The <code>gaiji</code> module Christian Wittern Brown University Women Writers Project
cwittern@zinbun.kyoto-u.ac.jp

Copyleft 2006 Christian Wittern

 Stanford Humanities Center

This is the source.

 automatically converted from presentation.odd conforming to yaps.odd conforming using p2y.xslt and p2y.perlMerged the two change sets below into this one document, also keeping 1 slide CW had deleted; fixed whitespace in the examples, Some content correctionsMinor typo correction
Characters and glyphs In a document printed on paper, a character and the glyph representing it form a single unit that can not be further divided While transcribing such a document with a computer, these have to be split in apart: in the abstract unit that is represented internally with a codepoint: the character (文字) in the shape that is associated with this codepoint: the glyph (字形) There are usually some generic shapes associated with a character, but they can vary considerably Two problems follow from this: Sometimes there is no character that can be used At other times, we want to identify the exact form that was used in the text In this unit, we investigate solutions to both problems
Representing a character that is not in Unicode

The represenation has two components: The definition of the character (in the teiHeader) The invocation of the character somewhere in the text These are available in the gaiji module
The definition of a new character

The TEI provides the following for defining a new character: In charDesc of the encodingDesc section in the teiHeader is a list of character definitions: The char holds a definition of a character a charName is a required child Additional properties of the character can be defined using charProp The invocation of the character somewhere in the text Additionally, a graphic representation of the character can be given in graphic A mapping to another character can be defined, a type is used to indicate the type of the mapping according to some typology.

More information can be found in Representation of non-standard characters and glyphs (which is currently Chapter 25 of the TEI Guidelines)

The character thus defined can then be invoked using the g element, the required ref has to point to the definition in the header.
Example: A character defined using the TEI gaiji module

Here is a character with its name and three properties: CBETA CHARACTER CB02596 composition [(禾*尤)/上/日] Mojikyo number M025240 entity CB02596 稽 ]]>

Differentiating between different glyphs of the same character

Sometimes, the character is in Unicode but with a different shape (異體字)

In this case, a similar mechanism can be used: Define a character in the header Give the intended shape there (as a graphic or point to a font file) Use the standard Unicode character in the text, but within a g element When rendering the text, the graphic can be used to display the character When indexing the text for search, the standard form can be used.

Sometimes there are more than one glyph of the same character in Unicode. In this case, the same method can be used, but additionally, the mapping can be used to indicate the relationship between these glyphs.
Associate a specific glyph with a character

Sometimes, it is desirable to associate a character with a specifc glyph instead of the set of generic glyphs which is the default in Unicode. In this case, a similar mechanism to the one above can be used, but it differs in two important points: Instead of a char, now define a glyph in the header of the document, the intended shape of the glyph can be stated by pointing to a font file or giving using a graphic element. In the text, the gelement is used, but now the standard character is used as its content.

This example used Ideographic Definition Sequences (IDS) as defined by Unicode to indicate the shape of the glyph: Variant of CJK U+884C IDS ⿰⺅亍 ]]> In the text, this would be invoked as follows: 行，北頭第一行]]>

As a result, the text can be used in different ways: When rendering the text, the graphic can be used to display the character When indexing the text for search, the standard form can be used.

Create a mapping between related characters

A slightly different problem arises from the fact that in some cases Unicode encodes not only characters, but also glyph variants. The following: 説 (U+8AAC) and 說 (U+8AAA) are different glyphs of the same character, but they each has its own separate Unicode value. Normally, only one of these should be used in a given text. Howeve, if for some reason it is necessary to use both, a mapping between these can be established. A similar method can be employed for other cases where one character has used in a ideosyncratic way as a replacement for another (異體字). To establish the fact that we consider these two as variants of the same character we could define the following in the header: Unified CJK 8AAC 說 ]]>