The Collected Works of Carl-Maria-von-Weber are being edited by the Carl-Maria-von-Weber Gesamtausgabe and aims at publishing all musical scores, letters, writings and diaries from Weber. The edition of the letters is currently undergoing fundamental reconstruction due to the desire to present this scholarly edition online (the text together with critical annotations). Therefore, a format had to be found and a workflow had to be established that allowed for transcribing, proof reading and online publishing.

First, the workflow will be presented, starting with the collection of the sources and the transcription directly to TEI P5. At present, no special versioning system ist used, so all changes to the transcription are being documented in the revision history of the file itself.

The encoding will be discussed departing from two major problems: the structure of the meta-data and the relation with other material such as databases, newspaper documents, playing schedules etc.

At present the meta-data-area of P5 is lacking some special fields which we need for the description of letters – and which ought to be grouped together: e.g. "addressee", "residence of the addressee", "position of the mentioned letter within the context of letters to the same or to other persons", "answering letter", "first publication of the letter". In this context we aim at a closer association with other running (or finished) projects and the TEI council since these problems arise for other projects as well. Hence a lot of work should be spent on documentary purposes.

The association with other material seems to be – on the surface – just a problem of some xlinks. Digging deeper, we come across all these nasty things as persistent identifiers and typologies which we would like to have (already) standardized but always have to rethink about as it is still work in progress.

In a second part the actual online-edition will be presented. The letters are part of the public website of the Weber-Gesamtausgabe (yet still hidden) which uses the content management system typo3. The index of the letters is built by a directory listing and enables the user to display either the TEI file directly or via HTML – the transformation to HTML is done by some typo3 extensions to neatly integrate the letters into the website-design and to give a more reader-friendly version.