[H O M E : Emblem Project Utrecht]

Emblem Project Utrecht

EPU Guidelines for encoding the bibliography

1 Introduction

These guidelines will be reviewed in this document:

1.1 Examples In The 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 make the examples legible extra tabs, spaces, or linefeeds may have been added.

Table of contents

2 Why Use a Bibliography?

The emblem book editions, which we intend to make, will need to point out specific literature. If bibliographic information is added into the editions themselves, it will cause repetition of a book’s details each time the book is referred to. Consequently, the duplication of information would probably be incorrect, because it would be copied several times. This process would be tedious and time consuming without a bibliography. want to bother with all of the detail each time we need to refer to the book.

The simple solution is to store all bibliographical data about the literature we have consulted in one place. When we need to refer to bibliographical data we will just point to this central location. As well, the bibliography can account for the literature we used in drawing our conclusions, allowing the reader to check our conclusions.

A single source of bibliographical data will enable us to do several other things.

3 What Should Be In The bibliography?

The bibliography should contain:

Web sites, CD ROM’s and microfilms are published items like any other. If we need to refer to them, they will be described in the bibliography.

The bibliography may contain items which we have not been seen or studied yet.

The level of detail in the description of bibliographical items should be sufficient to localise and identify the items. The bibliography does not aim at full description of source material.

4 Global Technical Principles

Besides the functional requirements outlined above, the following principles are also adhered to:

5 Technical Details

As stated above, we try to follow the Text Encoding InitiativeLinks to page outside of EPU. Where applicable, we will refer to the appropriate location in the Guidelines for more information.

5.1 TEI Elements: <bibl>, <biblStruct> and <biblFull>

According to the TEI Guidelines, three elements may be used to encode bibliographical information1. In the bibliography, we will use the <biblStruct> element to group the information pertaining to a single item. The same information might be encoded using the <bibl> element. However, the <bibl> element may contain text besides subelements like <title> or <author>. As a result, there is no guarantee its contents will be sufficiently structured for automatic processing.

The <biblFull> element is strictly structured and seems geared towards extensive description of source features, rather than the simple bibliographical information we are interested in. As well, the <biblFull> can not contain the <analytic> elements needed to describe journal articles.

5.2 TEI Elements: <analytic>, <monogr>, <series>

When describing books, articles, or a series we will follow the procedures outlined in Text Encoding InitiativeLinks to page outside of EPU, 6/10/2/1. Articles will be encoded as <analytic> elements; the journal or book containing the article will be described in a <monogr> element following the <analytic> element, both will be subordinate to the single <biblStruct> element. Where applicable, a <series> element may follow the <monogr> element to indicate the series in which the book was published.

To avoid repeating the same information several times the procedure will be streamlined, this is outlined below.

If we want to record that an article was published more than once, we will encode this using more than one <monogr> element within a single <biblStruct> element.

5.3 Scopes Within Bibliographical Items

A scope within a bibliographical item such as an article within a journal, or a page range in a book, will be encoded as a <biblScope> element2. We will use the <biblScope> element’s type attribute to encode, in detail, the meaning of the scope being described. For example:

<biblScope 
  type="ccyy/inst/mon/pb-pe">1969/184/dec/343-346</biblScope>

In the type attribute we will use the following values, separated by slashes:

band band
bk book (Latin: liber)
A reference to Book 10 within a bibliographical item will be encoded as
<biblScope type="bk">10</biblScope>
ccyy year (including century)
ch chapter (Dutch: hoofdstuk, Latin: caput)
col column (use plc for columns in the Patrologia Latina).
ed edition (used for referring to a specific version of the Bible)
embl emblem
Reference to an emblem within a work, if there is an unambiguous emblem number
epist letter, epistle
inst installment
instb-inste installment begin to installment end
Used for scopes in journals where issues have been joined. For example:
<biblScope type="vol/ccyy/instb-inste/pb-pe">
 56/1998/3-4/2-3
</biblScope>
item item in a list or collection of some kind (perhaps a bibliography)
itemb-iteme item begin to item end (to refer to a range of items)
l line
lb-le line begin to line end
Used to refer to a passage of verse
l1 level one subdivision of text where no more specific name is available
l2 level two subdivision of text where no more specific name is available
l3 level three subdivision of text where no more specific name is available
l4 level four subdivision of text where no more specific name is available
l5 level five subdivision of text where no more specific name is available
l1b-l1e level one subdivision range
l2b-l2e level two subdivision range
l3b-l3e level three subdivision range
l4b-l4e level four subdivision range
l5b-l5e level five subdivision range
mon month
For example:
<biblScope type="ccyy/inst/mon/pb-pe">
 1969/184/dec/343-346
</biblScope>
monb-mone month begin to month end.
note note
p page
pb-pe page begin to page end
pp unstructured page range (to be avoided)
E.g.:
<biblScope type="pp">13, 14-17</biblScope>
plv Volume in the Patrologia Latina
plp Page number in the Patrologia Latina
plpb-plpe Page number begin to page number end in the Patrologia Latina
plc Column letter (in capitals) in the Patrologia Latina
plcb-plce Column letter begin to column letter end (in capitals) in the Patrologia Latina
poem poem (in a collection of poems)
pt part (Latin: pars, Dutch: deel)
sect section (subdivision of chapter)
sectb-secte section begin to section end
ser series (Dutch: reeks)
sig sigil
sonn sonnet (in a collection of sonnets)
toappear indicates an article will appear in a journal at an unspecified future moment. The corresponding <biblScope> contents will be dummy.
For example:
<biblScope type="ccyy/toappear">2002/dummy</biblScope>
tract treatise, tractatus
unknown unknown (to be avoided)
v verse (e.g. in referring to the Bible)
vb-ve verse begin to verse end (to refer to a range of verses)
vol volume

5.4 References to The Bible

For the present, a single item in the bibliography refers to the Bible:

<biblStruct id="bible" ana="interpbco">
 <monogr>
  <title type="main">Bible</title>
  <title type="parallel">Bijbel</title>
  <title type="short">Bible</title>
  <imprint/>
 </monogr>
 <note type="webaddress">
  <xptr type="websitestart" doc="biblestart"/>
 </note>
</biblStruct>

The website we refer to enables direct access to a passage in the Bible by book and verse number. It also allows the reader to select from various translations of the passages.

The sites disadvantage’s are that there is no Dutch translation available, and that apocryphal books are missing. An advantage over other sites is that it is mindful of differences in verse, and chapter numbers between the Vulgate and other traditions.

When referring to a passage in the bible, we will use the standard modern chapter and verse numbers. The values to be used in the type attribute of the <biblScope> element are described in the table above: bk, ch, v, vb-ve, ed. For example:

To refer to Deut. 6
<xref doc="biblio" rend="ref" type="source" from="id (bible)">
 <bibl><biblScope type="bk/ch">de/6</biblScope></bibl>
</xref>
To refer to Deut. 6:1
<xref doc="biblio" rend="ref" type="source" from="id (bible)">
 <bibl><biblScope type="bk/ch/v">de/6/1</biblScope></bibl>
</xref>
To refer to Deut. 6:1-2
<xref doc="biblio" rend="ref" type="source" from="id (bible)">
 <bibl><biblScope type="bk/ch/vb-ve">de/6/1-2</biblScope></bibl>
</xref>

The ed value may be used (optionally) to indicate which version of the Bible we refer to. The versions used at present include:

vulg Vulgate
nonvulg non-specified non-Vulgate version

However, even when explicitly referring to the Vulgate, we will use modern chapter numbers. For example:

To refer to Ps. 119:8 in the Vulgate Bible 
 (there known as Ps. 118:8):
<xref doc="biblio" rend="ref" type="source" from="id (bible)">
 <bibl><biblScope type="bk/ch/v/ed">ps/119/8/vulg</biblScope></bibl>
</xref>

The Bible's book information is always in lower case. The abbreviations used for the Bible books were borrowed from the Online Bible. They are the following:

ge Genesis
ex Exodus
le Leviticus
nu Numbers
de Deuteronomy
jos Joshua
jud Judges
ru Ruth
1sa 1 Samuel
2sa 2 Samuel
1ki 1 Kings
2ki 2 Kings
1ch 1 Chronicles
2ch 2 Chronicles
ezr Ezra
ne Nehemiah
es Esther
job Job
ps Psalms
pr Proverbs
ec Ecclesiastes (Dutch: Prediker)
so Song of Solomon
isa Isaiah
jer Jeremiah
la Lamentations
eze Ezekiel
da Daniel
ho Hosea
joe Joel
am Amos
ob Obadiah
jon Jonah
mic Micah
na Nahum
hab Habakkuk
zep Sephaniah
hag Haggai
zec Zechariah
mal Malachi
mt Matthew
mr Mark
lu Luke
joh John
ac Acts
ro Romans
1co 1 Corinthians
2co 2 Corinthians
ga Galatians
eph Ephesians
php Philippians
col Colossians
1th 1 Thessalonians
2th 2 Thessalonians
1ti 1 Timothy
2ti 2 Timothy
tit Titus
phm Philemon
heb Hebrews
jas James
1pe 1 Peter
2pe 2 Peter
1jo 1 John
2jo 2 John
3jo 3 John
jude Jude
re Revelations
3ezr 1 Esdras (3 Ezra)
4ezr 2 Esdras (4 Ezra)
tob Tobit
jdt Judith
est Additions to Esther
wijs Wisdom of Solomon
sir Ecclesiasticus (Jesus Sirach)
bar Baruch
aza Prayer of Azariah
sus Susan
bel Bel
man Prayer of Manasses
1ma 1 Maccabees
2ma 2 Maccabees
3ma 3 Maccabees

5.5 The People Involved in Producing a Book

5.5.1 Authors and editors

An authour, according to <author>’s Links to this book in bibliography, 6.10.2.2 quidelines, are those with primary responsibility for the work’s intellectual content and any others involved in it’s production. For those who bear secondary responsibility the <editor> or <respStmt> elements may be used.

According to Text Encoding InitiativeLinks to page outside of EPU, 6/10/2/2 the <respStmt> element should be used for those with secondary responsibility that cannot be considered editors. However, the definition of the <editor> element in Text Encoding InitiativeLinks to page outside of EPU, 35/editor allows translators to be encoded as <editor>’s; also the definition of the role attribute on the element gives illustrator as a possible value.

The <editor> element is also more staightforward to use than the <respStmt> element, for this reason all persons are encoded with secondary responsibility as <editor> elements. The role attribute was then used to indicate the nature of their responsibility. The allowed values for the role attribute are:

  • editor (the default)
  • webeditor (a person who prepared the web publication of a work)
  • translator
  • engraver
  • drawings
  • foreword (writer of foreword or introduction)
  • parttrans (partial translator)
  • annotator

The list may be extended if the need arises. Preferably, an editor is given a single role.

5.5.2 More Than One Author or Editor

Each author or editor is encoded using a separate <author> or <editor> element (instead of using several <name> elements within a single <author> or <editor> element). There seems to be no explicit statement about this in the TEI Guidelines. The examples in Text Encoding InitiativeLinks to page outside of EPU, 6/10/2/3 and Text Encoding InitiativeLinks to page outside of EPU, 6/10/2/5 conflict. For the sake of simplicity we chose to use a separate element for each person involved.

5.5.3 Title Page Information about the Authors

In many cases, books have authors as well as information explicitly stated about who wrote them. In fact on the title page, they usually do. Vaenius's Amorum emblemata, for instance, says on the title page 'studio Othoni Vaenii'. Usually, we will not record these exact wordings in our bibliographical data. If the need arises, they may be encoded as <respStmt> elements.

5.5.4 Uncertainty about Authorship

A book’s authorship is sometimes unknown or debatable. We have added a status attribute to the <author> element to account for this situation. Its allowed values are certain (the default) and debated. For example:

<biblStruct id="august011845t" ana="interpbce">
  <monogr>
    <author status="debated"><name corresp="august01"/></author>
    <author status="debated"><name corresp="hugost01"/></author>
    <title type="main">De substantia dilectionis</title>
    <!-- etc. -->
  </monogr>
</biblStruct>

5.5.5 Corporate Authors

Corporate authors will be handled in the same way as regular authors; they will have an entry in the list of persons (cf. Referring to Other Elements), and on the book level they will be encoded as <author>’s.

5.5.6 Variant Spellings and Languages of an Author’s Name

The bibliography aims to identify the writers or editors involved in publication; it does not aim to record the exact spelling of their names in each publication. For each person, one spelling will be selected and associated with all of the publications he or she has been involved in. If significant variants exist (e.g. 'Otho Vaenius' and 'Otto van Veen'), both will be recorded in the list of names (cf. Referring to Other Elements). Those which will not be used will refer (using the corresp attribute) to the name which will be used in the bibliography. For example:

<list>
 <!-- etc. -->
 <item>
  <name id="schoon01" 
     type="pers">Schoonhoven, Florent van</name>
 </item>
 <item>
  <name id="schoon02" 
     corresp="schoon01" 
     type="pers">Schoonhovius, Florentius</name>
 </item>
 <!-- etc. -->
</list>

In this case, the items in the bibliography will refer to schoon01, and not to schoon02.

5.5.7 Publication Data

Data about a book’s publication is encoded in the <imprint> element. The <imprint> element may contain only structured information. We will use the <pubPlace>, <publisher> and <date> subelements. If a date is missing, we will use s.a., and if a place is missing we will use s.l.. The date (or year) of publication will be encoded in Arabic numerals, the place will be encoded using the modern English name.

Again, we do not intend to fully represent the publication statements as they may appear in the sources (as in Vaenius: 'Antverpiae. Venalia apud Auctorem. Prostant apud Hieronymum Verdussen. M. DC. IIX').

5.6 Referring to Other Elements

In a bibliography, often the same data has to be presented several times. An author’s name may be repeated on several items, or a journal title will be referred to in the description of several articles, etc.

It is clearly undesirable to repeat the information in the bibliography document(s). The main reasons to avoid this are:

  • Repeating the data allows for making mistakes, if not in entering the data, then in changing them. Unavoidably names (etc.) which were meant to be identical will end up being different.
  • Even if mistakes are avoided, having identical names will not make it clear to the reader that these names were meant to be identical and therefore, refer to the same person.

Wherever possible, we will not repeat information but rather refer to separate lists of names or titles containing the needed information.

The Guidelines which describe the following facilities to achieve this desired effect:

  • The copyOf attribute may be used on any element. It’s value identifies another element within the same document. Use of the copyOf attribute is meant to state that the element’s content is equal to the content of another element that is referred to. In effect, it creates an exact copy of this element. (Text Encoding InitiativeLinks to page outside of EPU, 14/6). Using the copyOf attribute avoids entering the same information in several locations. However, the attributes semantics are not quite what we have intended. Using the attribute says that both elements should have identical content, but does not say that the elements describe the same object.
  • The same objection holds for the sameAs attribute. It is used in the same way as the copyOf attribute and indicates the element content is actually equal to the content of the element to which it refers.
  • The corresp attribute, 'points to elements that correspond to the current element in some way.' Text Encoding InitiativeLinks to page outside of EPU, 14/4/1). For the application we need, 'correspondence in some way' is certainly applicable. The disadvantage of the corresp attribute is its vagueness. To an outside observer the kind of correspondence the attribute is meant to convey, is not always immediately clear.
  • The type of correspondence that needs to be established between elements of any kind, may be expressed explicitly using the <link> element. On the <link> element the type of correspondence (coreferentiality) may be indicated using the type atribute.
    This method, though effective, is rather cumbersome because the elements which are being linked both need to have id attributes.
  • Specifically for <name> elements there exist the key and reg attributes. The reg attribute makes it possible to add a regularised spelling to a <name> element. For our purposes it is not helpful because we would like to store regularised spelling only once, and not with each occurrence of the name.
    The key attribute is meant to contain a type of database key identifying the intended person, which is exactly what we would like to di. The disadvantages of using the key attribute are that it is unavailable on other elements (<title>'s); as well the attributes value is of type CDATA (rather than IDREF), which means that errors in referring to names will not be discovered when validating the document.

Weighing the advantages and disadvantages of the solutions outlined above, we have decided to encode the relation of coreferentiality between elements in the bibliography using the corresp attribute.

This mechanism will be used to associate personal names with items in a list of personal names, to quote journal titles in the description of articles and to quote series titles in the description of books. In descriptive content we may associate mentioned titles with their own entry in the bibliography.

An example of a description of a journal article:

<biblStruct id="portem011975a" ana="interpbcv">
 <analytic>
  <author><name corresp="portem01"/></author>
  <title type="main">Miscellanea emblematica</title>
 </analytic>
 <monogr corresp="spieglet"><imprint/>
  <biblScope 
     type="vol/ccyy/inst/pb-pe">17/1975/3/161-193</biblScope>
 </monogr>
</biblStruct>

You can see that the <author> element is referring to a personal name elsewhere in the bibliography. The list of personal names will contain:

<item>
 <name id="portem01" type="pers">Porteman, Karel</name>
</item>

The <monogr> element also refers to a journal listed in the in the bibliography:

<bibl id="spieglet"><monogr>
 <title level="j">Spiegel der letteren</title>
 <imprint/>
</monogr></bibl>

The two empty <imprint> elements used in the <monogr> elements are there to conform to the TEI dtd which demands their presence.

5.7 Categorisation, Keywords and Subjects

It might be desirable to add keywords to some or all items in the bibliography, and to categorise the items in one or more respects. When the bibliography grows in size, this may help in keeping it manageable and to locate relevant items.

The TEI <keyword>-mechanism is only available in the <teiHeader>. To encode categorisations and keywords we will therefore use lists of <interp> elements, grouped in <interpGrp> elements which indicate the relevant category. The bibliographical items will use the ana attribute to point to the <interp> elements. At present, this mechanism is used to broadly classify the nature of an item such as: a bibliography, (edition of) source material, emblem study, a technical study or another study. We have also used the mechanism to describe an item’s media type.

Subjects like persons or works, which are discussed in secondary literature will often be entries in the bibliography themselves. Relationships of this nature may be encoded using the corresp attribute on the <name> or <title> element, as discussed above.

5.8 Titles

A <title>’s level (level attribute) is given implicitly by the <title>’s use within the <analytic>, <monogr> or <series> element. It will not be used, except on journal titles (level attribute with value j), to distinguish journals from other <monogr>’s.

The type attribute will be used as follows:

main main title of work (mandatory). May occur only once.
sub subtitle (optional). May occur more than once.
parallel parallel title (optional). May occur more than once, it can also be used for works which have two titles, perhaps in two languages. If a parallel title occurs, it will be considered as an alternative for the main title, not the subtitle.
short a brief title added by us (recommended but optional). The short title will be used in constructing hyperlinks and other references to the item. It should contain the author’s last name and omit punctuation at the end of the title. It is not necessarily equal to the half-title or the title which appears on a book’s spine.

Titles of other works mentioned within the item’s description, as part of the title or in a description-type note, will be encoded as plain <title> elements. If the work which the title refers to is described within our bibliography, the corresp attribute will be used to associate the title to the relevant entry. If the bibliography describes several editions of the work being referred to, a reference will be made to the first included edition.

An example of a cross reference to a title from a <note>:

From the description of 'Princeliicke Deviisen':
<note type="description">Digital version of the Dutch edition, 
  which appeared in 1615. [...] Text, facsimile, 
  iconclass encoding of images. Hyperlinked to 
  <title corresp="paradi012000b">the French version</title> and 
  <title corresp="paradi011998a">the English version</title>.
</note>

An example of a cross reference to a title from another <title>:

<title type="main">De 
  <title corresp="veen--01amor">Amorum Emblemata</title> van 
  <name corresp="veen--01">Otho Vaenius</name> (Antwerpen 1608) 
  in de Britse en Duitse emblematiek
</title>

This last example also shows how a personal name should be cross-referenced from a title.

Using the titles in the bibliography, style sheets will construct hyperlinks and other references to the books in the bibliography. Generally, it will be undesirable for these references to end with a punctuation character. If a main <title> ends with a punctuation character which should be recorded in the bibliography, then it is desirable to add a <title> with type short without the final punctuation.

5.9 NOTES

<BiblStruct> elements may contain one or more <note> elements. These elements may be used for the following purposes:

5.9.1 Private Notes

Private notes are identified using the private value on the <note>’s type attribute. These notes are meant to contain internal project information which will not be made publicly available. This might include unstructured information about the bibliographical item or information about who owns a copy of an article, etc. The resp attribute will be used to indicate who entered this information.

5.9.2 Description Notes

Description notes are identified using the description value on the <note>’s type attribute. These notes are meant to contain a brief description of of an item or annotation to it, the notes should tell the reader what kind of information the book or article contains.

5.9.3 Catalog Notes

Catalog notes are identified using the catalog value on the <note>’s type attribute. These notes contain unstructured bibliographical information about the item, taken from for example, a University library’s catalog. This may be information which we have not independently verified. The resp attribute will give an abbreviation for the catalog ('NCC', 'BNTL') or the id attribute of the bibliography from this, stems information contained in the <note>.

5.9.4 Contents Notes

Content notes are identified using the contents value on the <note>’s type attribute. Content notes may contain a partial list of items (e.g. emblems) contained in the work being described. They may also contain references to parts of the work which discuss some aspects relevant to our editions.

5.9.4.1 Contents notes: listing items

A <note> of type contents may list items in the work being described using a <list> with <item>’s optionally containing an <xref> element if there is an online source for the relevant item. Some examples:

Including reference to an online resource:
<list>
 <item>
  <xref doc="hugo--01nyap2">
   Lord, thou knowest all my desire [...]
   <bibl>
    <biblScope ana="interpembl" type="p">2</biblScope>
   </bibl>
  </xref>
 </item>
 <item>
  <xref doc="hugo--01nyaembl1-3">
   Have mercy upon me, O Lord, for I am weak [...]
   <bibl>
    <biblScope ana="interpembl" type="bk/embl">1/3</biblScope>
   </bibl>
  </xref>
 </item>
 <!-- etc. -->
</list>
Without reference to an online resource:
<list>
 <item>
  De Ziel smeekt [...]
  <bibl>
   <biblScope ana="interpembl" type="embl">11</biblScope>
  </bibl>
 </item>
 <item>
  De Ziel, rustende [...], waakt [...]
  <bibl>
   <biblScope ana="interpembl" type="embl">14</biblScope>
  </bibl>
 </item>
 <!-- etc. -->
</list>

These are <item>’s which refer to emblems in emblem books. The <biblScope> element details the location of the emblem within the book. Its ana attribute may give the value interpembl to identify the item as an emblem, this is information which is used in formatting the 'parallels and sources' information on the emblem pages.

The <item>’s (or if applicable, <xref>’s) text content provides a title for the item. If there is an <xref>, its doc attribute identifies the web page where the item may be found.

5.9.4.2 Contents notes: indexing

A <note> of type contents can index part of the work’s contents. This may help our project as well as others, in locating relevant literature. An example can be found in the the description of Links to this book in bibliography:

<list>
 <item>
  <name corresp="cats--01">Cats</name>: 
  <bibl>
   <biblScope ana="interpdisc" type="pp">13, 69, 83</biblScope>
  </bibl>
 </item>
 <item>
  <name corresp="heinsi01">Heinsius</name>: 
  <bibl>
   <biblScope ana="interpdisc" type="p">83</biblScope>
  </bibl>
 </item>
 <!-- etc. -->
</list>

These are the <item>’s which associate personal names with scopes in the work being described, through the use of a <biblScope> element. The ana attribute has the value interpdisc to help the style sheets in formatting this information.

5.9.5 Web address notes

Web address notes are identified using the webaddress value on the <note>’s type attribute. The note will contain at least one <xptr> element of type websitestart. The <xptr> will give the address of the web site’s opening page. For example:

<note type="webaddress">
  <xptr type="websitestart" 
     doc="sluijt012001astart"/>
</note>

5.10 id Attributes

Some elements in the bibliography need unambiguous identification because we need to point to them, either from other files like our editons, or from within the bibliography (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 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 bibliography, we try to stick to the following guidelines:

persons first 6 positions of last name (padded with hyphens, if necessary, then a number in two digits. Examples:
<name id="hugo--01" 
   type="pers">Hugo, Hermann</name>
<name id="muller01" 
   type="pers">M&uuml;ller, Wolfgang J.</name>
<name id="muller02" 
   type="pers">M&uuml;ller Hofstede, Justus</name>
It is apparent in the example that diacritical signs on letters are ignored when taking the first six letters.
series use sensible abbreviation
journals use sensible abbreviation
bibliographical entry (<biblStruct>) id attribute identifying first author or editor, then the year of publication, then a letter. Examples:
alciat012000a
portem011975b
hugo--011971a
Many <biblStruct> elements still have an id attribute consisting of first author or editor, followed by the first four letters of the title. This has proved to be less than satisfactory, because many titles in the bibliography are quite similar.

5.11 Media

The TEI Guidelines do not recommend a structured way of encoding the media type (paper, CD Rom, web site or microfilm) in a bibliographical description. The Text Encoding InitiativeLinks to page outside of EPU, 6/10/2/5 suggests using <note> elements.

We chose to use a more structural encoding, using the categorisation mechanism described above. There will be a list of media types encoded as an <interpGrp>; the <interp> elements will be associated with the bibliographical entries using the <biblStruct>’s ana attribute. If no such association is made, it will be assumed that the entry refers to a publication on paper.

5.12 Web Sites

Web sites will be encoded using <monogr> elements. The web site’s <title> will be taken from, in descending order of preference, (1) Dublin Core metadata embedded in the web page, (2) the HTML <title> element (if this gives the impression of being used with care), or (3) a title visually presented on the web site’s opening page.

Other data such as: <author>, <editor>, and <imprint> information may also be found inspecting the site’s HTML source. The <publisher> is the person or institution that owns the web site.

For each web site the opening page is encoded in a <note> element with type webaddress, using an <xptr> element (cf. Web address notes). If desired, addresses of other pages on the site may be recorded using a <note> element with type contents (cf. Contents Notes).

5.13 Relations Between Items in the Bibliography

Between items in the bibliograpy, several kinds of relations may exist. For instance:

  • work mentions or discusses other work
  • work is a reprint of other work
  • work is or contains material form other work
  • work contains partial or complete facsimile of other work
  • work contains partial or complete translation of other work

We have not yet attempted to systematically enter this kind of information in the bibliography. Using <title>’s with a corresp attribute in the description-type <note>’s relations between items may be encoded, but not classified.

There is one kind of relation between items which we do store in the bibliography, when work is an edition of other work. To this end, a <linkGrp> element contains <link>’s between <biblStruct>’s where the first element is an edition of the second one. For example:

<linkGrp type="isEditionOf" 
    targOrder="Y" targFunc="edition original">
  <link targets="ayres-011998a ayres-011683a"/>
  <!- etc. -->
</linkGrp>

6 Some Issues That Have Not Been Satisfactorily Resolved

7 Practical Details

The bibliography is a TEI-conformant XML document. The upper level <text> element consist of a <body> with a number of <div>’s (divisons). These divisons are: