Encoding for ELTeC : a summary guide

The checklist

Here's a list of the distinctions your encoding should make explicit:

page breaks (<pb>) [optional]
section and chapter divisions (<div>)
headings, titles, footers (<head>,<trailer>)
paragraphs (<p>)
verse (<l>)
typographic salience (<hi>) [optional]

A level 1 encoding permits the following additional distinctions, all optional but recommended:

omitted material(<gap>)
editorial correction (<corr>)
chapter subdivision (<milestone>)
linguistic emphasis (<emph>)
code-shifting (<foreign>)
floating headings and titles (<label>, <title>)
quotation (<quote>)
authorial notes (<note>, <ref>)

A level 2 encoding may additionally contain:

<s> elements for segmentation within paragraphs
<rs> for named entities (type values to be determined)
word level tokenization using <w> elements which bear attributes pos, lem, join, msd (values all to be determined)

A note on page breaks

ELTeC transcription need not capture any aspect of a book's original appearance, including its pagination.
However,
- identifying a page number is useful when checking the encoding against an original
- identifying all page breaks can also facilitate some kinds of analysis.
Use <pb/> to mark the start of each new page or <pb n="x"/> for the page numbered x in the source.
If a word is hyphenated across a page break, re-assemble it on the page where it begins
Discard any catchword, running page header or footer, etc.

The text itself

Find the first and last chapters of the novel: everything before this is the <front>, everything after is the <back>, everything in between is the <body>.
Discard any non-authorial sections from front or back
Use <div>, with an appropriate type to mark remaining sections
- Use <div type='titlepage'> for the titlepage
- Use <div type='liminal'> for any other subdivision of front or back
- Use <div type='chapter'> for chapters
- Use <div type='group'> for groupings of chapters
- Use <div type='letter'> for epistolary novels
- Use <div type='notes'> within the back for notes

Headings and footers

Use <head> for each heading or title provided at the beginning of a section, and <trailer> for each provided at the end
Discard running titles or page headers

<div>...<p>Le président fit un signe et la musique de la garde républicaine joua la Marseillaise.</p> </div> <div type="chapter"> <head>II</head> <head>LE MARQUIS DE LA LANDE</head> <p>Dans la coquette garçonnière des Champs-Elysées...</p> </div>

Prose and verse

At level 0, the element <p> is used for any stretch of prose text followed by a significant line break. It is usually, but not necessarily, a complete syntactic unit.

The element <l> should be used for each distinct line of verse, where this is clearly not prose, as in for example a song or an epigraph.

<p>Eraû ceasurile 10. Un beţiv trecênd pe stradă cânta a lene: </p> <l>Vin drăguţă lângă mine: </l> <l>Drăguţă, drăguţă ...</l> <l>Căcĭc mi-e dor, mi-e dor de tine ...</l>  <p>Si românul, beţiv, ca si fanaraiotul păgân...</p>

At level 1, quoted paragraphs or quoted verse may be wrapped in a <quote> element.

<p>Eraû ceasurile 10. Un beţiv trecênd pe stradă cânta a lene: </p> <quote> <l>Vin drăguţă lângă mine: </l> <l>Drăguţă, drăguţă ...</l> <l>Căcĭc mi-e dor, mi-e dor de tine ...</l> </quote>  <p>Si românul, beţiv, ca si fanaraiotul păgân...</p>

Do not however use <quote> for quoted phrases within a paragraph: simply retain the quotation marks (similarly, for direct speech)
A citation e.g. giving the author of a quotation can be marked using <label> or as a free-standing <p>

Typographic salience

An ELTeC transcription need not capture any aspect of a book's original appearance, including its typography.
However, it is often useful to distinguish (e.g.) italicized words or phrases because this usage may have a semantic implication.
At level 0, we record simply that such words are typographically salient, using the <hi> element with no attributes
At level 1, we may instead capture the semantic implication, using e.g. <emph> or <foreign>.

For example...

Omissions and corrections

At level 0, material which has been omitted from a transcription is simply omitted, and any editorial corrections to the text are silently applied.

At level 1, we record the fact that something has been omitted using the <gap> element, and the fact that some text has been corrected using the <corr> element.

<gap desc="graphic"/>

<p>... and so <corr>indubitably</corr> preferable .. </p>

The same element may be used to mark where something missing in the source (eg opening quotation mark) has been supplied by the encoder:

<div> <head>I. Dika Lombardije.</head> <p n="SL-WIKI001351"> <corr>„</corr>Italija, kako lepa si, prostrana naša domovina! Kamorkoli se ozre oko, odpiraš mu nova čuda....dika Lombardije.“ — Tako je sanjaril sam s seboj pesnik Antonio Loschi...</p> </div>

Sub-divisions of chapters

Do not use <div> for subdivisions of chapters: instead mark the point at which the subdivision is indicated with a <milestone> element, using its unit attribute to indicate the kind of subdivision concerned, and (optionally) its rend attribute to indicate the appearance of the milestone.

At level 0, a milestone indication such as a row of stars may be retained as a <p>, but this should be changed at level 1.

At level 0:

<p>... stern-looking woman.</p> <p>* * * *</p> <p>Months had passed by .... </p>

At level 1:

<p>...stern-looking woman.</p> <milestone unit="subchapter" rend="stars"/> <p>Months had passed by... </p>

Groups of chapters

Use <div type="group"> for any collection of chapters which has its own heading, whatever it's called.

<body> <div type="group"> <head>Erstes Buch</head> <div type="chapter"> <head>1</head> <p>Ostrau ist eine kleine Kreisstadt unweit der Oder, bis nach Polen hinein berühmt... </p> </div> <div type="chapter"> <head>2</head> <p>Schon welkte das frischgemähete Wiesengras in der Mittagssonne... </p> </div> </div> </body>

Groups may nest as required. But take care that the structure is properly tesselated: the following, though valid, is wrong:

<body> <div type="group"> <head>Erstes Buch</head> <div type="chapter"> <head>1</head> <p>Ostrau ist eine kleine Kreisstadt unweit der Oder, bis nach Polen hinein berühmt... </p> </div> </div> <div type="chapter"> <head>2</head> <p>Schon welkte das frischgemähete Wiesengras in der Mittagssonne... </p> </div> </body>

Emphasis and code switching

At level 1, emphasized words in speech or narrative may be marked using <emph>.

You said <emph>what</emph>?

At level 1, code switching (use of a language other than that of the main narrative) may be marked using <foreign>

<p> -- Ite! Un <foreign>miş-maş</foreign>, zise Kiva, care intrand auzise spusele Zamfirei.</p>

The xml:lang attribute may optionally be used to specify the language switched to, but is not obligatory

Floating headings and titles

At level 1, headings or labels not attached to a division should be marked using <label>

<div type="chapter"> ... <p>Yo más le tengo por un aventurero impetuoso, y no me explico cómo no triunfó en España».</p> <milestone unit="section"/> <label>11 de marzo...</label> <p>«...Apenas se sabe ya de él. Mi sobrino no pudo tenerlo en su casa. </p> .... </div>

At level 1, titles of songs, literary works, etc. may optionally be marked using <title>.

<p>—Yo en tu lugar hubiera publicado en el <title>Diario de avisos</title> e sa carta para escarmiento de coquetas; pero aún estamos á tiempo.</p>

Authorial notes

At level 0, all notes are omitted. At level 1 :

Only authorial notes should be retained. Editorial or textual notes should be silently omitted
The content of the authorial note should be marked using <note>
The siglum indicating the point of attachment for the note within the text should be marked using <ref>
All <note> elements should be given a unique identifier (e.g. use the text identifier followed by "N" and a number) as value for their xml:id
All <ref> elements should indicate the note they refer to by means of their target attribute
All <note> elements should be collected together in the <back> element, inside a <div type="notes">, no matter where they appear in the source

<body> <div type="liminal"> <p>... a little easy cant about Sensation <ref target="#ENG18630_N1"/> Novelists; but in reality ...</p> </div> </body> <back> <div type="notes"> <note xml:id="ENG18630_N1">This slang term is not quite accurate as applied to me... </note> </div> </back>