[H O M E : Emblem Project Utrecht]

Emblem Project Utrecht

EPU Guidelines for Encoding Emblem Books

1 Introduction

These are the guidelines we will apply in digitising the emblems of Vaenius and other emblem writers. It is not complete, as it does not describe rules for all possible situations. When dealing with new books, new complications will arise and require changes or additions to the current rules. Even the books we have been digitising up to now have some phenomena awaiting a satisfactory solution.

Nevertheless, we feel it is important both for ourselves and for visitors to the site to state which principles and policies we have applied. For some readers it may give an idea about the possibilities of text markup using the XML/TEI scheme, and for others already aware of these possibilities, it may help to have a document describing the way we have applied them. These guidelines are also useful to our project so we can refer back to these rules in case of doubt.

Procedures for encoding the bibliography are dealt with elsewhere.

1.1 Examples In These Guidelines.

In these guidelines we will give a number of examples. It should be stressed that the examples are meant to illustrate the issues being discussed in the section where they occur. They are not necessarily complete and ready to be copied into the editions. To increase legibility extra tabs, spaces or linefeeds may have been added.

Table of contents

2 Principles

In making decisions about the type of markup to apply to our editions, we have used the following principles:

Some other considerations were made when deciding on an appropriate markup:

3 Transcription

The guiding principle in transcription is to faithfully representing what is in the source. While transcribing, we will try not to regularise or alter the source texts. Transcription results in a text will conform to the source in use of small and capital letters, spacing, use of u, v, w, i, j, y, interpunction and diacritics.

3.1 Unusual Letters and Diacritics

XML uses the characters of UnicodeLinks to page outside of EPU. This means nearly all possible characters can be encoded directly using their Unicode representation. Nevertheless, to enhance portability and to allow us to use non-unicode supporting software, we will represent all unusual characters (not being plain ASCII) using so-called 'character references'. The relevant DTD fragment contains most of the ones we will need. This works as follows:

é in the text
é the way we transcribe it
eacute the name in the DTD this refers to
é the Unicode representation of é

For signs not yet present in DTD, new ones will be introduced. For example:

e with macron:
in xml: &emacron;
in dtd: <!ENTITY emacron "&#x275">
q with '3' used as an abbreviation for 'que'
in xml: &q3;
in dtd:<!ENTITY q3 "que">

The last example illustrates what we will do when there does not seem to be a valid unicode representation of an unusual sign. The XML file will faithfully record the presence of the unusual sign and the DTD will give a provisional expansion ('que'). When a valid Unicode representation is available, it will be substituted for the provisional expansion. Even thought this procedure will also be used when a unicode representation is available, but most fonts do not seem to support it.

3.2 Apostrophes

The emblem texts use apostrophes to indicate a part of a word has been eliminated (as in Dutch "t'samen", or French "en l'aimant")

In Unicode, the apostrophe is rendered 'x2019'. Our DTD contains:

<!ENTITY apost "&#x2019;"> <!-- apostrophe -->

and we code:

en l&apost;aimant

A left single quote is represented as '&lsquot'.

3.3 Hyphens and Other In-word Punctuation

To build concordances of the words contained in the emblem texts we need to be able to distinguish individual words. Therefore, we should indicate which characters usually considered interpunction, are really within a word rather than between words. We will do this using the <c> element (character), and use its type attribute to indicate the way the character is used. Allowed values for the type attribute are:

shy soft hypen, hyphen used to split a word over lines
wi interpunction character within a single word
ws interpunction character between words

Examples (all from Amoris divini emblemata):

Soft hyphen:
sa<c type="shy">-</c><lb/>pio
Hyphen within word (French 'en-bas'):
en<c type="wi">-</c>bas
Hyphen between words (French 'di-ie', 'say I'):
di<c type="ws">-</c>ie

3.4 Structural or Ornamental Characteristics of Page Lay-out

The editions we will make are meant to facilitate access to the emblem's content. The user may always view the characteristics of page lay-out by consulting the facsimile. We will therefore not try to reproduce regularly used or ornamental typographical peculiarities (identation, italics and font sizes).

Lay-out characteristics that are apparently irregular or non-ornamental are encoded using the rend attribute, either on an existing element or on a <hi> element introduced for this purpose. Values of the rend attribute include:

ital italic
bold bold
large larger than surrounding font
sup superscript
sub subscript
sub subscript
small larger than surrounding font
smallcap small caps
indent1 indent 1 unit
indent2 indent 2 units
indent3 indent 3 units

Multiple values may be entered in a single rend attribute, and will be separated by spaces. We may need to introduce something similar to the Brown's Women Writers Projects' 'rendition ladders'.

Examples include:

<l rend="ital">italic line</l>
<l>normal line with <hi rend="large ital">largeitalic</hi> word</l>

3.5 Ornaments

We ignore non-textual page ornaments or adornments.

3.6 Missing, Unclear or Damaged Text

The transcription will not attempt to represent the larger pieces of missing text. If mottoes or epigrams are missing then the commentary will explain this situation to the visitor. If another text witness has been used to supply the missing parts, then this will be noted in the commentary. Smaller fragments of text that are missing will be encoded as <gap>. If the correct reading of a fragment is uncertain, this will be encoded using the <unclear> element.

3.7 Roman Numerals

UnicodeLinks to page outside of EPU defines special characters for ruman numerals (u2150 and up). We do not use these, but transcribe Roman numerals with the usual alphabetic characters (IVXLCDM and their lower-case counterparts). In those cases where the source uses reversed letters to write a Roman numeral, we ignore this and write

<num value="1612">MDCXII</num>
in the transcription in [Dedication]emblem or text edited on EPU website.

4 Editorial Enhancements and Normalisations

The basic principle in editorial additions to the source texts is that the source text should always remain available. To be able to present an easily readable version, we may add certain encodings. The editorial introduction to each book should state clearly which of the following encodings have been applied.

4.1 Normalisation of Spelling

We may normalise, for some or all, of the languages used in a book the use of u, v, w, i and j as follows:

<orig type="spelling" reg="overvloedich">oueruloedich</orig>

The type attribute on the <orig> element is a dtd extension. We need the attribute to allow us, when showing the text to the user, to selectively apply these normalisations. This will be the only level of re-spelling applied to the texts. We may add translation into Modern Dutch or English (cf. Translations).

With a normalisation in spelling, we may silently normalise the use of spaces.

4.2 Normalisation of Spacing

Spaces which are, according to modern usage, missing or superfluous may be added or removed using the <orig> element:

Removed: <orig type="space" reg=""> </orig>
Added: <orig type="space" reg=" "/>

We do not record minor variations in the width of spaces.

4.3 Normalisation of Interpunction

Interpunction may be changed to current usage, again using the <orig> element:

<orig type="interpunction" reg=","> /</orig>

For the present time the only form of interpunction which we will normalise is the slash ('/' or '//'), usually to be replaced by a comma.

With a normalisation in interpunction, we may silently normalise the use of spaces (as in the above example).

4.4 Abbreviations

Abbreviations may be resolved using the <abbr> element:

<abbr expan="civitate">civit.</abbr>

If some mark was used in the source to indicate an abbreviation was employed, this mark will remain present within the content of the <abbr>element ('ā' in the following example):

<abbr expan="hant">h&amacron;t</abbr>

To minimize overlap with other encodings abbreviations may be encoded within words. In the above example, this would result in:

h<abbr expan="an">&amacron;</abbr>t

4.5 Use of Capital Letters

Capital letters are to be transcribed conforming to the use in the source (cf. Transcription). We do not intend to normalise their use. To be able to use (capitalized) motto texts as a title or as a hyperlink, a separate list of titles will be built (cf. Emblem Titles).

4.6 Corrections

The source text may contain a number of misprints. They may be coded as follows:

corrected <sic corr="error">errot</sic>

5 Encoding Emblem Structure

5.1 Emblem Book Structure

An emblem book is a collection of emblems. Some emblem books (e.g. those by Vaenius) have a single subject, but many others do not. Each emblem can be viewed as an independent text and will often have been read as such. We therefore chose to represent the emblem book as a <group> of <text>'s. We introduced a type attribute on the <text> element to record the type as an emblem.

Introductory material (titlepage, dedications and panegyrics) will similarly be encoded in a series of <text> elements, now with the value front for the type attribute.

Back matter from the source (e.g. a colophon) will be encoded in a series of <text> elements, with the value back for the type attribute.

Emblem books that have multiple sections (e.g. Luyken's Jesus en de ziel or Suderman's collections referring to Hugo and Vaenius) will be encoded as nested <group>s. Each section will be a <group> in the top-level <group>. There may be introductory material <text> elements with type front) at the book level, these will be encoded as part of the top-level <group>) and at the section level encoded as part of the lower-level <group>.

The <group>'s <front> and <back> elements will be reserved for editorial material. The editorial introduction will be encoded as a <div> with type edintro. In the <back> element, we will store lists of titles, lists of entities representing images, lists of <link> and <interp> elements.

When encoding editorial material we will use the standards of TEI LiteLinks to this book in bibliography as much as possible.

5.2 Emblem Structure

In encoding emblem structure, we have tried not to make assumptions about the relative importance of the emblems constituents. We have therefore decided to encode each constituent as a <div> in its own right having the appropriate value (motto, pictura, subscriptio) on its type attribute. Besides the fundamental assumption about the constituents' equality, there are other reasons why we will avoid using the <head> element to encode mottoes:

  • There are emblems which carry both a header and a motto (the header being usually something like 'emblem 14'; see Emblem Books in LeidenLinks to this book in bibliography, pp. 130, 153)
  • There are emblems which seem to carry two mottoes (see Emblem Books in LeidenLinks to this book in bibliography, p. 141 )
  • Some mottoes may be short poems in their own right (cf. the Dutch motto in emblem or text edited on EPU website and in the Spanish version, this occurs in all emblems); using the <head> element to encode the motto would entail having to use an <lg> element within a <head>, which the TEI dtd's do not allow.

We will only use the <head> element in those situations where the source uses a heading which contains no additional meaning (such as a number). One situation where this may be useful is in encoding heads which just state the language of the epigram which follows (see the example in Emblem Books in LeidenLinks to this book in bibliography, p. 127 ). More specifically, the headings of the vernacular epigrams in Amorum Emblemata will be encoded as <div>'s.

What we do not want to lose in our encoding is that a particular heading and epigram belong together. It would be possible to express this relation between <div>'s using a <join> element (which creates a virtual group of texts). This seems like an error-prone process, difficult for subsequent processing, and not very common. We have decided to encode this grouping of <div>'s using higher level <div> elements.

Technically, this creates a hierarchy of <div>'s containing <div>'s. This hierarchy, results from the grouping of constituents in the source and is not supposed to reflect the relative importance of each constituent.

These general reflections do not answer the question as to how the constituents should be grouped when encoding an individual emblem book. No general answers will be given here. Does the Latin motto, in Vaenius's Amorum emblemata, belong to the emblem as a whole, or more specifically to the Latin quotations? Also, do the vernacular headings (now considered to be mottoes too) belong at the same level as the motto in Latin? And how about Amoris divini emblemata, which has no vernacular mottoes?

For Amorum emblemata we decided upon this structure (simplified, Latin motto grouped with Latin quotations):

<text id="va1608001">
  <pb n="..."/>
  <div lang="lat" id="va16080011">
   <div type="motto" id="va160800111">
    <p>[... text motto in Latin ...]</p>
   <div type="subscriptio" id="va160800112">
    [... quotations in separate div's ...]
  <div lang="eng" id="va16080012">
   <div type="motto" id="va160800121">
    <p>[... text motto in English ...]</p>
   <div type="subscriptio" id="va160800122">
    <lg>[... text epigram in English ...]</lg>
  [... other languages ...]
  <pb n="..."/>
  <div type="pictura" id="va16080017">
   [... reference to picture ...]

For Amoris divini emblemata we decided to group the Latin motto with all of the emblem texts, rather that just with the Latin quotations:

<text id="va1615001">
  <pb n="..."/>
  <div lang="lat" type="motto" id="va16150011">
   <p>[... text motto in Latin ...]</p>
  <div lang="lat" type="subscriptio" id="va16150012">
   [... quotations in separate div's ...]
  <div lang="esp" type="subscriptio" id="va16150013">
   <lg>[... text epigram in Spanish ...]</lg>
  [... other languages ...]
  <pb n="..."/>
  <div type="pictura" id="va16150016">
   [... reference to picture ...]

5.3 Page Breaks

Page breaks will be encoded using <pb> elements. The n attribute will be used to encode the page number. If no page number is present, foliation or signature will be used. If this too is missing, a number will be added between square brackets. For example:

<pb id="pb2" n="2"/>
<pb id="pbA2" n="A2"/>		
<pb id="pbcolon4v" n="[(:)4v]"/> 
	[The first emblem page in Amorum Emblemata]

The id attribute will serve as an anchor for associating page images with the pages (cf. Page Facsimiles).

This is not quite satisfactory because an editorial addition, such as the supplied page number, should be encoded using the <supplied> element. In this case however, the <pb> element itself represents the source (there actually is a page break, even if there is no page number). It is the attribute value which has been added but one can not mark an attribute value using the <supplied> element.

Alternatively, we could just omit the n attribute when there is no page number in the source. Even in this case some value would have to be shown to the user. This problem will be looked at again in the future.

5.4 Quotations

5.4.1 Separately Presented Quotations

The quotations (or texts presented as such) which Vaenius used to his emblems will be encoded each in a separate <div>. If the text is indeed a quotation then the <div> element will contain a single <cit> element with a <bibl> element (if the source contains a bibliographical reference), a <quote> element to contain the actual text and an <xref> element containing our own bibliographical reference.

Within the <quote> the quoatation text wil be encoded as a <lg> if it is in verse. In the case of prose, there is no extra element. For example:

 <bibl>August. in Psalm. 145.</bibl>
 <quote>Non vobis ... mundi.</quote>
 <xref><!-- our reference to a source --></xref>
   <l>Elige cui dicas, tu mihi sola places.</l>
 <xref><!-- our reference to a source --></xref>

Vaenius always gives his bibliographical references in the left margin. This is an example of a regularly recurring typographical feature without additional meaning. As such, it will not be recorded (cf. Structural or Ornamental Characteristics of Page Lay-out).

5.4.2 Inline Quotations

If a quotation occurs within running text, possibly within another quotation, the same elements (<cit> containing <bibl>, <quote> and <xref>) will be used. The <cit>'s rend attribute will be set to inline to indicate its location. Quotations recognized as such within the epigrams, will be encoded in the same way. In the case of Vaenius, the <bibl> element will be missing. If (in poetry) a quotation spans several lines it should be split in parts (because xml elements can not overlap). We will then code several groups of <cit>'s contaning <quote>'s and use the <cit>'s next and prev attributes to connect the quotations parts. The last <cit> element will contain the <xref> element to point to the source.

5.4.3 Sources

We will add explicit source references, so for the present time we will not add further markup to the original bibliographical references. The authors, titles and abbreviations used will be left unmarked.

For the way in which we encode our own references, see References to Sources. If the reference needs annotation (because it is doubtful or is different from the one given by the emblem author etc.) it will be given in a <note> with type source given at the end of the <quote>.

If a text is present in the group of presumed quotations, but does not appear to be a reference to a source by the emblem or a reference found by us, then it will not be encoded as a <cit> element, but as a plain <lg> or <p>.

5.5 Line Breaks

In prose line breaks will be encoded using an empty <lb> element. Cf. Hyphens and Other In-word Punctuation about hyphens used to split a word over two lines.

5.6 'Split' Emblems.

In some books, the constituents of a single emblem are not presented consecutively. A case in point is De la Feuille, where pages with texts from 12 or 15 emblems alternate pages with images. In other books, all picturae may have been bound together.

In these cases, information from multiple pages will be joined into a single <text> element, to help present the emblem as a unit. The <pb> elements will indicate the pages which the constituents stem from.

Where this has been done, the book's editorial introduction will have to explain the procedure.

5.7 Notes In The Emblem Sources

Original notes are encoded as <note> elements. The resp attribute's value orig will be used to distinguish them from editorial annotation.

The place attribute takes the following values:

below below the text, not necessarily at foot of page
foot at foot of page
left in left margin
right in right margin
accolade after accolade immediately following the line(s) annotated

5.8 Songs

Songs in the emblem subscriptio will be encoded as <lg> elements within <div> elements, as usual. The <lg>’s type attribute will have the value song.

Tune indications will be encoded using the <tune> element. The name of the tune which the tune indication refers to will be encoded as a <title> element.

Lines which are sung twice are often indicated using an accolade followed by (e.g.) 'bis'. This situation will be encoded as a <note> element with place accolade. The <note>'s target and targetEnd attributes will refer to id attributes of the relevant verse lines. The note's text will contain 'bis' for example.

Stanzas will be encoded as nested <lg> elements. Stanza numbers will be stored as value of the <lg>'s n attribute, but only if they are present in the source.

5.9 'Fixed' Emblem Parts

In some emblem books, the emblem's constituent texts carry fixed headings. For example, in Luyken's Jesus en de Ziel, the motto and epigram are followed by a group of Bible quotes invariably preceded by the text 'Goddelyk Antwoord' ('Divine response'). In other emblems, the subscriptio may be divided in constituents such as 'Zang', 'Toezang' and 'Toepassing'.

Each of these fixed headings will be encoded within a <head> element. However, it may also be desirable to record the implied text typology at the <div> level, this would allow its use in selective retrieval of text constituents. It is therefore recommended to use a type or subtype attribute on the <div> to store the information.

5.10 Original Tables of Contents

Original tables of contents will be included in the electronic edition. They are encoded as <text> elements (type="front"). The text will contain a single <div> with type attribute toc. The <div> will contain a <list> element which contains the actual table of content text. <ref> or <xref> elements will be used to hyperlink the table of content to the emblems and/or the page images.

5.11 Title Pages

We will encode title pages as <text> elements (type="front"). The text will contain a single <div> with type attribute titlepage. The <div> will contain a <titleBlock> element which contains the actual title page text.

The same procedure will be used for new title pages somewhere in the inside of a work.

5.12 White Lines In Prose

A prose text will generally consist of multiple paragraphs (<p> elements). If the text uses white lines to separate the parapraphs then these lines may be safely ignored. However, if some paragraphs are separated by white lines while others are not, then the presence of the white line should be encoded, using the <space> element.

The <space> element cannot occur directly in a <div> element. We will always enclose it in an otherwise empty, <p>. So:

<p> ... end of previous paragraph.</p>
<p><space dim="vertical" extent="1"/></p>
<p>Start of next paragraph ...</p>

6 Annotation

We recognize there will be several types of annotation. Part of the annotation will be stored as <note> elements within the emblem <text>'s <body>, while other types will be stored in <div> elements in the <back> of the emblem <text>. For that purpose we will use a <div> with type comment and a <div> with type translations.

6.1 Summaries

The summary will contain a very brief description of an emblem's meaning, preferably containing just a few words. It may be compared with what is termed the subject A Revised Format for the Index Emblematicus.Links to this book in bibliography, pp. 138-139.

The summary will be encoded as a <note> with type subject. These <note>'s will be placed in the <div> (type comment) in emblem's <back>.

Relevant words from the summary may be indexed.

6.2 Local Comments

Annotations to the epigram texts will be stored where they belong. They will be placed in <note> elements with type annot or source (source for annotating a quotation's source, annot for other annotations). If the emblem source texts contain notes, editorial notes will carry the resp attribute with value ed.

Notes which concern the epigram as a whole will be placed in a <note> element after the epigram's first word. Other notes will be placed after the last word to which they refer. Text from the epigram which is quoted inside the <note> will be marked up using the <mentioned> element. Notes will not be numbered and these numbers will be generated automaticaly.

6.3 Commentary

To each emblem we hope to add a commentary which may deal with the following issues:

  • The meaning of the emblem as a whole
  • The way in which the emblem's constituent parts contribute to the meaning
  • The sources of quotations, pictorial motifs, the motto, and combinations of these sources
  • Parallels in other emblem books and contemporary material
  • Uses which other writers or painters may have made of this material
  • Backgrounds to the emblem's intellectual content
  • Remarkable stilistic features

The commentary does not have a fixed order or arrangement. It is encoded within a <note> element with type attribute comment. The <note> element will be part of the <div> with type comment.

More detailed comments on subjects relevant to several emblems should not be given in the emblem-level commentary. They should be part of the editorial introduction to the book.

References to other emblems, paintings and literary works will be encoded using the techniques discussed in References. Whenever these works are not mentioned in the separate section on sources and parallels will be found below References to works mentioned, without claiming something about their status (source or parallel) will have a type discussed.

The commentary itself is a miniature essay and will use notes very sparingly (just to point to pertinent literature).

6.4 Sources

A list with sources and parallels to the emblem or its constituent parts will be placed within a <note> with type parallel in the <div> with type comment. The <note> consists of a <list> with <item>s. The <item>s refer to other works pointing either to <biblScope>s in works, in the bibliography or to other emblems edited on the EPU web site. The reference's type will be encoded on the <xref> or <xptr> element and will be either source or parallel.

Valid values for the <xref>'s type attribute are:

source the named work is a source for the present emblem or quotation
sourcefree (meant especially for quotations) the quotation is not exact
indirect (meant especially for quotations) the quotation is indirect (part of another quotation)
parallel the named work contains a similar text, motif or combination of these, but we do not claim it functioned as a source
discussed work mentioned, not necessarily a source or parallel (it might be mentioned to show how different it is from the emblem under discussion)

A brief text may be added to indicate the kind of parallelism between the present emblem and the work referred to.

For example (from emblem or text edited on EPU website):

<note type="parallel">
   <xptr doc="va1608" type="parallel" from="id (va1608051)"/>
   Metaphor of needle and thread applied [...]: 
   <xref doc="biblio" 
       rend="ref" type="parallel" 
       from="id (cats--011996a)">
     <biblScope type="embl">38</biblScope>

In this example the first is a reference (type parallel) to an emblem in Amorum emblemata, then a reference to an emblem by Cats. See References to Sources for a discussion on how to point to parts of items in the bibliography. The bibliography guidelines will expalin how to use of the <biblScope> element to point to scopes in works.

Another example, from emblem or text edited on EPU website, now referring to an emblem in the same book (see Referring to a Location in the Same Book on the use of <ptr> and <ref> elements):

<note type="parallel">
    Compare the vertical division in: 
    <ptr target="va161502" type="parallel"/>

6.5 Literature

With each emblem, we point to relevant literature. If this seems useful, the reference may be accompanied by a brief indication of an item's relevance. This list will also refer to the location of the emblem in Henkel and Schöne, EmblemataLinks to this book in bibliography. An example from Omnia vincit amoremblem or text edited on EPU website:

<note type="lit">
   <xref rend="ref" doc="biblio" from="id (henkel011976a)">
     <biblScope type="col">386</biblScope> 
   <xref rend="ref" doc="biblio" from="id (praz--011975a)">
     <biblScope type="p">89</biblScope> 

The list of literature is placed in a <note> element with type attribute lit. This <note> is part of the <div> with type comment. The <note>'s contents is a <list> with <item>'s.

6.6 Pictorial Motifs

Pictoral motifs are encoded using the mechanism of <interp> elements described in Motifs and Their Categorisations.

6.7 Henkel and Schöne

The emblem's Bedeutung as it may be given in Henkel and Schöne, EmblemataLinks to this book in bibliography, is copied into a <note> with type hs, again in the <div> with type comment. The description of the picture as given there is copied into a <note> with type hspict in the same location. An example from Au dedans ie me consumeemblem or text edited on EPU website:

<note type="hs">Verzehrender Liebesschmerz</note>
<note type="hspict">Wassertopf auf dem Feuer, 
	das Amor mit dem Blasebalg anfacht</note>

6.8 Translations

We may provide translations of the texts contained in the emblems (mottoes, quotations, epigrams) into modern English or Dutch. The purpose of the translations is to assist in understanding the text; aesthetic considerations are of minor importance here. Translations are in prose. In translating 17-th century English or Dutch to their modern equivalents, grammatical difficulties may be resolved.

Each translation of emblem constituent will be placed in a separate <note> element with type attribute translation. The <note> will be part of the <div> with type translations. The <note>'s lang attribute will be used to show the target language of the translation (cf. Languages). Within the <div> there will be a <linkGrp> element where <link> elements will associate the translation with the emblem constituents. One emblem constituent may be associated with several <note>'s (one for each translation).

An example from Au dedans ie me consumeemblem or text edited on EPU website:

<div type="translations">
 <!-- translations: -->
 <note type="translation" lang="dut" id="he16080021tr">
  Van binnen word ik verteerd.
 <note type="translation" lang="eng" id="he16080021tre">
  Inside I waste away.
 <!-- associating emblem constituents and translations: -->
 <linkGrp targOrder="Y" 
     targFunc="translated translation" type="translation">
  <link targets="he16080021 he16080021tr"/>
  <link targets="he16080021 he16080021tre"/>

6.9 Working Notes

While working on an emblem book we will need some place to store working notes and other information which may be useful at a later stage. These notes may be encoded in a <note> element with type attribute rest, placed in the <div> with type comment.

In the pages presented to external users this information will not be shown.

6.10 Quotations Within Editorial Text

When quoting emblem or other text within editorial text, we will markup the quoted text using the <mentioned> element.

When showing the texts, quotes will be added. There is no need to add them in the XML source.

Larger quotations within editorial text may be encoded using the <quote> element. In these Guidelines there currently is no provision for linking this type of quotation to its source, on the EPU site or elsewhere. If this need should arise, these Guidelines will be extended.

For example (from emblem or text edited on EPU website):

<p> ... Hoofts bijschrift kan men vergelijken met een vers van Cats:</p>
		<l>Soo haest u gunstigh oogh maer eens op my en viel:</l>
		<l>Ghy zijt mijn ander ick, de ziele mijner ziel,</l>

7 Associating Images With Emblems

7.1 The Pictura

The pictura is represented by its own <div> (type pictura (cf. Emblem Structure). A pictura image will be associated with the <div> through the use of a <figure> element which refers to the image.

For example (from emblem or text edited on EPU website):

In the XML file: 
<div id="va16080017" lang="lat" type="pictura">
  <figure id="va1608001pict" 
     ana="ip010002 ip010009"/>
In the DTD-fragment identifying the images: 
<!ENTITY va1608001pict 
   SYSTEM "../images/va1608/editie1608LIF/large/1.jpg" 

This example shows the <figure> element, using its entity attribute points to the description in the dtd which gives the file name asociated with the image. See Document Type Declaration for more information about the document type declaration in the xml files.

7.2 Engraved Emblem Texts

Text engraved in the pictura is transcribed in (possibly multiple) <p>'s within the <figure> element. For example (from emblem or text edited on EPU website):

<figure entity="va161502">

The <p> elements may be contained in the <figure>'s <figDesc> element.

This text may contain <lb> elements. For example (from emblem or text edited on EPU website):

<figure entity="va161501">

However, if an emblem text (motto or subscriptio) is both engraved and printed, such as the mottoes in Heinsius's Emblemata amatoria, this constituent will not be encoded within the <figure> element, but as a standard <div> element. At the book level a decision will be made whether the engraved or printed version will be transcribed. Differences between the versions will be noted in <note> elements.

7.3 Page Facsimiles

This section describes two procedures to associate page facsimiles and pages. The first procedure should only be used if there are multiple images to be associated with a single page (e.g. if we present images from multiple copies or editions). If there is only a simgle image per page, the second procedure should be used.

When page images are available they may be associated with the relevant page using a list of <figure>'s in a division in the <back> of the document (the type of the <div> will be pictlist) and a list of <link> elements which associate <pb>'s and <figure>'s through their respective id attributes. The <linkGrp> grouping these <link>'s will stored in the same <div> as the <figure> elements.

For example (from Amoris divini emblemata):

<div type="pictlist">
  <figure id="va1615pb1" entity="va1615pb1"/>
  <figure id="va1615pb3" entity="va1615pb3"/>
  <!-- etc. -->
  <linkGrp targOrder="Y" 
      targFunc="page facsimile" 
      targType="pb figure">
   <link targets="pb1 va1615pb1"/>
   <link targets="pb3 va1615pb3"/>
   <!-- etc. -->

The entity attributes again point to the descriptions in the dtd which give the file name asociated with the images. See Document Type Declaration for more information about the document type declaration in the xml files.

A simplified and preferred procedure is to associate the page facsimile and the page using the doc attribute on the <pb> elements. The doc attribute is an EPU extension. Its value is of type ENTITY and refers to an entity defined in the DTD.

For example (from emblem or text edited on EPU website):

In the XML file:
<pb id="pb002" n="2" doc="tekst002"/>
In the DTD:
<!ENTITY tekst002 SYSTEM  "../images/ca1627/small/tekst002.jpg" NDATA JPEG>

7.4 Thumbnails

This section describes two procedures to associate thumbnail images and emblems. The first procedure is deprecated, whil the second should be used for all new editions.

[Deprecated:] Thumbnail images can be associated with the relevant emblem <text>'s using the same mechanism as is used to associate pages and images.

For example (from Amoris divini emblemata):

<div type="pictlist">
  <!-- etc. -->
  <figure id="va161501thumb" entity="va161501thumb"/>
  <figure id="va161502thumb" entity="va161502thumb"/>
  <!-- etc. -->
  <linkGrp type="thumb" 
      targOrder="Y" targFunc="page facsimile" 
      targType="pb figure">
   <link targets="va161501 va161501thumb"/>
   <link targets="va161502 va161502thumb"/>
   <!-- etc. -->

[Preferred:] Thumbnail images will be associated with the relevant emblems using the thumb attribute on the <text> elements. The thumb attribute is an EPU extension. Its value is of type ENTITY and refers to an entity defined in the DTD.

For example (from emblem or text edited on EPU website):

In the XML file:
<text thumb="beeld002thumb" type="emblem" n="II" id="ca162702">
In the DTD:
<!ENTITY beeld002thumb SYSTEM  "../images/ca1627/thumbnails/beeld002thumb.jpg" NDATA JPEG>

8 Motifs and Their Categorisations

Interpretative encodings will be assigned to elements by associating them with pertinent <interp> elements. <interp> elements are stored in the document's <back> and grouped into categrories using <interpGrp> elements. The <interpGrp>'s type attribute will be used to describe the kind of interpretation intended. <interpGrp>'s may be nested to create hierarchies of interpretative encodings. Associating an emblem or emblem constituent part with an <interp> element will implicitly associate the constituent with higher-level <interpGrp>'s.

Association between <interp> elements and emblem constituents may be made in three different ways, explained below.

8.1 Associating <interp>'s using ana attributes


Using the element's ana attribute to refer to one or more <interp> elements (if more, the values should be separated by spaces, as in the example below). This method is still used to associate pictorial motifs, encoded as <interp>'s, and the <figure>'s to which they apply. For example:

On the figure element:
<figure entity="picC2a" ana="ip010171 ip010038">

In the back of the document:
<interpGrp id="ip01" 
     type="pictorial motifs">
 <!-- etc. -->
 <interp id="ip010038" value="fire"/>
 <!-- etc. -->
 <interp id="ip010171" value="candle"/>
 <!-- etc. -->

This method is no longer the preferred one because it complicates automatic processing.

8.2 Associating <interp>'s using <ana> elements

We have extended the dtd to allow the use of an <ana> element. The element's refers attribute may be used to point to a single <interp> element. If several <interp> elements need to to be associated with a single element, multiple <ana> elements should be used. This method has been used to associate some generic subject level encodings to Vaenius's emblems. For example:

On the text element:
<text type="emblem" id="va1608009">
 <ana refers="ip030002"/>
 <!-- possibly other ana elements -->
 <!-- other emblem contents -->

In the back of the document:
<interpGrp id="ip03" type="themes">
 <!-- etc. -->
 <interp id="ip030002" value="suffering"/>
 <!-- etc. -->

This method is currently the preferred one.

8.3 Associating <interp>'s using <ana> elements in pointers

In some cases <interp> elements need to be associated with, for instance, several lines of verse. In these cases <ptr> elements may be used to create a virtual grouping of these lines, cf. Referring to a Group of Locations in the Current Document. We have extended the dtd to allow <ana> elements within <ptr> elements in order to be able to associate <interp> elements with these virtual groupings.

This method has been used to associate petrarchan elements with passages from to Heinsius's emblems. For example:

In the back element of the emblem text element:
<linkGrp type="petrarchan elements">
 <ptr target="he16080032">
  <ana refers="ip04211"/>
  <ana refers="ip04271"/>
 <ptr target="he1608003205 he1608003206" targOrder="Y">
  <ana refers="ip042521"/>
 <ptr target="he16080034">
  <ana refers="ip04124"/>

The last <ptr> element in the example shows a single line of verse associated with a single interpretative encoding through the use of a single <ana> element.

In the first <ptr> element a single line of verse is being associated with two interpretative encodings by using two <ana> elements.

In the middle <ptr> element two consecutive lines of verse are being associated one interpretative encoding through the use of a single <ana> element.

8.4 Maintaining <interp> Elements

At present, <interp> elements are defined in each of the editions we have built. XML allows parts of a single XML document to be stored in an external file (entity). We really should use this facility because keeping all <interp>'s together, in a single file, would ensure consistency. However, XML software does not necessarily include these external files upon validating the documents.

A style sheet will be developed to ensure consistent <interp> definitions over multiple editions.

9 References

9.1 Referring to a Location in the Same Book

To refer to emblems or emblem constituents within the same book we may use either <ref> or <ptr> elements. The main difference is that in using a <ref> element, the editor adds the referring text (cf. the example); when using a <ptr> element, the referring text (if any) will need to be deduced by the style sheets.

The target attribute will use the id attribute of the constituent to which we want to refer. The type attribute may be used to indicate the type of relationship which is being encoded (cf. the list of allowed values in Sources).

For example:

... as may be seen from
	<ref target="va1615013">the Dutch epigram</ref>
	<ptr target="va161501"/>

This is a <ptr> element being used to point to an emblem; the style sheets will substitute a link to the emblem using the emblem's title. A <ref> element is used to refer to the Dutch epigram in the emblem. The style sheets will construct the text of the link using the words 'the Dutch epigram'.

9.2 Referring to a Group of Locations in the Current Document

Sometimes it will be necessary to refer to a range of locations. The <xref> and <xptr> elements may, in their target attribute, refer to multiple locations. This will to be employed when referring e.g. to some lines in a poem. There is an example of this in section Associating interp's using ana elements in pointers

However, when referring to multiple locations, it is more cautious to use multiple <ref> of <ptr> elements. Using a single <ref> or <ptr> element with multiple values in the target attribute will generally demand specific solutions in the style sheets and should be discussed first.

9.3 Referring to a Location in Another XML File

If we want to refer to an emblem (constituent) in an edition we have edited, or to an item in the bibliography, we will use <xptr> or <xref> elements ('extended pointer, extended reference'). The doc attribute defines the document which we refer to; it contains an name of an entity defined in the file. The from attribute defines the location in the external document. If the reference is to a location with an id attribute, then the from attribute is used as follows:

id (value-of-id-attribute)
. It is possible to refer to locations through other means than id attributes, but as yet, we have felt no need to use these.


<xref doc="va1615" 
   from="id (va1615023)">the Dutch epigram</xref>
<xptr doc="va1615" 
   type="parallel" from="id (va161502)"/>	

The first example refers to an epigram in Amoris divini emblemata. (Amoris divini emblemata is identified using the doc attribute 'va1615'; the epigram referenced in the from attribute ('id (va1615023)') which points to the element with id attribute 'va1615023' in Amoris divini emblemata).

For examples of how to refer to the bibliography see References to Sources.

9.4 Referring to Another File as a File

References to other (EPU) XML files will generally be considered references to the files' contents. When presenting these references to the user, they will be replaced by a reference to HTML content generated from the files.

When we want to refer to a qua file, for instance to the xml file containing this document, we will use an <xptr> or <xref> element with the doc attribute referring to the ENTITY which identifies the file in the extxml.dtd. The rend attribute's value file will be used to indicate the reference is made to this file. For example:

Reference to this file: 
<xref doc="techcoding" rend="file">the xml file containing 
				this document</xref>

This will probably only be useful in technical documentation.

9.5 References to Searches

From the editorial sections, it may be desirable to link to a query executed on the search page. Formal expressions which define a query on an XML file may be defined using Xpath recommendationLinks to this book in bibliography. Other XML query languages are being developed. When there is software available which can dynamically evaluate XPATH expressions, we will certainly want to use these query languages.

For the present time, we will use a simple mechanism in which we refer to a pseudo document, while passing a search string. The style sheets will deduce the correct HTML link. For example:

<xref doc="search" 
   from="foreign (book=va1615,index=pm,term=ip010034)">
      list of emblems

The example executes a search in Amoris divini emblemata for those emblems containing an anchor as a pictorial motif (execute this search). The doc attribute contains the value 'search', which has no other function than to tell the style sheets the from attribute's value should be interpreted as a search string.

The search string ('foreign (book=va1615,index=pm,term=ip010034)') will detail the book in which to search, the internal abbreviation for the index to be searched, and the term for which to search.

References to searches should use the <xref> rather than the <xptr> element.

9.6 Referring to EPU HTML Pages

9.6.1 Static HTML Pages

When we want to refer to the EPU (static) HTML pages, for example the feedback page, we will use an <xref> element, which uses the doc attribute to refer to the external source file for the HTML pages. (htmlpages.xml) Is where the from attribute identifies the <div> element in the source file which contains the HTML page's source. For instance:

Refer to the EPU feedback page from a TEI-encoded document:
<xref doc="htmlpages" from="id(feed)">feedback page</xref>

9.6.2 EPU Editions

To refer to editions of emblem books we made, for instance, to Amoris divini emblemata, we will just refer to the xml file containing the source of the edition. We will do this by using an <xptr> or <xref> element with the doc attribute referring to the ENTITY identifying the edition in the extxml.dtd file. The rend attribute's value edition will be used to indicate the reference is meant to be to the edition, not to some part of the source file. For example:

Refer to the edition of Amoris divini emblemata:
	<xptr doc="va1615" rend="edition"/>

9.7 Referring to HTML Pages

Our editions, at least initially, will be presented on the World Wide Web. One of the great things which the web offers is the possibility to hyperlink to other sites. We fully intend to take advantage of this possibility.

On the other hand, we do not want Web addresses to be used just anywhere. One reason is that the immature state of the Web causes frequent changes of address. Also, many resources will probably become available over the Web in the coming years. The way in which we refer to web pages should be manageable and easily adaptable.

We will use the following procedure:

  • Web sites are publications and as such are described in the bibliography.
  • In the bibliography, a web site like any other publication is assigned a unique id attribute.
  • All references to web sites will be through the bibliography.
  • In the bibliography, web addresses may be stored either as the the site's front page, or as a reference to a lower-level page on the site details can be found in bibliography encoding guidelines.
  • When referring to a web site the doc attribute of the <xptr> or <xref> element will be biblio, because the reference will be to the bibliography first.
  • The from attribute in the <xptr> or <xref> element will be used to point to the web site's entry (id attribute of its <biblStruct> element) in the bibliography.
  • If we want to refer to a lower-level page on the website we will add a <biblScope> element (within a <bibl> element) to the <xref> element pointing to the web site. The <biblScope> will describe the needed location on the web site in logical terms. The web address where this 'logical' location may be found will be defined in the bibliography, using an identical <biblScope> element (see the section about listing items in the bibliography encoding guidelines).
  • In this way there will be no difference between referring to a web site and referring to a location in any other book.

For example (from emblem or text edited on EPU website:

In the XML file:
<xref doc="biblio" 
    rend="ref" type="parallel" 
    from="id (alciat011995a)">
  <biblScope type="embl">161</biblScope>

In the bibliography:
<xref doc="alciat011995aembl161">
 Mutuum auxilium
  <biblScope type="embl">161</biblScope>

In the DTD:
<!ENTITY alciat011995aembl161 
    SYSTEM "http://www.mun.ca/alciato/161.html" 
    NDATA html>

In this way, the edition XML just refers to the bibliography (the 'biblio' value of the doc attribute), states which book or web site is meant ('id (alciat011995a)' in the from attribute) and names the scope in the book or site in the <biblScope> (emblem 161). The bibliography contains the same <biblScope> and in the <xref>'s doc attribute says the web site address is contained in the entity 'alciat011995aembl161'. Finally the DTD gives the true address 'http://www.mun.ca/alciato/161.html'.

9.8 References to Sources

The TEI GuidelinesLinks to this book in bibliography, 6.10.3 state a bibliographical reference should be encoded as a <ptr> or <ref> element, and not as a <bibl> element. In our case, with the bibliography in a separate document, this results in <xptr> or <xref> elements. The doc attribute will have the value biblio, the from attribute will point to the entry in the bibliography. This has been shown in several examples above. It may look like:

... the best study on this subject is 
<xref doc="biblio" 
    from="id (portem01inle)">
 the article by Porteman</xref>,
  where he demonstrates ...
... the best study on this subject is 
<xptr doc="biblio" 
    from="id (portem01inle)"/>,
  where he demonstrates ...

Somewhat surpsisingly, the Guidelines keep additional information, such as a page number, outside of the <ref> element (last paragraph in TEI GuidelinesLinks to this book in bibliography, 6.10.3). This is undesirable for our purposes, because we want to be able to trace references to the same location in a quoted work. Therefore, in this case we will deviate from the Guidelines and put a <biblScope> within the <ref> to indicate a location or scope within the work. This too has been shown in some examples above. Another one:

<xref doc="biblio" 
    from="id (alciat01embl1534)">
 Gratiam Referendam
  <biblScope type="embl">30</biblScope>

For the content and type attribute of the <biblScope> element, see the section about the <biblScope> element in the bibliography encoding guidelines)

10 Indexing

Indexes may be built, and have been built, on several encoded features (pictorial motifs, Petrarchan elements, etc). We have also built a process which automatically generates concordances of all words, by language, within the texts.

We may also decide to encode key words in the mottoes and emblem summaries. To accomplish this, words will need to be lemmatized. This may be accomplished as follows:

<p><w lemma="pietas">PIETATE</w> 
<w lemma="in">IN</w> 
<w lemma="parens">PARENTES</w> 
<w lemma="potior">POTIOR</w>.</p>

When we begin doing this, we will need rules to decide on a standard form (to be placed in the lemma attribute of the <w> word element) for each language concerned.

11 Other Encodings

11.1 Languages

The lang attribute may be used on any element to indicate the content's language. We will only use it if the element's language differs from its parent's language. Language codes consist of three letters and are based on ISO 639-2Links to this book in bibliography. At the highest level (the <TEI.2> element) we code:

Dutch language comments will need to be explicitly specify

Up until now, the following values were used:

dut Dutch
eng English
esp Spanish
fra French
ita Italian
lat Latin

The lang attribute is of type IDREF, and therefore refers to an element elsewhere which defines allowed values. For this purpose, <language> elements are stored in the <teiHeader>.

11.2 Names

We have not yet constistently tagged personal names. We choose to use the <name> element. We will use the key attribute to link the name to the already large list of personal names in the bibliography. This will give us a single list of personal names to maintain, including standard spellings of the names, without the need to use the <name>'s reg attribute. It also gives an easy way to link writers to their books, and most of the names referred to will be present in the bibliography anyway. The reference will look like this:

<name key="alciat01">Alciato</name>

For the list of <name>'s in the bibliography, see the section about referring to other elements in the bibliography encoding guidelines).

11.3 id Attributes

Some elements in the editions need unambiguous identification because we need to point at them, either from other files or from within the documents (e.g. to make clear relations between items). The TEI dtd's allow each element to have an id attribute which uniquely identifies the element within the document. The value which is used does not really matter, as long as it is unique and conforms to the limitation of the XML name datatype.

Nevertheless, because of mnemotechnical reasons, it is useful to have some conventions regarding the use of values for id attributes. In the editions, we will try to use the following guidelines:

emblems a code for the book (e.g. va1608), followed by a two or three digit serial number for the emblem
emblem constituents the code for the emblem followed by a hierarchically constructed number for the <div>'s. Cf. the examples in Emblem Structure. Because all <div>'s in the emblems will have an id, we do not need to assign id attributes to <lg> elements (there'll be only one <lg> within any <div>).
highest-level <text> element the book code (e.g. va1608)
verse lines the id of the higher-level constituent followed by 'l' and a line number.
<pb> 'pb' followed by the page number
However, if the <pb>'s n attribute contains a signature containing, for instance, '(:)4r', in the id attribute we may use something like 'pbcolon4r'.
<interp>/<interpGrp> 'ip' followed by a six digit number which shows the hierarchy involved (see the examples in Motifs and Their Categorisations).

11.4 Emblem Titles

The mottoes will be transcribed as they appear in the source (capitalised and in original spelling). Therefore, they are not very useful as titles or as referring text. We decided to create a separate <div> in the document <back> (with type titlelist) where useful titles are defined. In this <div> a <linkGrp> with <link>'s will associate emblem <text>'s and their titles. For example:

Title list:
 <item id="title001">Amor &aelig;ternus</item>
 <item id="title002">
  Perfectus amor non est nisi ad unum

<linkGrp type="reftitle" targType="text item" 
       targOrder="Y" targFunc="emblem reftitle">
 <link targets="va1608001 title001">
 <link targets="va1608002 title002">

When showing the editions, these titles should be clearly indicated to be editorial additions.

11.5 Numbers in Mottoes

In Heinsius's Emblemata Amatoria the mottoes contain an emblem number. It is desirable that these are marked up explicitly. We rejected the possibility to put the number in the <text>'s n attribute (or that of the corresponding <div>) because this hides how the number was present in the source.

We code the number (with following spaces and interpunction) in a <num> element. For example:

<div ... id="he16080011">
	<p><num type="textno">1. </num>Omnia vincit amor.</p>

12 To be Added or Modified in This Document

13 Practical Details

Each emblem book will be encoded as a single TEI-conformant XML document. The document name consists of the book identifier, followed by 'v' (version), then a version identifier. The extension used is '.xml'. For example:

Name of file containing Amorum emblemata:

13.1 Document Type Declaration

The document type declaration in the xml files will look like this:

Document type declaration for Amorum emblemata:
<!DOCTYPE TEI.2 SYSTEM "dtd/vaentei.dtd" [
	<!ENTITY % notations SYSTEM "dtd/notations.dtd">
	<!ENTITY % charrefs SYSTEM "dtd/charrefs.dtd">
	<!ENTITY % va1608figs SYSTEM "dtd/va1608figs.dtd">
	<!ENTITY % extxml SYSTEM "dtd/extxml.dtd">

The document type definition itself is the file vaentei.dtd. Several dtd fragments are copied into the document type declaration through the use of parameter entities:

notations.dtd gives NOTATION definitions for external entities
charrefs.dtd gives ENTITY definitions for character references
gives ENTITY definitions for figures used in this edition
extxml.dtd gives ENTITY definitions for external XML files

Note the absence of a file containing entity definitions for HTML pages. As stated above, these references will always be made through the bibliography.

13.2 Indentation

Wherever this is possible and helps readibility (without introducing undesired whitespace) the XML files will indent using the tab-character.