In principle every organisation, be it a private enterprise or a public agency, needs to reflect on the handling of documents governing internal as well as external actions. The concept of authority covers here a wide variety of actions, e.g. authorisation to enter into a contract, law enforcement, as well as the right of starting up a manufacturing industrial process. The underlying assumption is that these actions are derived from some kind of normative policy-making document.

With this in mind one may reflect upon the fact that open systems are a major development trend in today´s communications networks. One important concern is, therefore, the best ways in which to secure trails of authorisations, alterations included. More precisely, this is a matter of information security policies mirroring the norms that govern an organisation, such as who has right of access to what, without knowing beforehand who will be claiming this right of access.

This work package concerns the ways in which the use of standardised markup languages, in combination with advanced security-enhancing methods, may facilitate the management of version-dependent text entities over time. The primary application area is regulatory management. In a broad sense this comprises not only conventional legal documents such as rules and regulations of different kinds, but also documents, which in a long-term perspective may be of legal relevance to an organisation. In a future court trial this may be the case, for instance, with regard to documents containing authorisations for actions taken. A major point is the need to consider these issues in the early design of a system.

Information retrieval systems are designed to solve the very general task of finding documents given a short query. They are less typically designed to suit the needs of any specific users. Legal information, as well as information in every domain, is specific as to its nature, but even more importantly, legal information users typically have different needs from the general public. Information retrieval systems in general do not acknowledge this type of tailoring needs to specific domains, and are designed to conceal what they deem as spurious information from texts, rather than display it to readers even if it might be useful in some domains.

This work package deals with legal information usage from the system design perspective, and will deliver results on information seeking behaviour in the legal domain, a system design sketch based on those observations, and as a third item a prototype to experiment with what items obtained from state-of-the art text analysis can be used for the needs of legal information access. We expect issues of trust and authentication to be central for information access in this domain.

The integration of a legal perspective at a stage when a certain application has already been developed proves in practice often to be both timeconsuming and costly. It may in fact turn out to be impossible to implement good legal solutions when a technical design has been fixed and contracts have been signed. The attraction of considering document markup languages in combination with security-enhancing methods in this context lies in the need for transparent, content-dependent and context-sensitive management of legally relevant text units over time. The primary key words are thus legal awareness, XML, and security. The important criterion of success is early consideration of these critical factors. It is important to note that the aim is not to accomplish rule enforcement, but to provide a methodological infrastructure for tasks that require practical solutions. The major characteristics of this work package are those of conventional legally dogmatic ones. This means that focus will be placed on substantive law issues related to the area of information security. The EC Directives on E-commerce (2000/31/EC), E-signature (1999/93/EC) and Data Protection (95/46/EC) are of particular relevance here. The following serves as overall hypotheses:

• Security related legislative activities (within the areas of data protection, e-commerce etc.) ought to be combined with strategies for regulatory management.
• Technical means are vital in this process, but must be legally customised.
• There is a need to build legally founded trust infrastructures.

The SLIM Project is expected to provide refined methods of legal security enhancement based on the management of information in context. In other words, the main interest lies in data integrity in complex data structures. In terms of specific applications, the initial aim is, as mentioned above, to support regulatory management in general with an extension to activities related to e-business — for instance, successive drafting of contracts. Of particular interest here is the possibility of securing negotiated clauses during an electronic process of drafting. Results will be captured both in the form of a technical demonstrator that may be described as a proof of concepts and in conventional scientific documentation.