According to Wikipedia, “the semantic web is expected to revolutionize scientific publishing, such as real-time publishing and sharing of experimental data on the Internet.” The W3C HCLS group’s Scientific Publishing Task Force is going to explore how this could happen.

Currently, one describes experiments in a more or less ad-hoc way. The mapping between experiments, papers, and titles is… well, not the most consistent ever.

Do you want to know if the experiment you have in mind is done already? Good luck mining the literature. Although mostly everyone is well-versed on building queries in scientific search engines, the task is far from accurate.

Maybe the problem is in the way we write the literature. If we could write a description of every experiment in some kind of agreed format that both humans and machines understand, searches would be trivial.

An alternative would be to use an ontology to describe experiments. The ontology should not be too complicated to use. If a user feels overwhelmed by the large number of parameters required to describe an experiment, this user may hesitate to do it. Of course, every field would need to built its own ontology. The effort to integrate ontologies across fields may be titanic.

There is some progress in the direction of using named entity extraction as metadata already. For example, the pubmed interface gopubmed is above and beyond anything I have seen. It uses facets (left sidebar) to show metadata. I do not know the details on how it works, but going back to say Web of Science after gopubmed feels like going 5 years back in time. Is there any hope to have a similar interface for all scientific databases? I sure hope so.