Teaching TEI in Bern

At the invitation of Bénédicte Vauthier from the University of Bern (whose indefatigable enthusiasm, fund raising expertise, and hard work organizing the event, are all hereby gratefully acknowledged) I spent the last week teaching the French strand of a four and a half day long trilingual TEI training session, aimed primarily at people working on the encoding of modern primary manuscript sources, in collaboration with Christof Schoech and Alejandro Bia, who provided the German strand, and Elena Pierazzo who provided the Italian one. We did try a bit to co-ordinate our efforts so that the material covered in each strand was the same, though our methods of teaching, and some of the materials used naturally varied. The course also included a full day of invited presentations about a wide range of Swiss projects and a series of invited plenary presentations. The first day was hosted by the Swiss National Archive who showed us some of their collection, which includes a splendid exhibit of authors’ typewriters. More details of the whole event are, or will be, available elsewhere; there are also a few photos here. But mostly this blog entry just reports on the French teaching strand, and may be of most interest to others undertaking similar quixotic enterprises.

Each of the three strands contained five sessions, each held in a different teaching room (the French one equipped with PCs, the German and Italian ones with Macs) and combining a 45-60 minute presentation with a 60-90 minute scripted hands-on exercise. The programme, also common to all three strands, was as follow:

  1. Introduction to basic ideas of encoding, distinction between document and text, refresher on XML. In the practical session, students were shown how to use oXygen to create a simple document from scratch, using just the TEI- bare schema.
  2. Introduction to TEI, focussing on how to tag basic components of a text, brief overview of some commonly encountered TEI elements, using French examples. Practical session using  TEI Lite as schema. In the first part, students used oXygen to add tags to a plain text version of a sonnet by duBellay; in the second, they used oxGarage to convert a Word file containing a scene from Jarry’s Ubu Roi into TEI, and then used oXygen to improve its tagging.
  3. The TEI Landscape with a slide or two on each of the 23 chapters in the Guidelines, followed by a quick guide to using the TEI website as a means of exploring them. In the practical session, students first use oXygen to create a tei_all file, and then use Roma to create a schema which is better constrained to the encoding of modern mss.
  4. TEI encoding of editorial interventions and simple manuscript transcription. In the practical; students reverse-engineer a rather heavily edited version of the start of a Wilfred Owen manuscript into a more diplomatic TEI transcription in oXygen, using the schema they made in the previous tp (which is also used for the rest of the course).
  5. TEI Header, focussing on its use as a repository for metadata of importance to librarians and archivists, but also to scholars and end-users. In the practical session following students create a full msDesc for a manuscript letter, using oXygen to tag the various parts of a plain text catalogue entry with which they are supplied.
  6. TEI documentary editing, starting with facsimiles and zones, and moving on to include sourceDoc, line, mod, metamark, and other recent additions to the TEI family, notably change and @change. In the practical session, students used Inkscape to identify various zones within a page from a Durenmatt ms, transferred these to a basic TEI transcription of the same file, and then ran an XSLT conversion to combine these into an SVG file in which the transcript and the graphic were merged in such a way that a piece of javascript could display the changes documented in the TEI independently

Archives of all the materials I used for these sessions are available for download in PDF form (Exercices; Talks).  The TEI XML source is also available, under the usual CC-BY licence, from the MEET subversion repository.

In the final session, students were invited to apply their experience to their own materials, samples of which they’d been instructed to bring with them. Most of them set to immediately and started tagging, which was quite encouraging. I was able to do some hand-holding with the one or two students who had somehow failed to understand what was going on earlier, and the others all started banging away in earnest.

Some further comments on the sources I used for each of these talks, and some things I found it necessary to change follow….

  1. The pattern and content for Session 1 is pretty much traditional now. Even so, I managed to find some typos to correct in the exercise script — the keyboard shortcuts for the Windows, Mac, and Linux versions of oXygen are all subtly different, and my script is indecisive as to which it is using. I noted, as always, that it’s hard for beginners to understand that « I want a new line here » really means « I want to close this paragraph and start a new one ».
  2. I put the talk for session 2 together using some old talks introducing TEI Lite. I included a bit of the Punch exercise for old times’ sake, but mostly it derives from a very old talk introducing TEI Lite, spiced up with French examples taken from those in the Guidelines. Some of these are not tagged in quite the way I would have expected, or don’t demonstrate things as clearly as they might, so the whole lot probably needs to be reviewed for consistency. The first part of the exercise for this session went very well, with everyone suitably impressed by the effect of tweaking the CSS to render the document in « Mode Auteur ». The second part was less successful (though not everyone got to it through lack of time) largely because of improvements to oxGarage since I last revised the script; however, I remain convinced that this is the right place to introduce the possibility of starting from a Word document.
  3. The « TEI from chap 1 to chap 23 » talk in session 3 is quite a stretch to do in 40 minutes, especially if one is easily distracted, but remains useful since it motivates the following Roma practical. It’s essential to get across the notion of TEI modularity and this helps students see the point immediately. I based the tutorial on the  script used at this year’s Oxford Summer School, which I translated more or less faithfully, just removing some of James’s jocularities.
  4. The two talks on textual editing (sessions 4 and 6) with TEI were both derived from one I gave in 2011. The initial part, about encoding critical apparatus, seemed a bit redundant (especially for an audience including e.g. Jean-Louis Lebrave)  since we didn’t proceed to use the tags it introduces, and also because I spent too long on it; though it does provide some useful background. In the practical session, both Christof and I decided to use the Wilfred Owen examples used at this year’s Oxford Summer School, though with some  revisions. In the first exercise on transcription,  rather than making sense of an arbitrary non-XML transcription, students had to work out for themselves what was going on in the manuscript image, using the edited transcript provided purely for reference. They found this initially difficult, but once they’d realised that simply cutting and pasting would not be good enough, they became quite enthusiastic. During this exercise I also learned the interesting linguistic fact that the word raturer can be used for any kind of deletion, whereas the word biffer means specifically a deletion carried out by striking through something with a single line. This means that if I revise this exercise again (as I probably will) I shall be providing @rend values in French. (The word « stroked » used throughout the Oxford material seems to me just wrong, by the way. We’re not in the business of cuddling our texts. Au contraire).
  5. Session five’s talk on the TEI Header was more or less unchanged (except for a few typos) from the version I gave earlier this year, itself translated from plenty of other older versions. I think that providing the context for a <msDesc> is useful; the talk also provides an opportunity to talk about named entities, a topic which was otherwise missing in our programme. The practical was again based on one from this year’s Oxford Summer School, though again with some modification — notably that the manuscript described surely contains two <msItem>s, not one.
  6. For session 6, I used the second part of the textual editing talk more or less unchanged: the underlying concepts when defining zones seemed to go well, particularly once I’d admitted to being nul en math. The practical session was an entirely new one, made by translating Christof’s German version, which he produced on the basis of some notes from Elena, describing how she’d presented her Proust prototype at a conference in Australia earlier this year. In the event, one of the nicest moments of the entire course was seeing one of my students punching the air in excitement after having successfully opened the SVG output from this exercise in Internet Explorer. There was a small hiccup when I realised that the version of Inkscape we were using was in German, not English which made configuring it correctly rather tricky, but otherwise this all went far more satisfactorily than I expected. Which was, as they say, nice.