KeyBase is a database and web application for managing and deploying interactive dichotomous keys to flora and fauna. Based on concepts originally developed for the Lucid Phoenix project at the Centre for Biological Information Technology, The University of Queensland, KeyBase provides an environment where dichotomous keys, traditionally developed for print, can be more easily and effectively deployed and used.
Keys in KeyBase are arranged into projects. A project is a group of keys related in some way, usually being keys for a particular taxonomic group in a particular geographic region such as a country, state or bioregion. As well as being grouped by project, keys in KeyBase may be arranged in a taxonomic hierarchy. For example, one key in a project may be to families, others may be to genera within families and others may be to species within families or subspecies within species. KeyBase automatically links keys in a project together using the taxonomic scope of the key. A key to the species within genus A will be automatically linked by KeyBase to any instances of A in other keys within the project. In this way, keys in KeyBase are seamlessly linked together, allowing users to follow an identification through a KeyBase project from the highest to the lowest available taxonomic level.
A question that’s sometimes asked is “Why bother with dichotomous keys, when systems such as Lucid and DELTA allow us to create and deploy really neat, interactive, random-access keys?” The answer is that both traditional dichotomous keys and modern random-access keys have strengths and weaknesses, and both have a continuing role in identification.
Two principal strengths of dichotomous keys are that there are a lot of them available (because they are relatively straightforward and quick to create, especially for small to moderate-sized groups), and they are very efficient, accurate and informative when used effectively.
Both these are related to the information efficiency of dichotomous keys – for any given step in the key (except the first), only a subset of taxa need to be considered. This saves time and work in creating the key, and ensures that the author of the key can optimize each step in the key for the taxa in play, and readily ensure accuracy in applying that couplet to the taxa. In formal terms, these mean that dichotomous keys are less prone to Type II errors (falsely miscoding a taxon as not having a feature).
Another and important, though subtle, advantage of dichotomous keys is that they are great pedagogical devices, useful for learning about groups of organisms. Users quickly become familiar with the critical features of a group of taxa once a dichotomous key has been used a few times.
While dichotomous keys have been used and found useful for centuries, they are traditionally deployed in only one way – printed on paper. An underlying principle of KeyBase is that dichotomous are more interesting than that, and can be taught to be as interesting as the web allows.
For that reason, the keys in KeyBase are stored in a relational database, and this allows them to be deployed more flexibly. As a simple example, KeyBase keys can be easily represented as either bracketed or indented keys. More interestingly, keys can be deployed using the KeyBase Player. As an identification proceeds in the Player, KeyBase keeps track of and displays the taxa that remain in play and those that have been discarded. Using the Player it’s also easy to backtrack and “play” with the key.
One of the features of KeyBase keys is the ability to filter one or more keys. Any key, or any set of keys forming a project, can be filtered to produce keys to specified subsets of taxa. This allows you to ask questions like “What’s the key difference between these two taxa?”, or “I’m at this location, and have a list of species that are known to occur near here – can KeyBase give me a key to only those taxa?” Filtering makes KeyBase keys much more powerful and easy to use.
We have many ideas for future enhancements for KeyBase. These include: