Syd Bauman’s research while affiliated with Brown University and other places

This paper explores some of the ramifications of the inherent tension between the desire to analyze our own data and to share it with others — between our desire to be expressive and our desire to be conformant — particularly with respect to interchangeability vs. interoperability , positing that focusing on interoperability is to focus on sharing at the expense of analysis.

The Text Encoding Initiative's typed segment-boundary delimiter method is only one of several proposed mechanisms for handling overlap in TEI documents. HORSE (aka CLIX) defines a method by which an XML element is used normally when possible and as an improved version of the typed segment-boundary delimiter method when an overlap problem is encountered. A significant portion of the rules necessary for validation of HORSE markup can be expressed using Schematron. This, combined with an utter hack that can "HORSEify" the declaration of elements in a TEI Relax NG grammar, can provide a potential significant step forward in handling overlap in TEI documents.

This paper describes the preliminary results of a joint initiative of the TEI (Text Encoding Initiative) Consortium and the ISO Committee TC 37SC 4 (Language Resource management) to provide a standard for the representation and interchange of feature structures. The paper published in the proceedings of this workshop is in fact an extension of a paper published in the LREC 2004 proceedings, and about 50% are identical with it.

The TEI Guidelines provide little detail on how to encode a text within the physical structures of the book in which it is contained. This paper describes the physical structures of an early printed book and presents two methods for encoding a text within that structure through use of the TEI elements and .

This paper discusses the encoding of proper names using the TEI Guidelines, describing the practice of the Women Writers Project at Brown University, and the CELT Project at University College, Cork. We argue that such encoding may be necessary to enable historical and literary research, and that the specific approach taken will depend on the needs of the project and the audience to be served. Because the TEI Guidelines provide a fairly flexible system for the encoding of proper names, we conclude that projects may need to collaborate to determine more specific constraints, to ensure consistency of approach and compatibility of data.

The largest effort in the area of standardisation of computer encoding of language resources has been the Text Encoding Initiative (TEI), established in 1987. TEI chose as its underlying standard SGML (Standard Generalized Markup Language), and in the years before the inception of XML, a number of projects encoded their data according to some SGML DTD, TEI compliant, or otherwise. These projects could now benefit from migrating their data to XML. Apart from validation, the most compelling reason for migration is the scarcity of SGML-aware software and the abundance of XML-based tools and related recommendations. However, despite the fact that XML is a subset of SGML, migration is not a trivial process, especially in the case of large holdings of legacy language resources. This is why in 2002 the TEI Consortium established a Task Force on SGML to XML migration. The TF has now produced a number of reports that simplify and make explicit the conversion of SGML TEI (version P3) to XML TEI (version P4) documents. The reports are also relevant for a general audience of SGML users that are considering migrating their language resources to XML. This paper presents the recommendations made by the TF, concentrating on strategic considerations, the practical guide, and one case study, the conversion of the British National Corpus.

Where should attribute constraints live? In an external schema? In the document’s own metadata? In a separate file? Several possibilities are examined, raising lots of questions and offering a few answers.

Being able to execute an XPath against a corpus of XML documents is powerful. Being able to execute a second XPath from each node selected by the first is even more powerful.