Corporate news

You are one of the leading international experts on Topic Maps. What is your approach towards Topic Maps? What are your fields of interest?

One of the key properties, or sweet spots of Topic Maps over other approaches is that it successfully brings together people with machine processing. People can understand and relate to Topics and Associations, and can relate to the notion of organising content using topics. At the same time, Topic Maps has a generic data model that is extremely regular and easy to process in a consistent fashion.

All this makes Topic Maps a very powerful way to meet the expectations of large user groups who are interested in more powerful and expressive ways to organise and find content.

What are current developments concerning the Topic Maps data model, querying language and inferencing standards? Can you give a brief update from the technological perspective?

The Topic Maps Data Model has been stable for some time now so there is no change there. The main standardisation work has been around TMCL (Topic Maps Constraint Language) and TMQL (Topic Maps Query Language). Both of these are now at FCD (Final Committee Draft) in the ISO process.  TMCL is a constraint language as opposed to something like OWL, which is more about inference. We see many information modellers come from a data modelling background, such as ER diagrams or UML and they find it easy to use and understand TMCL as a way to define Topic Map schemas.

Do you have practical examples of current Topic Maps adoption in industry?

Without mentioning specific names we have seen a strong adoption of Topic Maps across the board. With large corporations in the areas of International Shipping, Telecoms, Oil and Energy adopting Topic Maps as a key component in both intranet and internet solutions. Along with this has been further adoption by government at the highest of levels.

Topic Maps have traditionally been designed for inhouse use for knowledge representation and management. But where is the web in Topic Maps?

Topic Maps fits in very well with Web 3.0, Linked Data etc. In fact not only can Topic Maps data be exposed and used as Linked Data but one of the core aspects of Topic Maps, the clarification of when URIs are used to identify a thing and when they are the address of some resource could be a useful addition to the Web 3.0 and Linked Data initiative. Networked Planet recently launched http://subj3ct.com, an aggregator of URIs that are used to identify subjects in the real world. The service stores and exposes identifiers, equivalent identifiers and links to online structured resources for a given subject. The service has taken the key aspects of Topic Maps and turned them into what we see as a unique service for Web 3.0 and Linked Data.

Topic Maps and the Semantic Web have lots im common from a conceptual perspective. Where do you perceive differences? Where do they intertwine?

Both communities are already borrowing from each other in many ways, even if it doesn’t seem like that much. Semantic Web advocates realised that they needed mechanisms to distinguish URIs for things from URIs for resources, they did it at the protocol level using a 303 redirect.

Topic Maps makes it part of the model, which I certainly prefer. Topic Maps has seen from the Web 3.0 work that it needs to do more on the web, hence things like http://subj3ct.com. We also see RDF vocabularies' such as SKOS which embody the higher level notions of Topic Maps.

What is to expect from Topic Maps in the future?

I think that the standards work will turn from being about standardising the data structures and syntaxes to standardising protocols to make the web of Topic Maps work more effectively. In general, I see Topic Maps moving more into the web and more into the mainstream. One of the key drivers will not only be the growing awareness by people of taxonomies but a realisation that, they are not enough and that they need something more.

About TMRA 2009

For more details please go to http://tmra.de/