Sounds like a very impressive project!
Melissa Terras wrote:
>The reason for doing this was to provide an AI
>character "agent" with data to make models of
>characters, and compare unknown characters to these
>models. Marking up the texts in XML provided a bridge
>between image and textual data. The resulting system
>propogates possible interpretations of unknown
>characters: a tool to help papyrologists read damaged
>texts. We have had some early success with this so
>I confess that my XML markup is not TEI compliant -
>yet. I think that it wouldnt take too much work to
>bash it into shape, and to change my data files. I
>guess my question to the list is, is there any point
>in making such a markup TEI compliant? (I am aware of
>the preservation of data issues). Is anyone else doing
>such work, or does textual encoding tend to stop at
>the character level? Is there room for markup on a
>stroke level in the TEI, or would I be making
>unnecessary work for myself ;)
I think it is fair to say that TEI markup in P3/P4 (the previous and
current editions respectively) stopped at the character level, although
there are mechanisms for encoding glyph variations.
The current revision of the TEI Guidelines (P5) has not yet addressed
issues such as use of SVG graphics.
One interesting possibility would be to encode the glyphs using SVG
(Scalable Vector Graphics, for more information see:
http://www.w3.org/Graphics/SVG/Overview.htm8) and declare them as NDATA
entities in your DTD and embed the entities directly in the text.
The reason I suggest SVG is that it is an XML format that describes
graphic content as a series of vectors. I have been interested in this
as a possible approach for recognition and searching of cuneiform texts
but have not yet had the time to test it. Vector analysis is fairly well
understood and there are a number of free software packages for vector
analysis. You should be able to specify deviations that would still
count as a "match" for your searches.
I think your research has the potential to open up new avenues for
encoding/analysis of ancient texts in particular. Keep up the good work!
Director of Research and Development
Society of Biblical Literature
[log in to unmask]
Co-Editor, ISO Reference Model for Topic Maps