"The Art of Open Archives" was an expert meeting for specialists in the field of information architecture, archives and online databases. Recently, many collections from archives and cultural heritage institutions have become accessible on the World Wide Web; this is also the case in the area of electronic art where a growing number of organizations (Ars Electronica, V2_, ZKM and many others) are creating web portals that contain information about their recent history.

But information on many topics - common themes for example, or oeuvres of artists - is dispersed over many different resources; users and visitors have to visit many resources simultaneously in order to get a complete picture about such subjects.

The issue of archival interoperability, or connecting several resources and creating good user interfaces for these connected, distributed databases, was the main topic of this expert meeting. During the discussions, though, many other issues were brought up, including the subject of "open archives" in a broad sense - allowing the community of users to contribute information to the various systems' resources.

Semantic Web, Open Archives and Dynamic Presentation

At the beginning of the meeting, three speakers - Ivan Herman, Janneke van Kersen and Lynda Hardman - presented current research and technical instruments in the area of annotated archives and interoperability.

First, the basic technologies were presented that can enable connections between web-based resources. The Semantic Web, according to Wikipedia, "is a project that intends to create a universal medium for information exchange by giving meaning, in a manner understandable by machines, to the content of documents on the Web." Semantic Web technologies such as XML (eXtensible Markup Language), RDF (Resource Description Framework), RDFS (RDF Schema) and ontologies described in OWL (Web Ontology Language) promise to offer an appropriate solution for making web-based resources, including cultural databases and archives, interoperable.

Ivan Herman, Head of Offices at W3C, the World Wide Web Consortium, gave a brief overview of the Semantic Web and its various standards and markup languages for describing and annotating resources. He described the various "layers" of standards with which semantic web content can be described, from low-level RDF to ontologies and ended with a variety of application examples of tools, projects and resources that already use or generate Semantic Web descriptions, including the Dublin Core, RSS, Adobe Photoshop and MuseoSuomi, a Finnish project for connecting museums' collection databases. Ivan's lecture slides will remain accessible at http://www.w3.org/2004/Talks/1111-Rotterdam-IH/.

Janneke van Kersen, consultant for Vereniging Digitaal Erfgoed Nederland (DEN - the Netherlands Digital Heritage Association) gave an illustrative presentation on the Dutch-oriented project "Cultuurwijzer" and related efforts to make Dutch cultural heritage resources interoperable. DEN plays a stimulating role here, and has chosen to work with the Open Archives Initiative Metadata Harvesting Protocol (OAI-MHP), an RDF- and Dublin Core-based interoperability standard that has already been in use for a few years in the world of scientific publishing. An advantage of this protocol, which can be situated in the low-level area of Semantic Web languages, is the fact that participating institutions maintain control over their own data.

In the application of OAI-MHP, several levels of problems are encountered - technical, organizational and content-based; for example, a sample survey within Dutch cultural heritage institutions has shown that only four organizations use the Dublin Core standard correctly. Another problem is the fact that there are major differences in description level between institutions; the archive world typically uses a hierarchical model (most notably ISAD(G)), while the museum world uses object-level descriptions for its collections.
The OAI-MHP is now based on XML - a format that is already very new to most heritage institutions - but it is to be expected that, in the future, RDF will come into the picture. DEN is investigating the possibility to include ontologies in this system; the CIDOC Conceptual Reference Model (CIDOC CRM), an ISO standard ontology for the cultural sector, may offer an appropriate solution here.

Next, Lynda Hardman, head of the Multimedia and Human-Computer Interaction theme at the Centrum voor Wiskunde en Informatica (CWI), gave a presentation that illustrated the potential for dynamic user presentations for collections and resources that use Semantic Web annotations. What is these formats' added value for the end user? Using the example of the Rijksmuseum's collections of Rembrandt paintings and information on the "chiaroscuro" technique, Lynda Hardman demonstrated various presentation tools, developed by CWI. These instruments make it possible to present information adapted to the user's profile and his devices, through targeted selection of content, choice of design and layout, and style of interaction. The so-called Cuypers system generates presentations based on these parameters. The Topia project goes even further than this by offering the user suggestions on classification.
Lynda's presentation is accessible from http://homepages.cwi.nl/~lynda/talks/2004/DEAF04/.

Roundtables and panel discussion

After these plenary presentations, a small information market took place, with representatives from several online archives and research projects in the field of digital culture and electronic art.

Next, three short parallel roundtable discussions were held; all three were summarized during the final panel discussion.

One roundtable offered a follow-up to the presentation by Lynda Hardman, focusing on the dynamic presentation of semantically annotated content. During the discussion, three major topics were touched upon:
1. Elderly people can contribute to a repository, but are often unexperienced with technologies involved.
After a short introduction round, the interests of the people present were inventorized. Gemeentearchief Rotterdam mainly struggles with interoperability of the seperate archives of different departments. Also they consider ways to sell and promote the contents to users, mainly elderly people, interested in personal histories.
In the m-cult project, Susanna Koskinen experienced the tension between public and private archiving. Elderly people can contribute greatly to a repository, but technology often forms a barrier for participation. M-cult experimented with interfaces between single clients and a large database-server, using automatic film editors, attaching metadata to the clips. 
2. Interfaces can be an obstacle for professional users.
Instituut Collectie Nederland has some experience with active annotations. Within the framework of the INCCA project, they experimented with the construction of an modern art repository by letting the artists/professionals annotate the contents themselves. The idea was to build a database of knowledge about modern art conservation and restoration techniques by involving the professionals in the field. Members of INCCA could contribute to the database by adding metadata about an artwork or about conservation techniques through a simple interface. Althoug the  interface of the public editor was easy to use, few people took the effort of adding  metadata. Too many or too complex metadata fields are an obstacle for professional users, but you also don't want to limit the amount of metadata, necessary for an rich database.
3. In the case of an archive open to the public, how to keep your data clean and how to configure it for the end user?
Opening up your archive to the public raises the question of how to keep your data clean and how to configure it for the end user. The choice between making a semantically annotated archive or an interoperable archive is a paradox.

The second roundtable looked more in depth at Semantic Web technologies and their applicability in a variety of contexts. One participant here remarked that his experiences with advanced web-based services were rather negative - not due to technology restrictions, but to the amount of work and complexity for human content providers, although more and more metadata can be generated automatically - by authoring tools and devices such as cameras, for example. Apart from that, a middle ground has to be found - depending on the application - between metadata provided by users and experts. It remains important to clearly state the goals; what do you intend to achieve?
Furthermore, the complexity of working with Semantic Web annotations can mostly be attributed to the challenge of describing and understanding a domain (for example, which concepts should be included in an ontology about electronic art?).
All in all, it is important to emphasize that the Semantic Web is still a young initiative that will evolve over the coming years.

Andreas Broeckmann was the moderator of the third roundtable, which focused on community wishes and needs. Diverse subjects were touched upon, such as the connection between physical and digital archives and the sustainability of systems. An important topic during the discussion was the issue of openness in the context of archiving - how to allow bottom-up participation from a community of producers and users. Perhaps, here, accuracy shouldn't always be the main goal - noise can be interesting as well.

The question was raised on how archives relate to the communities of artists and to users. Whereas interoperability was defined in a technical sense at the presentations, it is also, at least partially, a cultural phenomenon. In this way the question of open or closed systems returns: must database developers choose between the traditional archival approach or the open repository method? And does the latter offer more flexible innovation for ‘creative systemizers' or does it only confuse the concept of online archiving? The question of sustainability is important in the context of the strictly archival: do categories stick, say for twenty years? Since the categories themselves usually become invisible set in a certain framework, what kind of discourse is developed, say, for an online media history or European history and how might this be represented to the user?

For more free-flowing initiatives on the other hand, such as the runme.org initiative organized by Alexei Shulgin, sustainability becomes a piling of data, creating an archive on the run as it were: it enables users to propose keywords and objects. How bottom-up can these open archives be however? The accuracy of information, the complex and at worst idiosyncratic structure and the question of sustainability are problematic aspects of such free-flowing archives. Perhaps a ‘middle-up approach' might be a solution, selecting users as so-called "fan archivists", who help to sustain the above?  Information from a community could in this way be filtered to the (more official) archive. This does imply a gatekeeper position however, making editorial intervention necessary; something the semantic web is looking to avoid. Another option might be Peer-to-Peer, attracting users who are looking to exchange goods. A fan archivist might well send in work-in-progress on a project or a reaction to a specific event in exchange for archive material, for example. Though building channels for users and a systematics of interoperability will certainly prove an interesting challenge, it was clear that the strictly archival and the open repository need their borderlines.

Because of these very broad discussions, the meeting turned out too short for in-depth treatment of many pressing issues among the attendants; probably a series of consecutive presentations and workshops will be needed to cover the variety of subjects touched upon during "The Art of Open Archives".


This text is licensed under a Creative Commons License. License does not apply to images and comments.