 |
 |
On Jan 5, 2015, at 1:35 PM, Karen Coyle <[log in to unmask]> wrote: > 1) Everyone should read at least the first chapters of the Allemang book, Semantic Web for the Working Ontologist: > http://www.worldcat.org/title/semantic-web-for-the-working-ontologist-effective-modeling-in-rdfs-and-owl/oclc/73393667 +2 because it is a very good book > 2) Everyone should understand the RDF meaning of classes, properties, domain and range before beginning. (cf: http://kcoyle.blogspot.com/2014/11/classes-in-rdf.html) +1 for knowing the distinctions between these things, yes > 3) Don't lean too heavily on Protege. Protege is very OWL-oriented and can lead one far astray. It's easy to click on check boxes without knowing what they really mean. Do as much development as you can without using Protege, and do your development in RDFS not OWL. Later you can use Protege to check your work, or to complete the code. +1 but at the same time workshops are good places to see how things get done in a limited period of time. > 4) Develop in ntriples or turtle but NOT rdf/xml. RDF differs from XML in some fundamental ways that are not obvious, and developing in rdf/xml masks these differences and often leads to the development of not very good ontologies. +1 & -1 because each of the RDF serializations have its own advantages and disadvantages — Eric Morgan