your tools should ultimately produce. It provides a way for you to test assumptions and to identify patterns, and it is often far more useful in…
