I have been working on the XML encoding of the Vindolanda Tablets trying to update the website "Vindolanda Tablets Online" with the new tablets from the third book on The Vindolanda Writting Tablets, whilst adding extended functionality to the website. This includes work with contextual encoding and creating XML through PHP scripting.

This work has enabled me to create a new website for the Vindolanda Tablets, which is not finished yet. It uses the contextual encoding of the Vindolanda Tablets (fig. 1) to pull out the indexed words and allows the user to search through them using Ajax LiveSearch technology (fig. 2).

Over the last year I have been working on the XML encoding of the Vindolanda Ink tablets. The second publication of the Vindolanda Tablets (Bowman and Thomas 1994) which also incorporated some of the tablets from the first publication are available online as a resource, which allows users to search and view the tablets.

The online resource is attached to a set of XML documents (one for each tablets). These are available to download as a batch but they have no real use in the build of the website. 

At first I has assigned the task of figuring out how best to add the third Vindolanda Tablets publication (Bowman and Thomas 2003) while updating the existing XML to the newest EpiDoc Schema. The EpiDoc Schema is built on the TEI (Test Encoding Initiative) Schema, which specifies how to encode texts with XML. The EpiDoc Schema produces a framework to encode ancient documents with XML.

In image processing terms, what I'm trying to do is perform feature extraction from the images. A number of methods are available, but our images are such, that we need to get away from traditional methods such as simple thresholding, or plain classical edges and corner detection (Canny edge detection, Hough transform, etc...) The reason for this is that our images are extremely noisy (i.e., a lot of information in the image is irrelevant information), and the classical methods don't work well on noisy images. There is one method though that has the potential to help us extract these features we're looking to extract. This method is the so-called phase congruency method [1]. And it's the one I've been working on. [ ... ]

We know from various perceptive psychology studies that experts mobilise a very large and varied amount of knowledge when transcribing a text (see Melissa's book). Most literature on the subject is plausible but, according to Mike, probably wrong. The corpus of knowledge is definitely vast -- and certainly too rich for us to be able to represent it all. So where can we start? [ ... ]

This tutorial is on making an ontology for wine and it is very instructive.