Semantic Blogging Research

HP has been conducting research into fusing structured metadata into weblogs. They also have a demonstration blog set up to see their ideas in action. (Found via Open Access News.)

Semantic blogging exploits this same personal publishing, syndication, aggregation and subscription model but applies it to structured items with richer metadata data. The metadata would include classification of the items into one or more topic ontologies, semantic links between items (“supports”, “refutes”, “extends” etc.) as well as less formal annotations and ratings. There are several ways this more structured data could extend the power of blogging:

Discovery. At present is it not easy to discover either a channel of interest (e.g. “I would like to find blog channels about the semantic web”) or a collection of specific items of interest (e.g. “Are there any more blog entries describing this application idea?”).

Cross-linking. Current blogs support a single link between the channel record and the blogged item. By extending this mechanism to support linking between items (using a property hierarchy) we can create a network of topic interconnections that supports more flexible navigation. These links can themselves form part of the disseminated content – for example to represent the structure or scholarly discourse.

Flexible aggregation and selection. The current blog subscription mechanisms are in some ways both too fine (being bounded by the individual blogger’s channel of posts) and too coarse (e.g. I might like Ian’s technology channel but am only interested in the semantic web bits). The richer categorization and structure of semantic blog channels would make it easier for users to create virtual blog channels which aggregate across multiple bloggers but select from that aggregate according to other criteria such as topic (or community rating).

Integration with other sources and applications. The structured nature of semantic blog channels makes it possible to develop automated blog robots that can process and enhance the blogged items. For example, in the bibliography domain transducers would enable import and export via existing bibliography schemas like BibTex and automatic linking to large repositories such as CiteSeer.

There are lots of differing opinions as to whether the semantic web can actually be achieved. It’s good to see some actual research being done to shed some more light on this issue. As I’ve said before, I think metadata can still be useful in discrete communities and/or collections where there is some control and incentive to code accurately. I don’t think it will likely work for the web as whole given the web community’s tendency to game systems. A semantic blog network could be quite interesting for a community of researchers.