This page lists common questions and answers regarding SKOS aims and application. Your contribution is much appreciated here.
Feel free to add your questions (and answers!) directly below, or send suggestions to firstname.lastname@example.org, but please don't delete anything written by somebody else!
Also, this FAQ should cover a relatively broad audience, that may know a lot about Knowledge Organization Systems but not much about the semantic web, or vice versa, so bear this in mind when suggesting questions and writing answers.
- 1 Q: Where can I find the RDF/OWL files for the SKOS and SKOS-XL ontologies?
- 2 Q: What exactly is a Knowledge Organization System, and what is it good for?
- 3 Q: How do I publish a KOS on the semantic web?
- 4 Q: What's the difference between 'concept-oriented' and 'term-oriented' models of thesaurus structure?
- 5 Q: Can I use SKOS to publish a glossary or other type of knowledge organization system?
- 6 Q: Can I build a new knowledge organization system out of bits of someone else's?
- 7 Q: Can I use SKOS properties for other purposes?
Q: Where can I find the RDF/OWL files for the SKOS and SKOS-XL ontologies?
- See the RDF vocabularies section of the W3C SKOS site. The RDF files for SKOS are also accessible using content negociation from their respective namespaces, for instance http://www.w3.org/2004/02/skos/core#. Please note that a normal HTML browser will get an HTML page when accessing these namespaces!
Q: What exactly is a Knowledge Organization System, and what is it good for?
- A Knowledge Organization System (KOS) is a set of elements, often structured and controlled, which can be used for describing (indexing) objects, browsing collections, etc (see SKOS cases). Typical examples of KOS are thesauri, classification schemes, subject heading lists, taxonomies... KOSs are common in cultural heritage institutions (libraries, museums) or in any scientific discipline, such as biology, which has a specific interest in naming and classifying.
Q: How do I publish a KOS on the semantic web?
- [Adapted from the ESW wiki] To publish a KOS such as thesaurus on the semantic web, follow these steps: (1) generate an RDF description of the thesaurus' content (2) publish the RDF data on the web.
- The first step means creating a file or set of files that is an RDF-based serialization of the thesaurus itself. The SKOS RDF schema provides most, if not all of the RDF classes and properties you will need for this task. The SKOS Primer is the best place to start to learn more about this schema. For more in-depth guide to generating RDF-based serializations of existing thesauri, for both standard and non-standard thesauri, and from a number of existing formats (including XML, relational tables), please see the tutorials, presentations and papers section of the W3C SKOS site, or the documentation page of this wiki. Re-using an existing schema such as SKOS, rather than designing your own from scratch, is highly recommended, as it promotes interoperability and re-use of data and software tools.
- The second step, 'publishing the RDF data on the web', can be done by using any standard HTTP or FTP server to serve the RDF files (as you would with any HTML or XML file), or via a dedicated RDF server. Note that a very effective way of publishing RDF KOS data over the web is to follow Linked Data principles. A very good tutorial can be found in the Best Practice Recipes for Publishing RDF Vocabularies.
Q: What's the difference between 'concept-oriented' and 'term-oriented' models of thesaurus structure?
- A 'term-oriented' thesaurus uses terms (words or phrases) as its primitive elements, and asserts relationships between those elements, such as a 'term equivalence' link between animal and creature. The main elements of 'concept-oriented' thesaurus are more abstract concepts, which are aimed at capturing meanings beyond lexicalizations. In practice, while a term-based thesaurus would directly relate two terms with a same meaning using an equivalence link, a concept-based vocabulary would represent the same word or phrases as mere labels of a single concept. The SKOS model is clearly concept-oriented.
Q: Can I use SKOS to publish a glossary or other type of knowledge organization system?
- There is no theoretical objection for using SKOS in a much wider range of scenarios than the thesaurus case that originally motivated its design. Glossaries or folksonomies, for instance, may be ported to SKOS. Please bear in mind though that this might require some departure from the original knowledge structure to fit the SKOS model, or, alternatively, the creation of a specific 'profile' or extension of SKOS to fit the vocabulary at hand.
Q: Can I build a new knowledge organization system out of bits of someone else's?
- Re-using concepts from one vocabulary in another one perfectly fits the SKOS approach to representing and sharing knowlegde organization systems over the web. In particular, a concept may be asserted to belong to several concept schemes, as explained in the SKOS Primer
Q: Can I use SKOS properties for other purposes?
- SKOS properties were designed with re-useability in mind. This holds especially for the documentation properties (skos:note and its sub-properties like skos:definition, skos:example, etc.). The lexical labels skos:prefLabel, skos:altLabel and skos:hiddenLabel provide a means for labeling arbitrary resources, indicating that one single label per language should be used as the preferred one. Furthermore, preferred labels can be used to unambigously represent an item within the scope of a given concept scheme, data set or application. The rdfs:domain of all these properties is not restricted to skos:Concept.