We've been working on best practice and a scalable framework across disciplines for metadata.

It is apparent, and has been for a long-time, that people do not like adding metadata to files. This is not only true of researchers, but anyone. Photos are a good example, not many people tag their photos, and those that do tend not to use more than one or twof tags.

There is, however, a lot of excellent work on taxonomies and ontologies, and a move towards linked data, that requires a significantly increased level of metadata.

Our conclusions are that this is a long-road, and that we should try and encourage people to at least tag their data with a minimum amount of metadata. We can then extend this to add value to not only the data itself, but also to the researcher's time. There must be an incentive for the individual, preferably at the time of deposit, or else it is extremely difficult (impossible?) to convince researchers that spending time on this activity is valuable.

In light of this we have proposed a three-level metadat strategy, which is outlined below. We are working to automate the core metadata tagging through the operating system/application software.