Encode a Primary Source Transcription


As an encoder working on a primary source document, your main job is to represent the original source document as faithfully as possible. In other words, you are classifying different components of the document (front matter, body, back matter, title page, chapter, etc.), and describing how things appear (small-caps, italic, centre-aligned, and so on). The overriding concern here is to tell the truth. Transcribe the source text, tag it to represent its structure and components, and describe how it appears using CSS properties and values in the @style attribute.
If you are encoding a primary source transcription from EEBO-TCP, refer to the instructions for preparing an EEBO transcription.
If you are encoding a semi-diplomatic primary source transcription for the MoEML Library, refer to the instructions for encoding a library text.


Encode your primary source text in six passes. Generally, you want to do the block-level elements first and then proceed to the inline elements (see w3schools.com). Most of our encoders check the transcription while they are encoding the block-level elements. While some variation in work flow is acceptable, you always want to anticipate the possibility (especially with long documents) that someone else may have to pick up where you have left off. Proceed as follows with primary documents:
  1. Add mark-up to identify what Jerome McGann calls the linguistic codes of the work (McGann 13). These codes include the content divisions in a work (title page, dedicatory epistle, introduction, books, chapters, sections, tables, indices, colophon), the basic organizational blocks within those divisions (paragraphs, stanzas, lines), and illustrations. Open, rename, and save the appropriate template that will guide you through this first pass. In this pass, you will encode the <teiHeader>, a process described in the document Encode the <teiHeader> in MoEML Files. Mark up the literary codes in the <text> element as described in the template and below.
  2. Check transcription, make corrections, and supply gaps.
  3. Add mark-up to identify what McGann calls the bibliographical codes of the work (McGann 13). These codes—functions of the way the printed book is produced—include page beginnings, line beginnings, running titles, hyphens, catchwords, and signature numbers. We describe typographical features (italic letters, size of font, dropped caps) using the @style attribute on the appropriate element. If there is no appropriate element, we use the <hi> element with a @style attribute (see more below). If the work you are encoding is light on toponyms and names, you might want to add the MoEML place and name mark-up before you add the bibliographical mark-up.
  4. Add mark-up to identify those features of the work that pertain to the MoEML mission. Those features include names of places (toponyms) and of people.
  5. If applicable, add musical notation. Though there are no documents in our current collection that feature musical notation, MoEML plans to add musical notation to future documents.
  6. Add mark-ups to conserve the styling of the primary source. In primary source transcriptions, we aim to conserve the original styling of the early modern printed book or manuscript being transcribed to the extent possible given Cascading Style Sheet (CSS) and browser standards.

Pass 1: Basic Document Structure

The primary division in your text is between the <teiHeader> and the <text>. In the <teiHeader>, you include all the document metadata (information about the document). In the <text>, you include the entire transcription of the document itself.
The <text> element is usually divided into three components:
      <!-- Front matter, such as title page, dedications, etc. -->
    <div type="chapter">
      <!-- The body of the document. -->
    <div type="colophon">
      <!-- Back matter including indexes, appendices, etc. -->
We will look at each of these in turn.

Encode the Front Matter

The front matter in a work includes the title page and other preliminaries. If the work you are encoding includes a dedicatory epistle, a letter to the reader, an introduction, a table, and/or a frontispiece, these paratexts belong in the <head> element. We will work through the encoding of such front matter.

Encode a Title Page

For our front matter, we’ll use the encoding of a title page in the short document called The Cold Year. You can see the original page on the EEBO site.
In our first pass through the text, we transcribe the text and tag the components of the title page. We use the <titlePage> tag to enclose everything. This is a specialized tag with a small number of child elements. (See the TEI documentation on <titlePage> for more information.) Here is our first pass:
  <pb facs="http://eebo.chadwyck.com/fetchimage?vid=21023&page=1&width=1200"/>
      <titlePart>THE COLD YEAR.</titlePart>
      <date when-custom="1615" calendar="mol:julianSic" datingMethod="mol:julianSic">1615</date>
      <titlePart type="sub">A deepe Snow: In which <lb/>Men and Carrell haue perished, <lb/>To the generall loſſe of Farmers, Graſiers, Huſ<lb type="hyphenInWord"/>bandmen, and all ſorts of people in the Coun<lb type="hyphenInWord"/> trie, and no leſſe hurtfull to Citizens.</titlePart>
      <titlePart type="desc">Written Dialogue-wife, in a plaine familiar talke betweene a <lb/>London Shop-keeper, and a North-Country man. <lb/>In which, the Reader ſhall finde many <lb/>thinges for his profit.</titlePart>
      <figDesc>Woodcut illustration of citizens and animals in the snow.</figDesc>
    <docImprint>Imprinted at London by W.W. for Thomas Langley <lb/>in Iuie lane where they are to be ſold.<date when-custom="1615" calendar="mol:julianSic" datingMethod="mol:julianSic">1615</date>.</docImprint>
First, note that the <front> element begins with a page beginning, and the <pb> tag points to the EEBO page-image using our special URI scheme with the prefix moleebo. This is documented elsewhere. After the page beginning comes the titlePage element, and this is the only other component of the <front>.
Inside <titlePage> are these elements:
  • <docTitle>
  • <docDate>
  • <figure>
  • <docImprint>
In any given title page, these may occur in almost any order, although it’s most likely that the titles will appear near the top, and the imprint information near the bottom. There are many other elements that may appear in a title page, including <imprimatur>, <epigraph>, <byline> and <docAuthor>, but these do not happen to appear here.
Notice that inside the <docTitle>, there are <titlePart> elements. Most title pages (especially from this period) have a multiplicity of title components, which we can distinguish by means of the @type attribute, which can take the values "alt", "desc", "main", "part" or "sub". Also note that <lb> tags are necessary both within and between <titlePart> elements; because different types of title may share the same line, these cannot be block elements, so line beginnings must be explicit.

Encode the Main Text in the Work

The main text of the work goes in the <body> element within <text>.
We use the <div> element to identify the structural divisions within the main text. A <div> can have a @type attribute. It can also have an @xml:id attribute.
When you are encoding transcriptions of primary sources, give the <div> an @type. Most of the time, the transcriber or project director will give you the value for the @type attribute. Some of these values will be recognizable to you as the building blocks of the genre you happen to be encoding (chapter, book). Some of these values are custom mol:values that we have created (e.g., for the mayoral shows).
<div type="dedication" xml:id="stow_1598_dedication">
  <head>TO THE RIGHT Honorable, the Lord Mayor of the <ref target="mol:LOND5">Citie of London</ref>, to the communaltie, and <hi>Citizens of the ſame, <name ref="mol:STOW6">Iohn Stow</name> Citizen, wiſheth long health and felicitie</hi>.</head>
In addition to <div> elements, all text within the <body> element should be properly encoded using TEI tags. Groups of prose should be tagged using the <p> element. Verse should be tagged using <lg> element for the group of lines, with the <l> element wrapped around each line. In certain cases, it may be necessary to use Cascading Style Language [CSS] to describe uniquely styled phrases or clauses.

Encode the Back Matter

Most of the primary works in the MoEML library do not have back matter. Use the <back> element only for postscripts, tables of contents, colophons, and material that is clearly secondary to the main text of the work.

Pass 2: Interpolations and Supplied Characters (Gaps)


When transcribing, editors may come across illegible characters for various reasons. For example:
  • the page has been cropped or the leaf is damaged;
  • the film, scan, photo, or facsimile is unclear;
  • there is ink bleedthrough from the type on the other side of leaf;
  • there is un-inked or over-inked type.
These missing characters need to be represented in our transcriptions if possible. Attempt to decipher them and mark them up with the <supplied> element using the @resp, @evidence, @source and @reason attributes. For the @resp attribute, give your MoEML @xml:id. For the @reason attribute, choose a value from the following table that explains why the text was illegible:
Value Explanation
"bleedthrough" Heavy type on reverse side of page obscures text.
"broken-type" Type appears malformed or fractured.
"damage" There is dirt on the page, tearing, etc.
"error-in-original" Illegibility resulting from misprint/typesetting error.
"faded-ink" Illegibility resulting from faded ink on the page.
"gap-in-inking" Type not sufficiently inked.
"gap-in-transcription" Gap in transcribed material. Use only when facsimiles are unavailable.
"ink-smudged" Smudging dating from the original print process.
"lost-folio" Gap resulting from missing sheet of paper.
"omitted-in-original" Gap resulting from typsetting error (i.e., no type where there should be).
"original-cropped" Original page has been cut in a way that removes part of the text.
"scan-cropped" Facsimile photograph does not include the whole page.
"scan-unclear" Facsimile photograph is not clear, out-of-focus, etc.
"unclear" Fallback value to be used if there is not a more precise value.
Once you have chosen an appropriate value for the @reason attribute choose an appropriate value for the @evidence attribute: if your interpolation is based purely on context, then choose "internal"; if it is based on an external source such as another edition, then choose "external", and use the @source attribute to supply a mol:uri link to the @xml:id for that source in our bibliography.
Finally, add an editorial note with a prose explanation, including another @resp element identifying you as the author of the note.
Here is an example of how to properly use the <supplied> element:
<p>Expecting <supplied resp="mol:JENS1" reason="gap-in-inking" evidence="internal">e<note type="editorial" resp="mol:JENS1">Gap in inking: missing letter obvious from context.</note></supplied>uery day</p>
If a page has multiple missing characters that are caused by the same reason (such as a torn page or cropped scan) then add an editorial note to the first <supplied> element on the page and leave the following <supplied> elements without editorial notes. Here is an example of the editorial note that you would place inside of the page’s first <supplied> element:
<p>we were <supplied reason="damage" evidence="external" source="mol:EEBO2" resp="mol:ROTH4">ea<note type="editorial" resp="mol:ROTH4">Torn page: all missing characters on this page are supplied from EEBO copy.</note></supplied>ſed at the beſt</p>
The transcribers for EEBO-TCP have been cautious and have left gaps where they were uncertain of a reading. Those transcribers are not early modernists; we can often supply the gaps simply by consulting the page images and exercising our better knowledge of early modern texts. We can also consult other editions of a text to see the choices other editors have made. When we supply text that has been omitted from an EEBO-TCP transcription but is clear enough in the page-image to be transcribed without further investigation, we do not use the <supplied> element. We are not doing an edition of the EEBO-TCP transcription, so their omissions due to error or caution are not pertinent.
However, sometimes even MoEML editors cannot decipher missing characters. In this case, the self-closing <gap> element should be used with the attribute @reason. This @reason attribute has all the same values as the @reason attribute for <supplied>.
For example:
<p>Who ſ<gap reason="faded-ink"/>ted with my dull variety</p>


Editors may also encounter misprinted text in primary source documents. For example, the following passage from Thomas Middleton’s The Triumphs of Honour and Industry in which her in the text proper is misprinted as het: On the toppe of the Caſtle, Honor manifeſted by a faire Starre in his hand, Religion with a Temple on het head, Piety with an Altar, Commiſeration with a melting or burning Heart (sig. C1r; emphasis added).
When encoding this passage, we must mark up and, if possible, fix this misprinting. Tag the misprinted text using the <sic> element. If you can confidently interpolate the misprinted text, nest the tagged text string inside a second <choice> tag. Within this <choice> tag, add your interpolation and tag using the <corr> element with a @resp attribute. The value of this @resp attribute should be your "xml:id". The passage would therefore be marked up as such:
<p>On the toppe of the Caſtle, Honor manifeſted by a faire Starre in his hand, Religion with a Temple on <choice><sic>het</sic><corr resp="mol:VIRA1">her</corr></choice> head, Piety with an Altar, Commiſeration with a melting or burning Heart.</p>
Suppose, however, that the encoder could not confidently interpolate the misprinted text. In such an instance, they would simply tag the misprinted text using the <sic> element:
<p>On the toppe of the Caſtle, Honor manifeſted by a faire Starre in his hand, Religion with a Temple on <sic>het</sic> head, Piety with an Altar, Commiſeration with a melting or burning Heart.</p>

Pass 3: Encode Bibliographical Codes

Once you have encoded the basic structure of the work and its linguistic codes, then you will go back and add mark-up to indicate how the text is disposed in the physical space of the book. Your mark-up tells the truth about the material book that is the documentary witness for our edition.

Encode Line Beginnings and Hyphens

Use a self-closing <lb> element to indicate where line beginnings occur in the front matter, main text, and back matter of a work. For example, consider the following address from the dedicatory epistle of Thomas Dekker’s Troia-Nova Triumphans, or London Triumphing:

To the Deſeruer of all thoſe Honors,
which the Cuſtomary Rites of this Day,
And the generall Loue of this City beſtow vpon
him; Sir Iohn Svvinerton, Knight, Lord
Maior of the renowmed City
of London.
(sig. A2v)
To transcribe the lineation of this passage in XML, we must include a self-closing <lb> element after every line of text:
<p><lb/>To the Deſeruer of all thoſe Honors,<lb/> which the Cuſtomary Rites of this Day,<lb/> And the generall Loue of this City beſtow vpon<lb/> him; Sir Iohn Svvinerton, Knight, Lord<lb/> Maior of the renowmed City<lb/> of London.</p>
Line beginnings in early modern primary sources often occur in the middle of words. In such instances, the line beginning is preceded by a hyphen, which signifies to the reader that the word continues on the next line. Because semantically significant hyphens also appear in early modern texts, as in compound adjectives and hyphenated nouns, we must determine whether the hyphens at the end of lines are compositorial (added by the compositor to indicate that the word continues on the next compositorial line) or semantic (intrinsic to a compound word).
End-of-line hyphens present four different encoding scenarios:
  1. A single, non-hyphenated word split over two compositorial lines without a compositorial hyphen:
    In this case, higher should have a hyphen to indicate that the line beginning does not also indicate a space between two different words. Normally, a line beginning without a hyphen functions semantically as a space, so we need to designate for the processor that the encoded line beginning should not function as a space. We make such a designation by using the @rend attribute with "hidden" value.
    May pole high<lb rend="hidden"/>er than the
  2. Semantic hyphen that the compositor does set with the punctuation piece:
    Note here the difficulty in distinguishing compositorial from semantic hyphens. One could argue that the word with the hyphen is really one word (i.e., Blackfryers) that happens to be broken between two compositorial lines, in which case the hyphen would be compositorial. However, because instances of hyphenated Blacke-Fryers unbroken by a line beginning appear elsewhere in the text, we conclude that this is a semantic rather than compositorial hyphen, and therefore the hyphen character should be encoded as follows:
    The <ref target="mol:BLAC8">Blacke-<lb/>Fryers</ref>ref called
  3. Compositorial suppression of a semantic hyphen that would normally be part of the word. In this case, encode a simple line beginning with no hyphens and attributes.
  4. Compositorial hyphen that serves only to split a single, non-hyphenated word over two compositorial lines:
    Because the hyphen is only placed at the line beginning to indicate that Barons is a single word, encode the hyphen with the @type="hyphenInWord" attribute and value as follows:
    Maior and Ba<lb type="hyphenInWord"/>rons of this
Note that <lb> elements are not considered to be white space between the text they separate. If a space would be between the two letters separated by a line beginning, include white space after the <lb> in your XML. If there is no space, as in the hyphenated cases discussed above, do not include a space in the XML.
Hyphens may also occur in catchwords at the bottom of a page. These are not actually line beginnings since often the word appears in full on the following page. In these cases, simply transcribe the hyphen character.
You will also occasionally see instances like this:
in which a multiline quotation is signalled by the use of quotation marks or commas (more commonly the latter, as shown in this example). The quotation should of course be tagged with a <q> element, but the use of these punctuation marks should also be encoded using the <pc> element immediately following each line-beginning (<lb>) element, as follows (other tagging has been removed for clarity):
<lb/><pc type="lineQ">,,</pc>of England, ſendeth gréeting to the Sheriffes & all his miniſters, <lb/><pc type="lineQ">,,</pc>as alſo to all his louing ſubiects, French and Engliſh of London, <lb/><pc type="lineQ">,,</pc>Know ye that I haue granted to God

Encode Page Beginnings

Use a self-closing <pb> element to indicate where page beginnings occur in the front matter, main text, and back matter of a work. For example, look at MoEML’s diplomatic transcription of Anthony Munday’s Metropolis Coronota, in which page beginnings are marked by a horizontal line. These digital page beginnings reproduce the material page beginnings observed in the EEBO facsimiles of the original pageant book. To reproduce the pagination of the pageant book in XML, we inserted a self-closing <pb> element wherever page beginnings occurred. For example,
<p><lb/>Ere we returne to ground agen, <lb/>Seeing iolly Christmas drawes ſo neere, <lb/>When as our ſeruice may appeare, <lb/>Of much more merit then as now, <lb/>Which doth no larger ſcope allow,</p> <pb/> <p><lb/>Then that which is already done; <lb/>Your loue, my Lord, ſo much hath won <lb/>Vpon the Fryer and his Compeeres, <lb/>As we could wiſh to liue whole yeeres, <lb/>To yeeld you pleaſure and delight, <lb/>Be it by day, or be it by night.</p>
Note that, if you are encoding a transcription based on facsimiles (page-images) from either Early English Books Online (EEBO) or the English Broadside Ballad Archive (EBBA), you should provide links to the facsimiles for users to access. Note that EEBO is behind a pay-wall, so only some users will be able to access the links. Alternatively, EBBA is open-source, meaning that all users will be able to access the links.

Link to Facsimiles in EEBO

Each facsimile in EEBO has been assigned a document identifier and a page number (note that the page numbers assigned to the facsimiles will almost invariably differ from the material book’s page numbers). You can determine the document identifier and the page number of any given facsimile by analyzing the http:// address. For example, consider this facsimile of Anthony Munday’s Metropolis Coronata, which has an http:// address of http://eebo.chadwyck.com/fetchimage?vid=13311&page=3. Its document identifier is 13311 and its page number is 3.
To link to an EEBO facsimile, add a @facs attribute to the <pb> element immediately preceding the content depicted in the facsimile. The value of the @facs attribute should contain in sequence:
  1. the prefix moleebo:, meaning MoEML link to EEBO
  2. the document identifier assigned to the facsimile
  3. a pipe character (|)
  4. the page number assigned to the facsimile
The following example demonstrates how to encode a link to the aforementioned facsimile of Metropolis Coronata:
<pb facs="http://eebo.chadwyck.com/fetchimage?vid=13311&page=3&width=1200"/>
Whenever you insert a <pb> element, link it to the appropriate facsimile in this way. In most cases you will have two <pb> elements pointing to the same facsimile because EEBO displays a two-page spread in each facsimile.

Link to Facsimiles in EBBA

Unlike EEBO, each facsimile in EBBA has been assigned only a document identifier. Page numbers are not necessary because each facsimile depicts an entire broadside ballad (document) in one image. You can determine the document identifier for a given facsimile by analyzing its http:// address. For example, consider this facsimile of ballad entitled London’s Praise, or, the Glory of the City, which has an http:// address of http://ebba.english.ucsb.edu/ballad/22002/image. Its document identifier is 22002.
To link to an EBBA facsimile, add a @facs attribute to the <pb> element immediately preceding the content depicted in the facsimile. The value of the @facs attribute should contain the prefix molebba:, meaning MoEML link to EBBA, followed by the document identifier assigned to the facsimile.
The following example demonstrates how to encode a link to the aforementioned facsimile of London’s Praise, or, the Glory of the City:
<pb facs="http://ebba.english.ucsb.edu/ballad/22002"/>
Because broadside ballads were printed in two pages on a single broadside sheet, most transcriptions of broadside ballads will require two <pb> elements that both link to the same EBBA facsimile.

Encode Forme Works

Most pages in early modern books will include text in both the head margin and bottom margin of the page. This text is often repeated on multiple pages because some of it remains in the form when the rest of the type is dumped out. We call this kind of text forme works, and it includes running titles, page numbers, signature numbers, and catchwords. Tag this content using the <fw> element with @type and @style attributes assigned to unique values for running titles, signatures, and catchwords. The mayoral pageant books usually do not include page numbers. Note that forme works text must be transcribed and checked for each page on which it is found. Do not assume it will be the same on every page.

Running Titles

Most early modern books will have running titles atop each page of textual content. Tag each running title using the <fw> element and a @type attribute with a value of "header". Add also a @style attribute with CSS values to indicate where the running title appears; for example, a @style attribute with a value of "text-align: center" indicates that the running title is centered. Note that you are not indicating how we want the transcription to be rendered; you are describing how it appears in the primary source that you are transcribing. You might think of your styling as descriptive CSS rather than prescriptive CSS.
<fw type="header" style="text-align: center;"> Metropolis Coronata. </fw>


Most early modern books will also have signatures in the forme works at the bottom of the recto side of some leaves (usually the first three leaves in a gathering of four). Tag each signature using the <fw> element and a @type attribute with a value of "signature". You are transcribing the signature exactly as it appears in the source, not extrapolating a signature for citation purposes. For example, the first leaf in the B gathering is usually signed B. (For citation purposes, we would extrapolate the information B1r if we were quoting text found on the recto side of that leaf.) In most instances, you will also want to include a @style attribute with a value of "text-align: center" so that the signature will be placed in the centre of the footer.
<fw type="signature" style="text-align: center;"> A 3 </fw>


Most early modern books will also have catchwords in the forme works at the bottom of each book page. A catchword anticipates the first word of the following page. Tag each catchword using the <fw> element and a @type attribute with a value of "catchword". In most instances, you will also want to include a @style attribute with a value of "text-align: right" so that the catchword will be placed in the right-hand corner of the footer.
<fw type="catchword" style="text-align: right;"> euer- </fw>

Encode Forme Works in Longer Texts

Since forme works are by nature similarly laid-out and formatted throughout a text, when you are encoding a longer text, it makes sense to use <rendition> elements in the header to control the appearance of them, rather than repeating the same @style settings on page after page. This section shows a worked example of how to use <rendition> and its @selector attribute to describe the appearance of forme works in a Stow 1598 text.
One advantage of using <rendition> is that the encoded transcription is much less cluttered by CSS. The following example shows the encoding of the top-of-page forme works on a verso and a recto page:
<!-- Verso: -->
<pb n="64" facs=""/>
<fw type="pageNum" place="top-left">64</fw>
<fw type="header" style="margin-top: -1em;">Of Orders and Cuſtomes.</fw>
<!-- Recto: -->
<pb n="65" facs=""/>
<fw type="header">Of Orders and Cuſtomes.</fw>
<fw type="pageNum" place="top-right">65</fw>
There are two things to note:
  • The items are encoded in the order in which they appear on the page; so for a verso, the page number appears first since it’s in the top left, and for a recto, it appears after the running header because it’s in the top right.
  • No CSS is needed, because the layout and appearance will be handled using <rendition> elements in the header.
This is how those rendition elements will appear. First, the running header, which is the simplest:
<rendition scheme="css" xml:id="stow_1598_orders_runningheader" selector="fw[type='header']"> display: block; text-align: center; font-size: 120%; margin-bottom: 1em; font-family: Georgia; text-indent: 0; </rendition>
Often, when using a <rendition> element, we point to it from an element in the text using the @rendition attribute. But here we don’t need to do that; instead, we can use the @selector attribute to point from the <rendition> to all the elements in the text to which it applies. The content of @selector is a CSS selector, which basically says all <fw> elements which have @type="header". The CSS then centers the text, sets its size and family, and overrides any text-indent value it might have inherited from its ancestors. It also provides a margin-bottom value to separate it from the following text on the page.
Next, we need to style the top-left and top-right page numbers respectively:
<rendition scheme="css" xml:id="stow_1598_orders_lpagenum" selector="fw[type='pageNum'][place='top-left']"> display: block; margin-top: 0; margin-left: 0; margin-right: 0; margin-bottom: -1em; font-family: Georgia; text-indent: 0; text-align: left; </rendition>
This <rendition> selects page numbers which are at the top left (i.e. verso page numbers) and styles them appropriately. The key value is margin-bottom: -1em. This ensures that the page number is pulled down onto the same line as the running header that follows it.
For the recto page number, we have to do the reverse:
<rendition scheme="css" xml:id="stow_1598_orders_rpagenum" selector="fw[type='pageNum'][place='top-right']"> display: block; margin-top: -2.5em; margin-left: 0; margin-right: 0; margin-bottom: 1em; font-family: Georgia; text-indent: 0; text-align: right; </rendition>
Here, we set the margin-top to -2.5em in order to pull the page-number up onto the same level as the running header; note that the running header has a margin-bottom of 1em, and its font-size is large, so the negative value has to be higher than for the verso page-number.
We can take the same approach for signatures and catchwords. Here the situation is slightly different because catchwords appear on almost every page, but signatures do not. This is the approach:
<!-- In the header: -->
<rendition scheme="css" xml:id="stow_1598_orders_catchword" selector="fw[type='catchword']"> display: block; text-align: right; text-indent: 0; </rendition>
<rendition scheme="css" xml:id="stow_1598_orders_signature" selector="fw[type='signature']"> display: block; text-align: center; text-indent: 0; </rendition>
<!-- In the text: -->
<fw type="signature">F</fw>
<fw type="catchword" style="margin-top: -1em;">ſayings</fw>
Notice that we add a local margin-top value to the catchword to pull it up onto the same line as the signature. This is only necessary when the signature exists, so we don’t put it in the <rendition>; we just add it when we need it.

Encode Last-word Wraps

The printer of primary source may occasionally wrap the last word in a line onto the white space near the right-margin of the next line. For example, observe how the printer of Thomas Middleton’s The Triumphs of Truth (1613) formats the following two lines:
Enuy Learne now to ſcorne thy Inferiours, thoſe moſt loue (thee,
And wiſh to eate their Hearts, that ſit aboue thee.
(sig. B3r)
The wrapped word is not a forme work, so it requires a different encoding practice. To encode a last-word wrap, tag the last word in the line using a <hi> element with a @style attribute. In most cases, the value associated with the @style element should be "position: relative; top: 1em; left: -2.5em;". This CSS value essentially means relative to the element’s normal position, put the element 1em lower, and 2.5em to the left. The previous Middleton passage should therefore be encoded in TEI-XML as follows:
<lg style="margin-left: 3em; font-style: italic;">
  <l><name style="font-style: normal;" ref="mol:ENVY1">Enuy</name> Learne now to ſcorne thy Inferiours, thoſe moſt loue <hi style="position: relative; top: 1em; left: -2.5em;">(thee,</hi></l>
  <l>And wiſh to eate their Hearts, that ſit aboue thee.</l>

Pass 4: Encode MoEML Dates, Names, and Toponyms

Finally, you need to tag all dates, names, and toponyms.

Pass 5: Styling and Conservation

In primary source transcriptions, we aim to conserve the original styling of the early modern printed book or manuscript being transcribed to the extent possible given Cascading Style Sheet (CSS) and browser standards.

Apply Style to Document Margins

The presentation of margins in our diplomatic transcriptions of primary sources happen at three levels:
  1. In the <text> element at the beginning of the document.
  2. As a "margin-left" value under the @style attribute of the <p> and <lg> elements, as well as their subsidiaries.
  3. As a "text-indent" value under the @style attribute of the block level elements for a single indent.
The properties defined in the <text> element constrain the width of the presented document. Consider the <text> element in LOVE8.xml (Thomas Middleton’s The Triumphs of Love and Antiquity), for example:
<text style="width: 28em; padding-left: 5em; padding-right: 5em;">
Using the measurement of twenty-eight em (an em being a correlative sizing—one em in a twelve point typeface would be twelve points), we limit the width of LOVE8.xml. This is then centred with "padding-left" and "padding-right" so that the document is not flush left on the MoEML site. With this higher-order margin creation in effect, further lower-level margins work inwards from these definitions. Essentially, the margins ordered in the <text> element create the outline of the physical page, and further margins create the difference between page-edge and text as it appears throughout the document.
Throughout the document a range of margins are both possible and present, hence the need for lower orders of encoding. Margin creation at the page level is impractical because our document hierarchy is conceptual (chapters, sections, paragraphs etc.) rather than page-based, so there is no page container. Therefore we opt for margin creation at the block level: paragraphs and line groups. The following set of <p> elements in LOVE8.xml serves as an example:
<p style="margin-left: 3em;">Harts, 6. Bucks, and a Tun of Wine, to make mer<lb type="hyphenInWord"/>ily, and this Noble feaſt was kept at <ref target="mol:DRAP2">Drapers Hall</ref>. <lb/>An 1463. <name>Q. <hi style="font-style: italic;">Elizabeth Grey</hi></name>, his Wife, Daughter to <lb/>to <name><hi style="font-style: italic;">Richard, Wooduile</hi>, Earle <hi style="font-style: italic;">Rivers</hi></name>, & to the Dutcheſſe <lb/>of Bedford, ſhe was Mother to the Lord Grey of <lb/>Ruthen, that in his time was Margueſſe Dorſet.</p> <p style="margin-left: 3em; text-indent: 1em;"><name ref="mol:RICH3">K. <hi style="font-style: italic;">Richard</hi> the 3</name>. Brother to <name ref="mol:EDWA6"><hi style="font-style: italic;">Edward</hi> 4</name>. D. of <lb/><hi style="font-style: italic;">Gloceſter</hi>, and the House of <hi style="font-style: italic;">Yorke</hi>.</p>
You will notice the lack of styling attributes for the right margin. We define the left side for lower-order styling, but the right is defined negatively by the left definitions and the initial mark-up in the <text> element. When additional margins are needed within pre-defined block elements, continue to use "margin-left". Remember that this compounds across levels, so that a paragraph with a margin of two em within a <div> defined with a four em margin will have a final offset of six em.


For longer primary source documents with consistent styling, such as A Survey of London, we use <rendition> elements to declare CSS values that we can reuse throughout a document. For instructions, please see Use the <rendition> Element and @rendition Attribute.


Cite this page

MLA citation

Holmes, Martin D., Janelle Jenstad, Zaqir Virani, Tye Landels-Gruenewald, Sarah Milligan, and Lucas Simpson. Encode a Primary Source Transcription. The Map of Early Modern London, Edition 7.0, edited by Janelle Jenstad, U of Victoria, 05 May 2022, mapoflondon.uvic.ca/edition/7.0/encoding_primary_sources.htm.

Chicago citation

Holmes, Martin D., Janelle Jenstad, Zaqir Virani, Tye Landels-Gruenewald, Sarah Milligan, and Lucas Simpson. Encode a Primary Source Transcription. The Map of Early Modern London, Edition 7.0. Ed. Janelle Jenstad. Victoria: University of Victoria. Accessed May 05, 2022. mapoflondon.uvic.ca/edition/7.0/encoding_primary_sources.htm.

APA citation

Holmes, M. D., Jenstad, J., Virani, Z., Landels-Gruenewald, T., Milligan, S., & Simpson, L. 2022. Encode a Primary Source Transcription. In J. Jenstad (Ed), The Map of Early Modern London (Edition 7.0). Victoria: University of Victoria. Retrieved from https://mapoflondon.uvic.ca/editions/7.0/encoding_primary_sources.htm.

RIS file (for RefMan, RefWorks, EndNote etc.)

Provider: University of Victoria
Database: The Map of Early Modern London
Content: text/plain; charset="utf-8"

A1  - Holmes, Martin
A1  - Jenstad, Janelle
A1  - Virani, Zaqir
A1  - Landels-Gruenewald, Tye
A1  - Milligan, Sarah
A1  - Simpson, Lucas
ED  - Jenstad, Janelle
T1  - Encode a Primary Source Transcription
T2  - The Map of Early Modern London
ET  - 7.0
PY  - 2022
DA  - 2022/05/05
CY  - Victoria
PB  - University of Victoria
LA  - English
UR  - https://mapoflondon.uvic.ca/edition/7.0/encoding_primary_sources.htm
UR  - https://mapoflondon.uvic.ca/edition/7.0/xml/standalone/encoding_primary_sources.xml
ER  - 

TEI citation

<bibl type="mla"><author><name ref="#HOLM3"><surname>Holmes</surname>, <forename>Martin</forename> <forename>D.</forename></name></author>, <author><name ref="#JENS1"><forename>Janelle</forename> <surname>Jenstad</surname></name></author>, <author><name ref="#VIRA1"><forename>Zaqir</forename> <surname>Virani</surname></name></author>, <author><name ref="#LAND2"><forename>Tye</forename> <surname>Landels-Gruenewald</surname></name></author>, <author><name ref="#MILL2"><forename>Sarah</forename> <surname>Milligan</surname></name></author>, and <author><name ref="#SIMP5"><forename>Lucas</forename> <surname>Simpson</surname></name></author>. <title level="a">Encode a Primary Source Transcription</title>. <title level="m">The Map of Early Modern London</title>, Edition <edition>7.0</edition>, edited by <editor><name ref="#JENS1"><forename>Janelle</forename> <surname>Jenstad</surname></name></editor>, <publisher>U of Victoria</publisher>, <date when="2022-05-05">05 May 2022</date>, <ref target="https://mapoflondon.uvic.ca/edition/7.0/encoding_primary_sources.htm">mapoflondon.uvic.ca/edition/7.0/encoding_primary_sources.htm</ref>.</bibl>