The Text Encoding Initiative: A Theoretical Standard for the Encoding of Electronic Texts

Invited lecture (HTML slide deck) delivered at the SSHRC / University of Saskatchewan Image, Text, Sound and Technology (ITST) Conference, Saskatoon, 15 May 2004.

The talk introduces the Text Encoding Initiative Guidelines to an audience assumed to have a basic familiarity with structural markup, SGML, and XML. It sets out what the TEI is — a modular XML/SGML standard for the scholarly description and exchange of texts, structural rather than presentational, descriptive rather than generative, and universal in intended scope — and what it is not, contrasting format markup with structural markup through a worked bibliographical example. The bulk of the lecture is an assessment of the standard's advantages and disadvantages for humanities scholars. In its favour: a carefully designed textual grammar built by experts across markup, humanities computing, library science, and editorial studies; a well described, consensus-based benchmark against which individual encoding decisions can be argued; an appropriate level of detail for most scholarly purposes; and ease of repurposing. Against it: structures that are complex and hard to standardise, leaving even TEI-encoded texts difficult to exchange between projects without significant revision; an orientation towards the capture rather than the creation of texts, with a largely static scheme poorly suited to describing process; a structural approach that can be non-intuitive and that forces disambiguation where disambiguation is itself contested; and a documentation burden that can make simple print features verbose to encode. The lecture closes with practical guidance on when to adopt the TEI and how deeply to encode, recommending richness over economy where the outcome is uncertain, close adherence to the published standard, and reversible local changes where the standard genuinely does not fit.

This lecture was one of two O'Donnell gave at the same conference on the same day. Slides 5 to 13 are shared verbatim with the companion lecture, “The Electronic Cædmon's Hymn: A Single Scholar, Multiple Text Electronic Edition and Archive”, which is deposited as a separate record; only the title page and slides 1 to 4 are distinct to this talk.

This record contains the original HTML slide deck (zipped, unmodified) and a slide-by-slide PDF rendering of fourteen pages. The PDF was generated by Claude Code from the HTML originals.