To dramatically improve the efficiency and usability of applications implementing any kind of information sharing, and to radically simplify their development, we have to step aside a little from the actual applications. By doing so we can realise that the goals and problems of digital libraries, global directories, file sharing and searching, gateways, various messaging apps, and of the PKI can easily be mapped to database terminology. It's hard to find an application area - and thus a protocol or API used - on our PCs or on the Internet whose function cannot be described by the patterns of authentication, right management, querying and updating, metadata and active tasks, and whose knowledge cannot be contained in tables. The idea is not new: let's create an infrastructure in which we can uniformly and easily access any local or remote data source, and use it instead of various application specific interfaces. Beyond simplifying access we enable further utilisation and combination of information coming from these sources - and in general the development of real life applications relying on the flow of information. Most apps do so...

No question it's a tough thing to do. In our opinion not only technological but ideological departure is necessary from some of the current trends. We see the solution in utilising the methods of proven relational database methodology. We'd like to view sources and - it's important - their metadata as relational entities, without modifying their internal structure. This idea, although not new, will probably spark a lot of questions from experts. To satisfy this in our presentation, beside dwelling more deeply into our ideas, research and possible results, we'll compare all this to other approaches - like the ones evolving around XML, on hierarchical database ideas, or on object oriented RPC (Corba, DCOM), and to some existing standards like ODBC and iFS. We'll show the logic behind our choices, and some typical/expected strategic pitfalls, those we try to avoid. We'll try to prove why the old model of tables, fields and relations (with a modern twist) is the right choice - through real life examples - some well-known ones from the NIIF organisation too. In addition through the through the Web's example we'll briefly explain why free software is not only an ideological but a strategic choice for a project like this.