Semantic Desktop - Nepomuk recommendation and searching

The project has 2 major parts - Metadata processing and developing applications using this powerful metadata.
Processing Metadata of Nepomuk
     Right now, Linking of resources is not done automatically by Nepomuk.We need to create tags for each resource manually. Since the content of all files are already indexed and saved as part of the Nepomuk metadata, we could use this to automatically tag all the resources.

Say, we have multiple files talking about 'technology', tags could be created based on their content and link them to each other accordingly. Similarly, resource linking could be carried out using other metadata automatically.

     Finally, this might lead to a more semantic desktop, where all resources are inter related.A lot of new functionality can be added using this information.

Voice/text based desktop search engine
    The user asks ( or types in ) a question, and he is provided with all the files/resources which match the criteria.
eg. He may ask 'Which movie did I watch last Saturday'
     He will be provided with the list of movies he watched that day.

Natural Language processing will be used to process the question, and using metadata stored in Soprano ( part of Nepomuk Framework ) the results are generated.A common voice to text converter could be used for audio queries.

People with physical disabilities could be benefited like this. It would make the desktop usage more interactive adding a real life perspective to desktop usage.

Intelligent file browser
    Plugin for file browser which suggests files/resources which are similar to the file currently being used/selected.
eg: Say user selects a song of a particular band, then he would be suggested songs from the same band or any other metadata.