From the general point of view, the objective of the research documented in this work is to apply the potential of Knowledge Representation in the Web context. The important thing here is to detect the problematic issues detected in current web applications and to take profit from the new possibilities of Semantic Web tools. More concretely, the idea is to apply Semantic Web technologies to electronic commerce of multimedia contents and to the management of the copyright of these contents. Both application fields are deeply interrelated so they will be considered in conjunction in this work.
The Digital Rights Management and Rights Expression Languages analysis presented in the Rights Expression Languages chapter has highlighted some of the limitations of existing DRM initiatives and their associated RELs. First, there is a potential problem with the rights expression languages patent by ContentGuard. This patent covers any usage rights grammar. Therefore, any approach that is not syntax-based can potentially overcome the limitations imposed by such a generic patent.
Moreover, language grammars are not powerful enough to model some copyright statements, e.g. fair use, and they also make implementation more difficult due to their expressive limitations. They capture just a part of the language specification while the rest is captured in non-formal form. The captured part is the language syntax while the semantics are described in human-readable form. Thus, the semantics should be interpreted from this non-formal specification each time a REL processor is implemented or maintained.
Another important point is about Digital Rights Management Systems. As it has been shown, in order to build trusted systems, they must impose strong control over the end-user device and avoid piracy. However this is a very intrusive approach that leads to unsatisfactory user experiences, even in completely "legal scenarios". For instance, there can be interoperability problems between devices and contents governed by different DRMS or actions authorised by copyright law can be blocked by a DRMS, e.g. private copy.
This is also due to the expressive limitations of the technologies employed to date. RELs do not take into account the legal framework where the DRM systems that employ them operate. The copyright framework is too complicated to be captured by syntax-based tools. Until now, RELs are based on languages that just model the set of actions that can be authorised or prohibited to be performed on a given content for a given context. This usages model must be complemented with the legal framework that effectively governs it in the last term.
Finally, as it shown in the Control section, RELs have a problem with the level of detail they must support in order to cope with usage control. They have to specify every possible allowable usage in their grammar, or at least provide expansion points for this. However, at last, these expansion points must be completely detailed in order to fully specify the concrete action at hand. This is due to the lack of generalisation mechanism that is inherent to the language grammars used to formalise the RELs.
Therefore, the objective of this work is to face the disadvantage of existing Digital Rights Management Systems, which are due to the limitations of syntax-based rights expression languages. In order to face this objective, the state of the art has shown that knowledge representation techniques allow facing the previous inconveniences with guarantees. They come from a long tradition, previous to the appearance of computers, but that has shown its usefulness in the computers era.
The application context is also important and it must be taken into account. The Internet, and more concretely the World Wide Web, is the place where the more exigent requirements for DRM are being placed. Consequently, it is necessary to face the previous problems with this context in mind. As it has been also shown in the Web Technologies chapter, the World Wide Web has some limitations. The Semantic Web is conceived as a way to overcome them employing knowledge representation techniques, as it has been detailed in the Semantic Web chapter. Therefore, it seems appropriate to base this work on the Semantic Web in order to make a contribution to REL and DRM that incorporates the advantages of knowledge representation and is able to put them naturally into practice in the context of the Internet and the World Wide Web.
To summarise, the objective is to build an ontology that captures a great part of the copyright domain semantics and allows modelling rights expression that overcome current RELs limitations. Moreover, the target ontology is a Semantic Web one, which would smoothly integrate in the World Wide Web environment. This copyright ontology is going to constitute the main outcome of this work.
To conclude, it is important to note that despite the disadvantages detected in current REL initiatives, there is a great amount of work that has been carried out as it has been shown in the Rights Expression Languages chapter. A syntactic approach based on XML Schemas has been employed in order to define most of these RELs. These initiatives have performed an extensive analysis of the digital rights management field and, though the resulting grammars for the RELs capture some of the semantics highlighted during the analysis, they are not intended to formalise these implicit semantics. Consequently, they remain hidden when just syntax-based tools are employed.
An additional objective of this work is to take profit from all the hidden semantics in the existing REL grammars. In order to do that, it is necessary to formalise them using convenient tools, e.g. web ontologies and Semantic Web tools. A methodology for XML semantics reuse is going to be applied to the XML Schemas that define these grammars. This way, their implicit semantics can be made explicit and reused in order to complement the copyright ontology. Moreover, they can be also used to validate the copyright ontology.
The more appropriate context for this work is the Semantic Web, because it offers an open, heterogeneous and potentially global framework for knowledge management. By automating knowledge management at the web scale, humans can obtain a greater benefit from the knowledge they globally possess. Machines can help them manage this enormous amount of information if they have access to a significant part of its intended meaning.
The working hypothesis is that knowledge representation techniques in the Semantic Web, which allow taking a semantics-based approach, will provide new opportunities for rights expressions representation and digital rights management. First of all, with these knowledge representation tools, it is possible to avoid basing DRM solutions on rights language grammars that that are very likely to be subject to patent limitations.
Moreover, these semantics-based tools have a greater expressive power. Therefore, it is possible to capture the semantics that otherwise remain implicit, formalise them and make them easily available from the computer point of view. The first consequence of this is that it will facilitate enormously DRM tools development. Additionally, it will make possible to take profit from the expressive power in order to build generalisation structures that facilitate dealing with rights situations considering just the level of detail they require.
Moreover, if we go one step further the taxonomical tools that make generalisation structures possible, there are ontologies, detailed in the Ontology section. Ontologies are so expressive that they will be able to capture a great amount of the underlying legal framework and combine it with the usage models typical in RELs initiatives. Therefore, it will be possible to develop a copyright ontology that takes into account copyright law together with the common usage patterns of copyrighted content.
Ontologies have the additional benefit of facilitating evolvability and interoperability. Therefore, a copyright ontology can be defined with the required level of detail for a given application context and evolve later in order to cope with new situations and requirements. And these new requirements are going to appear for sure due to the dynamism of digital technologies and global networks and markets.
On the other hand, it is quite unlikely that there is going to be just a one-fits-all solution for rights expressions representation. Therefore, interoperability is going to be a key issue and ontologies an opportunity. As ontologies do not constraint the way things are written down, i.e. the grammar, but just what are we talking about, i.e. the semantics, it is easier to interoperate. A copyright ontology will thus also facilitate interoperability among different RELs. Moreover, it will be easily enriched with the semantics that will be reused from existing initiatives, which will facilitate the development of the copyright ontology, its validation and enable it as a key tool for DRM interoperability and integration.