Just as web applications can benefit from the fact that a topic map stores the application schema as data, so may any other application which must deal with structured information. Again, for the developer the advantages are:
A data model which typically matches the decomposition of application design into a set of interacting objects.
A data model which can be modified simply by altering the data which provides the application schema, and without the need to re-compile or re-populate database tables.
A single API for accessing the data.
These advantages can be applied in two ways. Firstly, even when the complete data model of the application is known at design time, the use of topic maps makes it easier to modify the design as development progresses and to later refine the design without requiring users to upgrade their database. However, the flexibility of the topic map approach also enables applications to be developed in which the complete application data model is not known at design time but is configured or created by the end user.
Bravo, developed by GlobalWisdom Inc. is a knowledge management tool which combines both explicit and automated categorization of documents with user-feedback to determine the relevance of documents to any given query. The operation of Bravo is more complex than an automated categorisation of documents, as the system learns what documents are relevant to a particular query by both explicit feedback from the users (where a user rates a result for its relevance) and by implicit feedback (where the system determines the document's relevance to the user's query based upon what actions they do or do not perform on the document). Additionally the Bravo system enables individual users to be recognised as experts in particular subject areas and for other users to modify their results sets to promote the documents that the 'expert' claims to be most relevant. What is different in Bravo is that this expert may never have explicitly marked documents as relevant, instead the user is filtering the result set according to the expert's profile, which, in turn, is built through the expert's interactions with the software. This would include many documents that the expert may never have seen, but that Bravo has, using the profile, identified as relevant.
The information about documents; the concepts related to the documents; the 'experts' in the different conceptual areas; and the relationship between concepts are all stored by Bravo using a topic map. Bryan Thompson, CEO of GlobalWisdom, says "We use a topic map engine to provide a sophisticated and scalable information architecture. This serves as a differentiator for us and helps us to address more upscale markets.". GlobalWisdom chose not to implement their own topic map engine but instead to make use of the K42 engine from Empolis. "You could consider us a value-add for a topic map engine, and visa-versa" comments Thompson. Naturally, other means of representing the information structures required by Bravo are available and were considered - including directory servers and database systems, but topic maps were chosen because of the richness of the basic information architecture that they provide.
For GlobalWisdom, using topic maps as their underlying architecture and choosing to use a third-party engine has enabled them to bring a complex, scalable solution for to market more rapidly than they would otherwise have been able to. For customers of GlobalWisdom, although they need never be aware of the topic map technology that hides beneath the Bravo application, the topic or concept-centric organisation provides faster, more accurate access to relevant information, reducing the amount of time that people need to spend searching for the information that enables them to do their work.
| Up: Topic Maps - A Practical Introduction With Case Studies | |
| Previous: Web Application Development | Next: Application Integration |