Tools for Managing Taxonomies (or Thesauri, or Ontologies)
Taxonomy consultants, such as those at Earley & Associates, may be the ones who develop a taxonomy for an organization, but the organization's own staff will ultimately be responsible for maintaining it, so the question arises what tool or tools should be used the maintain that taxonomy and perhaps further develop it. A taxonomy may be implemented in a CMS, in SharePoint, or with search (Google Search Appliance, FAST, etc.), but these systems do not have taxonomy management components.
An interest in taxonomy tools was evident by the number of chat-based questions that my colleague Seth Maislin and I received from participants in this week’s Taxonomy Community of Practice Call, Cross-Mapping Taxonomies, which we jointly presented. There is a need for tools that do more than merely enabling manual adding and deleting of terms. Mapping two taxonomies is something that only a few tools support, but there are many other day-to-day taxonomy management activities that also require specialized taxonomy management software.
This week several Earley & Associates consultants, including myself, participated in a special training on Smartlogic Ontology Manager, a good example of full-featured taxonomy management software. The question arises: is this taxonomy management or ontology management software?
If we look at competing software products, we see various designations:
- Wordmap – Taxonomy Management Software
- Data Harmony – Thesaurus Master
- Synaptica – Knowledge Management System
- SchemaLogic – metadata management software
Trying to choose a taxonomy management system by evaluating price, support and services, integration capabilities, data interoperability standards, and the inclusion of some very significant add-ons for indexing or auto-classification, is already a huge task. To consider whether the product’s emphasis should be taxonomy, thesaurus, knowledge management, metadata, or ontology can add another level of confusion.
Does the software’s self-described type, subtitle name, or tagline mean it’s really different from the others? Yes and no. Let’s face it: the distinctions between controlled vocabulary, taxonomy, thesaurus, ontology can be blurred, and sometimes these go by other names, such as knowledge organization system or metadata model. Nevertheless, product names and taglines are very carefully chosen to reflect the targeted use. For the most part, though, I see these products as “crossovers” in some way or another.
Most products cross over between taxonomies and thesauri, whereby you can omit associative (related term) relationships in a taxonomy but include them, along with scope notes, to create a thesaurus. Even the low-end product MultiTes does that. (I’ve also seen a tool that enforced top-down structure building, so this would be more a taxonomy tool than a thesaurus tool.) Whether supporting the creation of custom semantic relationships alone is enough for a taxonomy/thesaurus tool to cross over and become an ontology tool, however, I am not so sure. SmartLogic’s Ontology Manager, though, is definitely a crossover between a thesaurus management and an ontology management tool. Not only can the user define relationships, but there is also the option to set up classes whose use is strictly tied to designated inter-term relationships. Class thus conforms to the ontology use, but could also be used simply as a means of subcategorizing a thesaurus.
There are pure ontology tools out there, a well-known one being the open source Protégé. It’s such a pure ontology tool, that it doesn’t even have a feature for nonpreferred (synonym) terms, requiring some workarounds to include such terms. I know of a company planning to implement it as their taxonomy management system, because they like the fact that it’s open source and they have an engineer who has learned to like it. But I would caution not to use a tool just because it’s free. Protégé is difficult to learn to use, and with multiple people managing the taxonomy, multiple people may need to get trained, and Protégé’s training (at Stanford University) is expensive.
Ultimately, you need to determine what kind of knowledge management system/model you need to build. Can it be a simple hierarchical taxonomy? Does it need the support of nonpreferred terms? Should it have associative relationships as a thesauri? Does it need customized semantic relationships between various classes of terms, as an ontology? It might make sense to start out building your controlled vocabulary with an easy-to-use, inexpensive tool such as MultiTes and then trade up when you need further capabilities. Smartlogic’s Ontology Manager, for example, allows importing data from various formats including MultiTes files.