The future plans for this work can be classified in three lines. First, to complete some of the research and development lines already initiated. This future work line is mainly related to the mapping of differente REL initiatives to the Copyright Ontology. Second, to develop a semantics-enabled DRM system that fully exploits the Semantic Web approach to DRM. And third, to add new features to the semantics-enabled DRM system build up by the conjuntion of all the contributions of this work.
As it has been said in the Evaluation chapter, it is still necessary to complete the mappings from the MPEG-21 REL and RDD Ontologies and from the ODRL Ontology to the Copyright Ontology. This mappings should be complemented with a mapping from Creative Commons (CC). Although CC is a very simple rights expression language based on a simple RDF Schema, it is the more widely used one nowadays. The main Creative Commons features that seem the reason for its success are its simplicity, its inclusion of copyright terms and the availability of human-readable versions of the licenses in natural language.
The NewMARS web application is the current implementation of the Copyright Ontology framework in order to build a semantics-enabled Digital Rights Management System. As it has been shown in the Semantic DRM System section, it is a quite complete implementation of a Semantic DRMS. However, there are some functionalities that are pending a proper implementation. The main caveat is to build it on top of a semantic metadata store that scales up to great amounts of metadata and which provides a complete solution for OWL-DL reasoning and Semantic Web Rules support.
A part from the previous future lines that are geared toward completing already initiated work features, some new features, which would be desirable for a Semantic DRM System, have been detected. First of all is to improve the security and trust of the Semantic DRMS. Another interesting point is to mimic one of the Creative Commons features, i.e. to provide a natural language version of the rights expressions.
The inclusion of security issues that transform the e-commerce application into a trusted one has not been initiated yet. This requires the application of Web of trust ideas and Public Key Infrastructure technologies. The primary intention is requiring that each actor taking part in the systems have its own digital certificate. This certificate, with its corresponding private key, will be used for digitally signing all the statements done by this actor (agreements, offers, assertions…) so responsibility can be tracked later and even produce contracts.
The plan is to apply digital signatures at the level of the RDF graph segmentation already employed in NewMARS in order to facilitate browsing metadata, as it is detailed in the corresponding Metadata Retrieval section. A similar approach has been already applied in the DBin project [Tummarello05].
Another interesting feature is to provide a natural language version of the rights expressions in order to facilitate human-user interaction with the Semantic DRM System. This is valuable feature of the Creative Commons initiative. However, in the CC case, the REL is quite simple and the human-readable licenses are not generated autommatically, they are generated a priori and they just provide the legal framework for the license, i.e. they are not personalised for the current copyrighted content.
The idea is to take profit from the linguistic foundations of the Copyright Ontology, i.e. the use of verbs and case roles as the modelling building blocks. Due to this linguistic base, it is possible to generate Controlled Natural Languages [Schwitter05] from the Copyright Ontology expressions. Many different natural languages can be supported. Starting from a mapping from the Copyright Ontology to the lexical resource Wordnet [Fellbaum98], it is possible to autommatically retrieve the lexemes corresponding to Copyright Ontology concepts in English but also in other languages, for instance those present in EuroWordnet [Vossen98].