Letters are the basic unit in alphabetic script encoding and transcription, and opering with Leabhar Cheanannais source is necessary to distinguish between characters and glyphs: as UNICODE defines it and MENOTA Guidelines remember us characters are “the smallest components of written language that have semantic value”, while glyphs are “the shapes that characters can have when they are rendered or displayed” (cf. Unicode 4.0, ch. 2.2).

So process of encoding and trascribing Kell's text is such operation of see a series of individual glyphs in source document, recognizing them and connecting these glyphs to the characters at mine disposal

Unicode standard based on assumption that a great number of glyphs should be considered as different representation of the same character, so Unicode just distinguish between small (minuscule) and large (majuscule) characters, such as "a" and "A"; Unicode does not regard italics (or bold type) as individual characters

Medieval manuscripts were written in the Latin alphabet from the very beginning and the basic inventory is thus the characters a-z / A-Z, supplemented with a number of new or borrowed characters, several ligatures and a variety of diacritical marks.

There was also a large number of abbreviation marks in use, especially in Old Icelandic and Old Norwegian manuscripts (Cfr. Menota Handbook 2.0, ch. 2): some abbreviation marks behave as ordinary characters in the sense that they occupy a separate position on the base line, but, on the other hand, many components of ordinary characters are diacritical, i.e. placed above or through or below another character, and thus akin to typical abbreviation marks; this means that rules for transcribing ordinary characters and abbreviation marks should be somehow identical.

Diacritical marks are often seen as forming an integral part of a base line character and the whole encoded as a single character. This applies to accent marks, such as the one above “e” in fig. 2.3. This combination of a base line character and a combining mark can be encoded as a single character, in Unicode referred to as LATIN SMALL LETTER E WITH ACUTE and the hexidecimal code value 00E9. As we shall see below, it is possible to decompose this letter in Unicode and refer to it as a combination of LATIN SMALL LETTER E and COMBINING ACUTE ACCENT. We would like to emphasize that both encodings are strictly equivalent.

Abbreviation marks, on the other hand, are usually treated as separate characters and encoded as characters in their own right. From a purely graphical point of view, the distinction between the acute accent in “é” and abbreviation marks such as the “zigzag” mark and the bar, both exemplified in fig. 2.3, is far from obvious, but the semantics are different. The acute accent may in some manuscripts be used to signify length, but it is often used quite freely, sometimes only to distinguish one minim character from another. Abbreviation marks have a definite (if sometimes ambiguous) meaning and can be expanded into one or more characters; the zigzag mark above “h” in fig. 2.3 signifies “er”, and the bar above “n” signifies another “n”.

Following TEI p5 Guidelines rules and Menota guidelines suggestion I ecnoded each characters firstable according to the position and the dirction of writing.

If character belonged to ordinary Latin alphabet set a-z or A-Z I always encoded as such, while character outside Basic Latin range and non Unicode Characters I always encoded by Unicode codepoints and by entities.

Alphabetical characters with diacritical marks, e.g. “é” if there were Unicode codepoint available I encoded by directly use of Unicde codepoint (here “é”), while if tehre weren't any codepoint available then I encoded as a base line character + one or more combining marks. Thus the character “é” would be encoded as “e” + “&combacute;” here the latter entity meaning COMBINING ACUTE ACCENT.

By using entities it is possible for me to define as many characters as I believes are necessary for transcription, and to update my entities corpus on the fly.

By encoding specification <?xml version="1.0" encoding="UTF-8"?> all characters in the Unicode Standard can be encoded with their Unicode code points (as here I encoded "<" and ">" character), without resorting to entities.

According to TEI p5 mandatoria I declared entities in a separate list named characters.txt, a file which constitute a complete updatable list of Kell's characters, in future available for other encoding pourposes; I linked each entity to code points defined in the MUFI character recommendation, as made and suggested in Menota guidelines, so that if my Kell's project is displayed with a fully compliant MUFI font, all entities will be displayed correctly.

Here small e.g. entities declaration, relative to Folio 34 rectum or Chi Rho page encoding:

Please notice entity declaration of <!ENTITY ndesdes "&#xe19;"> - as character "&ndesdes;" retrived from Unicode Thai sets, and entity <!ENTITY tedes "&#x4ad;"> for character "&tedes;" retrived from Cyrillic range of Unicode set.

Fonts Junicode or other fonts sets necessary for correct and complete browsing of Kell's characters should be retrived from MUFI fonts recomandation page; please use Mozilla Firefox to browse my Kell's pages.

Please notice that may some fonts aren't browsable correctly or may not dislpayed at all because of language sets extensions missing in your OS language settings.