- Schedule, Tuesday, 2 July: TEI customisation, analytic encoding
- Seminar 3: Customisation
- Seminar 4: Transcription and facsimiles of primary documents
- Seminar 5: Analytic and interpretive encoding
Today will feature an introduction to TEI customisation for scholarly editions, a tutorial on connecting image facsimiles, and a workshop on analytic and interpretive encoding.
Schedule, Tuesday, 2 July: TEI customisation, analytic encoding
|9.30||Seminar 3: Customising TEI with Roma and RomaJS [TSG]||Presentation|
|11.30||Seminar 4: Facsimile surfaces and zones [TSG]||Presentation; Practice|
|14.00||Seminar 5: Analytical and interpretive encoding [CO]||Presentation|
|15.00||Exercise 5: Structuring your critical judgments|
|16.00||Library Time||Senate House Library|
- General grasp of using Roma to customise TEI.
- Facility with encoding surfaces and zones on facsimile transcriptions.
- A set of principles for interpretive encoding.
Syd Bauman, “Overview of TEI Customisation”.
Chapter 17 of the TEI Guidelines (for analysis).
Chapter 23 of the TEI Guidelines (for customising the TEI).
Seminar 3: Customisation
Access the slides here.
Notes on TEI and ODD [CO]
In What is the TEI? Lou Burnard asks: “How should you go about choosing just the parts of the TEI you need? How should you communicate the particular TEI encoding choices you have made to others so that such integration remains possible?”
Schema: the document’s grammar. Say you want a rule: every poem should be a collection of
<l> tags with
@n (numbered line attributes), and that each group of
<l>s should be grouped under an
<lg> tag specifying that it is a numbered stanza (
@n again) and a
@type attribute for indicating the length of the stanza (e.g., couplet, tercet, quatrain, &c.).
Or suppose in a documentary editing project, we are encoding letters, and we want to make sure each
<date> tag has a
@when attribute, or that each
<p> has an
@xml:id (a good idea), or that each
<persName> has a
@ref to point to the
@xml:id of the person.
A schema will codify all of those rules such that the xml document will not be valid unless each is followed. The ODD validates the TEI document.
How to customise?
tei_all: Simple, yet not recommended, way is to use the TEI All (or, put another way, the null) schema. This says that you will use all available modules in the TEI.
tei_lite: also, simple; also not recommended (by me, at least): this selects about 50 of the most common elements used in about 90 percent of existing projects.
Both tei_all and tei_lite documents can be selected when you open a new file in oXygen (to do this, select File > New > Framework Templates > TEI P5).
There are three good ways to create schema for your project:
- Relax ng schema (built into TEI files on oXygen).
The TEI has elements such as
<classSpec> (and others) combine formal XML declarations for inclusion in a DTD or Schema with detailed documentation and examples, for inclusion in a technical manual about the encoding scheme being specified. For this reason, a document using these elements is informally known as an ODD
ODD files (One Document Does it all) are TEI customisation files that serve as a source for the derivation of
- a formal TEI schema
- readable documentation of the project’s TEI encoding choices
Now, going back to oXygen, open an ODD template file by choosing New > File > Framework Templates > TEI ODD > ODD Customization. What you see looks like a standard TEI file. It even has an TEI-XML namespace and a
<?xml version="1.0" encoding="UTF-8"?> <TEI xmlns:xi="http://www.w3.org/2001/XInclude" xmlns:svg="http://www.w3.org/2000/svg" xmlns:math="http://www.w3.org/1998/Math/MathML" xmlns="http://www.tei-c.org/ns/1.0"> <teiHeader> <fileDesc> <titleStmt> <title>Title</title> </titleStmt> <publicationStmt> <p>Publication Information</p> </publicationStmt> <sourceDesc> <p>Information about the source</p> </sourceDesc> </fileDesc> </teiHeader>
But what makes the ODD file is the
<schemaSpec> element, which lists elements specifying which TEI modules we want as well as the elements and attribute rules.
<body> <schemaSpec ident="oddex1" start="TEI"> <moduleRef key="header"/> <moduleRef key="core"/> <moduleRef key="tei"/> <moduleRef key="textstructure"/> </schemaSpec> </body> </text>
<moduleRef> refers to a TEI module. There are currently 22 modules in the TEI Guidelines that match to a chapter. So, for example, the “header” module corresponds to Chapter 2 of the TEI Guidelines on the “TEI Header”.
What is happening above is simply a listing of four TEI modules to be included: “header”, “core”, “tei”, and “textstructure”. These are all required for TEI conformance, which is why they are in the ODD template file. Other optional modules include:
<moduleRef key="namesdates"/> <moduleRef key="transcr"/> <moduleRef key="analysis"/>
A module element makes available all the declarations allowed within the schema. Including the the
@except attributes will change this default structure.
So suppose you choose the optional module
<moduleRef key="analysis"/> but you only want to use certain elements. You would check for the available elements in the module by going to the appropriate chapter in the Guidelines, which happens to be Chapter 17. The end of the chapter includes a helpful module analysis, which indicates:
Module analysis: Simple analytic mechanisms Elements defined: c cl interp interpGrp m pc phr s span spanGrp w Classes defined: att.global.analytic att.linguistic
This tells you what elements are available and what class these belong to. But if we only want to use phrase- and word-level elements (
<s>, <phr>, and <w>) in our document (and none of the interpretive stuff), we could write a
<moduleRef key="analysis" include="s phr w"/>
Or, if you’re really clear about the elements you want, exclude what you do not want:
<moduleRef key="analysis" exclude="c cl interp m pc span spanGrp"/>
Burnard notes a very good example of how the EpiDoc Guidelines in TEI constrained their encoding with ODD. The EpiDoc Guidelines wanted to constrain the
@type attributes in
<div> elements. And for very good reason: EpiDoc is the standardised set of guidelines for encoding classical documents which require a controlled vocabulary. So within their
<schemaSpec> you will notice:
<elementSpec ident="div" mode="change" module="textstructure"> <attList> <attDef ident="type" mode="replace" usage="req"> <valList type="closed"> <valItem ident="apparatus"> <desc>to contain apparatus criticus or textual notes</desc> </valItem> <valItem ident="bibliography"> <desc>to contain bibliographical information, previous publications, etc.</desc> </valItem> <valItem ident="commentary"> <desc>to contain all editorial commentary, historical/prosopographical discussion, etc.</desc> </valItem> <valItem ident="edition"> <desc>to contain the text of the edition itself; may include multiple text-parts</desc> </valItem> <valItem ident="textpart"> <desc>used to divide a div[type=edition] into multiple parts (fragments, columns, faces, etc.)</desc> </valItem> <valItem ident="translation"> <desc>to contain a translation of the text into one or more modern languages</desc> </valItem> </valList> </attDef> </attList> </elementSpec>
One could also add a new element, which you could do today by just adding an
<elementSpec>, but you also need to consider whether the non-TEI elements belongs in a certain class. You need to think about content models and how other elements are defined within those models.
Burnard: “one of the purposes of the TEI Guidelines is to guide encoding practice … it tells you how to communicate what you have done to others.”
Why are we doing this? Aren’t we jumping ahead a bit on the TEI?
Answer: Possibly, but the TEI is very difficult to learn in a few days (in a few years, even), and with ODD you engage with the Guidelines with a critical eye. Moreover, it is inevitable that projects will decide to encode different things in different ways. For example, some projects might consider it very important to label each rhyme word within a poem’s lines with
@label, and they will want to include
@rhyme attributes in their supervening
<!-- from Alexander Pope's *Essay on Criticism* --> <lg type="couplet" rhyme="ab"> <l>Speak silent when you Doubt your <rhyme label="a">Sense,</rhyme></l> <l>And speak, tho' Sure, with <hi rend="italic">Seeming</hi> <rhyme label="b">Diffidence.</rhyme></l> </lg>
Suppose I want a new element that is not available in the TEI? Like, say, an
<alliteration> element for the poem above?
<elementSpec ident=”alliteration” mode=”add”> <classes> <memberOf key=”att.interpLike”/> </classes> </elementSpec>
You will now be able to have a valid
<alliteration> element that you have added to an “interpLike” class, which is to say interpretive class.
Other poetry projects will not use any rhyming identifiers, as they may simply want to record, say, textual features such as variant readings. Burnard also gives a good example of the various ways editors will encode person names.
The ODD approach allows you to communicate exactly what encoding decisions you made, how you constrained them for consistency, and how you envision what is important in your document model. This benefits other researchers doing similar projects as well as the TEI, which is community-driven.
The TEI also offers a web application called Roma to build ODDs from the element and attribute specifications covered above.
For more information on writing ODDs, consult Syd Bauman and Julia Flanders’s teaching materials.
Seminar 4: Transcription and facsimiles of primary documents
Access the slides here.
Access the exercise file here.
Click here for the spoiler file.
Seminar 5: Analytic and interpretive encoding
Access the slides here.
Here is a detailed (and in-progress) example (courtesy Steven Olsen-Smith and William Newmiller’s course exercise at the US Air Force Academy) infrastructure of interpretive identifiers in the
<back> element of Stephen Crane’s Black Riders.
<back> <div type="interpretations"> <p> <interpGrp type="theme"> <interp xml:id="theme-lifedeath">life and death</interp> <interp xml:id="theme-religion">god and religion</interp> <interp xml:id="theme-mind">psychology and mind</interp> <interp xml:id="theme-humanrelations">human relations</interp> <interp xml:id="theme-other">other</interp> </interpGrp> <interpGrp type="figureofspeech"> <interp xml:id="fig-symbol">symbol</interp> <interp xml:id="fig-image">image</interp> </interpGrp> <interpGrp type="reference"> <interp xml:id="ref-object">object</interp> <interp xml:id="ref-human">human</interp> <interp xml:id="ref-deity">deity</interp> <interp xml:id="ref-animal">animal</interp> <interp xml:id="ref-concept">concept</interp> </interpGrp> <interpGrp type="description"> <interp xml:id="desc-sound">sound</interp> <interp xml:id="desc-color">color</interp> <interp xml:id="desc-condition">condition</interp> </interpGrp></p> </div> </back>
If you consult an encoded poem, notice how we use linking mechanisms to point to the
@xml:ids of various interpretive concepts.
<div2 type="poem" ana="#theme-lifedeath"><head>II</head> <ab> <lb/><s>Three <w ana="#desc-condition">little</w> birds in a row <lb/>Sat <w ana="#desc-sound">musing</w>.</s> <lb/><s>A <w ana="#ref-human">man</w> passed near that place.</s> <lb/><s>Then did the little <w ana="#ref-animal">birds</w> nudge each other.</s> <lb/><s>They said, "He thinks he can sing."</s> <lb/><s>They threw back their heads to laugh, <lb/>With <phr>quaint countenances</phr> <lb/>They regarded him.</s> <lb/><s>They were very curious, <lb/>Those three little birds in a row.</s> </ab> </div2>
- Consult either the Bunting notebook, the Moby-Dick xml file, or your own project file.
- Identify some interpretive or analytic aspects of a section of the text.
- Encode some linguistic elements to identify syntactic structures.
- Create a new
<back>that includes the ids and a list of your interpretive concepts.
- Encode the interpretive aspects in the text transcription and link them to your ids in the
Recall how yesterday, you included a “metadata” folder within your project directory. That folder includes various authority files: personography data, object data, and interpretive data, and so on. To use stand-off markup, enter the same pointing/linking technique in the text file but instead link to an external link with a filepath:
<s>Three <w ana="../interpretations.xml#desc-condition">little</w> birds in a row <lb/>Sat <w ana="/interpretations.xml#desc-sound">musing</w>.</s>
This means that you are storing interpretive elements within the
<body> of a separate XML file in your metadata sub-directory; and the XML above just calls upon the
@xml:id of a given
<TEI xmlns="http://www.tei-c.org/ns/1.0"> <teiHeader> <fileDesc> <titleStmt> <title>Interpretative Entities</title> </titleStmt> <publicationStmt> <p>Project Authority List of Interpretive elements.</p> </publicationStmt> <sourceDesc> <p>Information about the source</p> </sourceDesc> </fileDesc> </teiHeader> <text> <body> <div type="interpretive"> <interpGrp> <interp xml:id="revenge">Revenge, mostly dealing with Ahab's rage.</interp> <interp xml:id="diffidence">Diffidence, usually being the uncertainty or mistrust of oneself in the ship-mates.</interp> <interp xml:id="metaphor-animal">Comparisons to animals.</interp> </interpGrp> </div> </body> </text> </TEI>
You could of course do the same for personography data (many do, in fact). It might be wise to condense your external references by creating abbreviated pointers with
@key URIs. If you are working with several documents (i.e., individual files) that are calling upon the same authority lists of people, places, objects, interpretations, etc., then you would want to store them separately and use the stand-off approach in each file. There are more complicated approaches to hosting other structured data externally using
<xi:include> elements, which are explained in Chapter 16 of the TEI Guidelines.
The Guidelines also make the following helpful distinctions:
- source document (to which the stand-off markup refers, in XML or plain text)
- internal markup (pointing to tags within the text)
- stand-off markup (either outside of the source document and pointing in to it the data it describes, or alternatively is in another part of the source document and points elsewhere within the document to the data it describes)
- external document (containing stand-off markup that points to a different, source document)
- internalize (creating a new XML document with external markup and data integrated with the source document data, and possibly some source document markup as well)
- externalize is a process applied to markup from a pre-existing XML document, which splits it into two documents, an XML (external) document containing some of the markup of the original document, and another (source) XML document containing whatever text content and markup has not been extracted into the stand-off document; if all markup has been externalized from a document, the new source may be a plain text document.
Proceed to Day 3.