In and out of Tags

The Wittgenstein Archives at the University of Bergen organises a session (90 minutes) to present some experience and new results of its work which it believes will be of broad interest for the text encoding community in general. The session will consist of three papers:

"Quality Assurance In Between Tags"; Maria Sollohub
"Attributes: A Problem", Claus Huitfeldt
"Attributes: A Solution", Peter Cripps
The first paper starts by pointing out that whereas in discussions of text encoding much attention tends to be placed on encoding techniques, code structures and syntax, the task of ensuring the quality of text material contained between code tags is of equal importance for projects making use of text encoding. For example there is a need to spell-check multi-lingual encoded documents while remaining in primary format, to be able to check that particular elements satisfy requirements concerning format or use of specific terminology etc. Using the experience of The Wittgenstein Archives over the last five years as a point of reference, typical problems are described and solutions are suggested. The solutions presented are implemented in the Archives' own encoding scheme, MECS, but need not be MECS-specific. The issues raised are of relevance to all text encoding projects.

The second paper first sets out to understand text structures in terms of relationships between textual objects and their properties. A decision to encode certain properties with generic identifiers and others with attributes implies a distinction between primary and secondary properties which is often useful, but which may also sometimes have undesirable consequences. Finally, certain kinds of properties are identified which seem to require attributes, yet cannot be represented properly with SGML's attribute mechanisms.

The third paper sketches a new and more flexible syntax for the handling of attributes. This mechanism allows for attributes to qualify not only generic identifiers but also other attributes. It distinguishes between two types of attributes, open and closed, and two kinds of attribution, weak and strong. The open/closed distinction belongs to an attribute's definition, whereas the strong/weak distinction refers to two ways in which an attribute may be applied.