1 Preface

Mark Olsen
ARTFL Database
University of Chicago
1050 E. 59th St.
Chicago, Illinois 60637

A few months ago, when the TEI published its Guidelines, and you saw the 1300 pages, and hefted the seven pounds, of the Guidelines for Electronic Text Encoding and Interchange, you wrote me, as you may remember, words to the effect: "Your bricks landed on my desk today. Is there a Cliff Notes version? A bare-bones TEI, without any of the optional stuff, just the absolute minimal TEI encoding scheme?"

This is my attempt to provide you what you asked for --- but only half of my effort to provide you with what I think you really need. The half I don't think you were asking for, though other people have, is a sort of `Pocket-sized Guide to the TEI': a version of the TEI encoding scheme which is small enough to be understood without too much trouble, but large enough to do reasonably serious work with, and powerful enough to suffice for most people's work encoding electronic text, most of the time. Lou and I have discussed this at some length, and a `pocket-sized TEI' (aka `TEI Lite') is now documented in a little paper called "An Introduction to TEI Tagging" (document number TEI U5).

You didn't ask for a TEI Lite, though: you asked for something even smaller and more austere: you asked me to isolate the absolute minimum set of TEI tags, without which it's difficult to imagine making any useful electronic text nowadays at all. That is what I have done in this document.

Note, however, that what you have in your hands is emphatically not an attempt at a realistic markup scheme for real use in encoding new texts. It is a definition of a toy markup language: the absolute minimum is not necessarily a useful minimum. In particular, although this tag set may conceivably suffice for the translation of ARTFL texts and other pre-existing data into TEI form, still I think that when you set about creating new electronic texts, you would be crazy to limit yourself to the textual features listed here, and I hope that, despite your well publicized antipathy to any rational scheme of text markup, and despite the ample measure of craziness which your friends all know and treasure in you, you won't do such a silly thing.

The tag set defined here is simple enough that you should be able to get familiar with it in half an hour, become proficient in it in an afternoon or so, and outgrow it completely in a day or a week or two. And it is a clean subset of the full TEI encoding scheme, so that when you do outgrow this bare-bones tag set, and start (as I hope) looking at TEI Lite and the full TEI markup language, you will already have a firm grasp of the basics of TEI encoding, and can easily integrate the additional tags into the mental framework you built while assimilating this bare-bones TEI scheme. In order to encourage you, and other readers who share your predilection for craziness, to move eventually to the full TEI markup scheme, I mention periodically in this outline the tags in the TEI header and the TEI core tag set which are not included here, so that you will know what you're missing.

I have to confess that Lou is skeptical about the definition of this bare-bones TEI subset. Like me, he thinks that it won't be useful for serious encoding of real data, but he disagrees with my belief that it may nevertheless be useful to those encountering the TEI for the first time. I hope it will be useful, by (a) reducing the clutter so you can see the basic outlines of the TEI scheme more clearly --- the tags included here are the ones everyone is going to need to use --- and (b) demonstrating, by a reductio ad absurdum, how reducing a tag set to this size (it's about the same size as the first version of HTML) forces one to omit too much material which can be useful in the encoding of virtually any text, and which is absolutely essential for dealing rationally with some texts. Lou thinks I am dreaming; time will tell.

So: here is the bare-bones TEI subset you asked for --- may you read it in good health, and may it prove useful in showing you how to translate your existing data into TEI form, and extending your existing software to handle TEI data. (N.B. maledictions will rain on your head if you implement support for this subset but not for the full TEI DTD. And what's more, you'll deserve every malediction in the book.) Use it to encode some simple exercises in SGML and TEI tagging. A few exercises should suffice to persuade you that you'll need a larger scheme (e.g. the full TEI scheme) for serious encoding of texts you hope anyone will work with. Use TEI Lite instead of standard TEI if you must, but don't limit yourself to the skeletal tag set (perhaps I should say, cadaverous tag set) sketched here. Even you aren't that crazy.

Best regards,

Michael