Some of what you want to do could be covered (or has, in certain 
projects) been covered by the use of @rend on various elements 
describing what is going on. So for example a special feature of an 
abbreviation could be marked as <abbr rend="insular">, or a single 
character within <c> or <hi>. There are, of course, no special 
attributes for saying "this character is in a different colour ink", or 
"this word is in majuscule" to represent a special meaning, which might 
be useful.

The only projects I know of that have tried to express this kind of 
information are (1) the early DLV which had an XML transcript in TEI, 
where <handShift>s were used to indicate change in ink colour, <abbr>s 
were typed into sixteen or so categories, and so forth. No TEI 
extensions, though. (2) EpiDoc, where detailed comment on physical 
appearance of letters is significant because most epigraphic texts are 
so short and obscure: we mostly use <hi> to mark tall letters, or 
strikethrough, or overbar, etc.

This will be particularly important for a project that I am about to 
get involved with: Byzantine Seals in TEI, because sigillographers are 
interested in every single feature of letter forms, including four 
variant alphas (that have to be encoded by the same Unicode character). 
For this we shall have to come up with some way of marking the 
letterform and other special features in XML and then displaying them 
as variant unicode font glyphs in the diplomatic transcript.

I too would be very interested to know if other projects have codified 
this kind of material already (because I hate reinventing wheels...)



Quoting Juan Garces <[log in to unmask]>:

> Does anybody have experience with or know about encoding palaegraphic
> observations about manuscripts on the word or even the character level? I am
> interested in marking up a series of such observations, such as different
> versions of the same letter, for which there is only one unicode codepoint.
> I am aware that some kinds of observations, like "graphically distinct from
> the surrounding text", are catered for by the existing tagset, but is there
> a general interest in expanding/adapting/creating tags for such a kind of
> markup?
> Thanks for all responses in advance.
> Best wishes,
> Juan

Dr Gabriel BODARD
Inscriptions of Aphrodisias
Centre for Computing in the Humanities
King's College London
Kay House
7, Arundel Street
London WC2R 3DX

Email: [log in to unmask]
Tel: +44 (0)20 78 48 13 88
Fax: +44 (0)20 78 48 29 80