The Ibsen Letters - and Beyond

Øyvind Eide
Centre for Ibsen Studies
University of Oslo
Drammensveien 42
Oslo, N-0242

This paper will explore the possibilities for combining TEI encoding and database approaches in order to present a structure of metadata about letters, and demonstrate how these metadata can be used to integrate information from different sources in order to generate TEI documents "on the fly". We will also show how this system can be extended to include not only letters wrritten by a single person, bu also letter exchanges between different persons.

In January 1998, the Manuscript project at the Centre for Ibsen Studies at the University of Oslo was started. The goals of the project are as follows:

The Manuscript project works in close co-operation with two other projects which are also located at the centre, i.e. The International Ibsen Bibliography[1] and the Henrik Ibsen's Writings[2]. The digitization part of the Manuscript project does not aim at digitizing a particular collection of documents from one archive, but the writings of Henrik Ibsen wherever they are located. The manuscripts for his plays, poems, articles and speeches constitute a relatively small number of items located in 10-15 collections, whereas ownership of approximately 2.700 known letters, telegrams and dedications[3] is spread between 100-150 different libraries, archives, museums and private collectors.

We have developed a digital register of all known letters based on a printed register from 1978.[4] The purpose of the digital register is to:

The use of TEI for metadata is, along with formats like EAD, MARC and Dublin Core, found in several other projects.[5] In the first version of the register, it was a single SGML file using a small subset of TEI Lite. The links to the facsimiles were contained in the register for each letter, but there were no links to transcriptions, as the aim of the project was to make facsimiles, and not transcripts. But, as we will see later, this kind of information is included at a different level.

As in other projects using TEI for metadata, we soon faced a file size problem.[6] We decided to convert the TEI file into a database, but we wanted to continue using the TEI format for delivery. We use PERL scripts to link the database to a HTTP interface. The search page is written in HTML, while the user can choose to have the result of a search delivered as a TEI lite document or as a document containing a HTML table.

This system presents the results of our project. But printed editions of most of the letters exist in electronic forms. It is important for the users to have access to as much source material concerning Ibsen as possible, provided the quality of the material is of an acceptable standard.

Therefore, we made an agreement with the Documentation project[7] to include searches in their database of classical Norwegian literary texts as a part of our presentation. This made it possible to include a link to a text version of each letter, together with the link to the facsimiles.

This system can be described using the library catalogue metaphor: The register database corresponds to the card catalogue. A search in this database will produce "cards" pointing to the online resources. Some questions regarding integration of the different source materials will be discussed in this context.

A main problem with publishing letters as described above is that it does not show clearly in which way the letters are part of an exchange between two persons. Letters to Ibsen from some persons, e.g. the author Bjørnstjerne Bjørnson, are also published in digital form by the Documentation project. We will demonstrate a user interface that shows the two participants in a letter exchange as a TEI document presenting the letters in chronological order, with facsimiles and text when they are available. Some of the difficulties we faced in this work will be discussed.

Most letter exchanges will be part of a vast international web. No single institution, nor single country, can publish a complete set. Possibilities to use our approach for integration of collections and registries of letters will be discussed. A long-term goal is a system where all digitally available letter registers and collections worldwide are available through an integrated search system. This would be an important step in the direction of a digital library with an organization based on content, not on the locations of paper collections.


  1. <>

  2. <>

  3. This number includes letters that we know the existence, but not the whereabouts, of. We believe most of them to be lost.

  4. Henrik Ibsen's brev : kronologisk registrant med adressatregister m.v. / ved Øyvind Anker. - Oslo, 1978.

  5. The William Blake Archive <> uses TEI for manuscript description. See also three papers from DRH'97 <>: "Integrating data and metadata in the digital library: SGML and other approaches" by MacKenzie Smith, Harvard University Libraries; "EAD at the Library of Congress: a progress report" by Lee Ellen Friedland, Library of Congress; and "Linking word and image: two TEI-based imaging projects" by Richard Gartner, Pearson New Media Librarian, Bodleian Library, Oxford.

  6. "SGML solution can obviate many of these problems: [...] It can, however, lead to extremely large documents when the collection is large." MacKenzie Smith, op.cit.

  7. <>