The first portion of the NRMIG meeting featured a presentation on taxonomy development by Laura Dorricott, Project Delivery Manager of Taxonomy services with Dow Jones. Her presentation can be viewed at: http://presentations.ala.org/index.php?title=Sunday,_January_25
- Taxonomies are part of an “evolutionary path” featuring the following elements:
Dictionaries & flat lists –> Structured authority files –> Hierarchical taxonomies –> Controlled vocabulary thesauri –> Ontologies. - Taxonomies form the building blocks for ontologies; ontologies are semantic representations of the real world in all its rich diversity.
- The purposes of controlled vocabulary include translation, consistency, indication of semantic relationships, hierarchical relationships to assist browsing, search and retrieval (precision and recall). The return of investment comes when you enable someone to save time and increase productivity.
- In response to a question: folksonomies are actually in a different class than taxonomies, as they generally consist of social tagging that are not typically hierarchical.
- Keyword searching has many drawbacks and is insufficient for information seeking purposes. Taxonomy helps people filter out the noise and discover relevant information needs regardless of what they’re labeled.
- Search and navigation are not alternative, but complimentary solutions that serve audiences well considering the multiple viewpoints they bring to the table.
- When building taxonomies and controlled vocabularies, you must account for ambiguity (polysemes), synonymy, semantic relationships (hierarchical and associative), facets, warrant, structures, metadata (controlled vocabulary).
- In response to a list of standards, Diane Hillman recommended that SKOS (Simple Knowledge Organization System) be added to that list(see http://metadataregistry.org for more information).
- What is the rationale for creating a thesaurus, which is usually labor-intensive and requires specialized software when it grows beyond a certain size? A revealing statistic is that 40% of corporate users can’t find the information they need to do their jobs on the intranet. Companies have many information retrieval issues which result in loss of productivity and profits. There is significant value in using controlled vocabularies, including improving productivity, reducing costs, gaining competitive advantages, driving usage, driving cultural change, and leveraging information management skills.
- In response to a question about corporate versus library taxonomy, she revealed that enterprise-wide taxonomies usually feature around fifteen terms up top, fairly broad-based and not very deep. However, if the client offers information about information, they will have subject based taxonomy, such as libraries have. You can see many examples of taxonomies at: http://taxonomywarehouse.com/
- In terms of what experience is needed, Laura responded that she most frequently hires people who have experience building taxonomies. Sometimes you do need subject expertise. The client should be involved in all stages and she uses a live session that the client can view.
- What prompts companies to hire taxonomists? They often reach a point where they have excess information and their situation becomes urgent, especially considering the increasing regulations about managing documents. It is expensive to do manual indexing of all documents and a taxonomy provides a framework for both auto-classifying and manual indexing. Additionally, acquisition of companies often results in a need for taxonomies.