"This paper presents creation of a digital library of previous IAPS proceedings. A large number of conferences resulted in an interesting output of scientific papers and it makes sense to have this wider and easier available to the audience. The currently used interface of the IAPS Digital Library was created in the framework of an EU-IST project, called SciX (Open Scientific Exchange of Information). First results on the gathering of initial content will be presented as well as related experiences. The SciX-project is funded by the European Commission in order to demonstrate the feasibility of alternative models of scientific publishing made possible by the Internet. The project includes both theoretical work in making a formal model of the scientific publishing process, to be used as a basis for studying the life-cycle costs of alternative business models, and demonstrators of functioning e-prints archives.The concept of self-organisation refers to the key idea that repositories would not be maintained by professional publishers, but by associations that do not have a commercial interest in managing the repository. The repositories should be created in such a way that they could be managed with a minimum of human effort. A major effort in managing these repositories is the organisation of the content - how to group papers into categories, how to create hierarchical directories, like Yahoo, across the papers in the repository, etc. Not unimportant is the management of the users and maintenance of the consistency of the data. An important aspect of SciX is applying text-learning techniques to automatically categorize the content. However, self-organisation means that decisions towards data entry and extensions of available information are not taken by the SciX-consortium, but remain in the hands of the association in charge. Furthermore, a splitting of workload can take place among a number of collaborating activists, so that an accumulation of smaller tasks is feasible. The criterion of "ease to use" has to be mentioned as a prerequisite, both for the maintainers as the endusers. Proprietary software packages normally do not meet this criterion; data entry and viewing have to preferably be performed from any internet browser and furthermore, a minimal level of computer literacy should be required. In many cases the editing of the data is being handled by non-librarians, which means that data will not necessarily be imported into the Digital Library as "clean" and "proper" as a library would prefer. Maybe from a librarian's point of view a "quick and dirty" approach is not preferred, but in this context for the enduser, hardly less effective. Furthermore, as the full text version is attached to the record the justification of the need for an ultimate precise citation is narrowed.The content providers behind a repository are responsible for what kind of information they input and how detailed or secured the pieces of information are in the projected repository. This does, for example, not cause serious problems in an academic context as long as reviewed publications (i.e. conference proceedings) are recorded. However, individual submissions with a variety of publication sources may lead to the installation of an advisory board who supports decisions concerning the relevance of entries, insurance of quality control, topical relevance, etc. For the moment, this is not planned by IAPS, as the inclusion of previous proceedings has the highest priority.Hardly any data from the early days of IAPS has been archived in a digital format. E-prints were therefore produced by means of scanning and then applying OCR to the scanned images for the creation of metadata (authors, title and summary). The "reuse" of archived digital data sources, which may be in general only available to more recent publications, is more efficient as additional human effort is minimal. In order to allow for as much re-use of existing digital material, SciX is not only aiming at tools that support digital libraries, but will support large parts of the publication process including the support for a conference organisation, reviewing, etc. The setup of an interface for submission and review for IAPS 2004 has to be mentioned in this respect.In the framework of the IAPS17-conference (2002) first ideas on setting up a Digital library were presented. However, further research had to be conducted and the rough estimation of publication output turned out to be larger as assumed. Before 1982 nearly 3.000 pages were published. Between 1982 and 2002 the output turned out to be around 9.000 pages. It can be stated, that the IAPS materials published before 1994 are rather difficult to trace back and digitalisation would allow for wide accessibility. Easy access to IAPS publications offers a great potential for new members, such as interested researchers from Eastern Europe."