TEI Markup for paintings

For my presentation in Zagreb, I was experimenting with the surface element in the TEI markup. The only thing that stumped me was how to tell the TEI document that multiple JPEGs represent the same information, possibly at different points of resolution.

Just like that this image

or this image

are both  part of this image

and some things are easier to see in the one and in the other.

Drawing cool Graphs Quickly

In the process of working out how to sketch the knowledge for my DH talk in Zagreb, I found a Graphviz website that allows one to experiment with the DOT material immediately.

WebGraphviz has a type-in box for the graph description and a display area below where a simple click of [Generate Graph] will output the results. I usually keep the DOT Language Guide close at hand in the adjacent tab. GraphViz supports such a rich set of output formats, but unfortunately, Google Slides (nor Blogger) will not handle SVGs correctly ... :P

Here a rather sizeable graph from my presentation for Zagreb.

Lishi Website

Completely forgotten that I used to work on this (here my polyptique interests took their departure).
Probably would need to request a password again at this point in time.

IDP Reasoning System

IDP is a reasoning system of the DTAI at the University of Leuven. There is a web-interface that can be used for experimentation. The system differs from the ASP format in specific ways, though it competes in the tri-annual ASP competitions (for a download of problems see here) It supports representations such as Abstract Dialectical Frameworks.

digital humanity tools

I looked at WordHoard today, a tool spearheaded by Martin Mueller of NWU, one of the grandmasters in the TEI text-encoding initiative (he was also behind the MONK project).

I was struck by some of the cool aspects of WordHoard, such as the stemming and POS infrastructure and the Bean scripting window, as well as some of the weirdness (MySQL-backed hibernate objects that require full rebuild for every XML edit?).

Some of the linguistic services available from MONK are incorporated in the MorphAdorner that is still available on BitBucket (with cool online examples hosted at NWU). There is also a collaboration with the Lucene-based BlackLab index.

The TEI community seems to be big fans of the Oxygen XML editor, but I can handle a lot of pain for 240US$, to be quite honest about it.

I also noticed that the next TEI conference is in Vienna, in September, and that they have a call for papers out now. Maybe I can think of something cool to write on.

Dissertation Organization

I remain convinced that in some sense a dissertation, just like a master thesis, is an exercise in knowledge organization before it is anything else. The degree in the Social Sciences and Humanities is given to those people who can demonstrate that the can take a large body of knowledge, enter it, trace its connections to their respective lairs, and emerge with a 300-400 page record of that journey that documents their successful traversal.

Thus, developing the infrastructure required to get to the point of completing that traversal successfully is a key element of the overall effort. Just as in the case of the master thesis, having developed such infrastructure is an enabling tool for the next stage of the journey---whether that is the habilitation or the professional career of the researcher, is a separate matter.

With the first 100 page draft of my overall project, I was able to convince myself that the abilities tested for in the master thesis are still available; I dusted these off, so to speak. The approach of re-deriving things from the primary sources does not only speak to the archival-centric focus of the effort, but also forms a bulwark against the mountain of secondary literature, by definition average on average, and an inoculation against easy interpretations, especially in the partisan literature.

The overarching effort requires some form of dashboard, that makes it easy to

• track the incoming literature
• track the "processing" phase of the identified literature, at minimum
• obtaining (downloading, ordering, purchase, ...)
• bibliographical indexing (BibTeX)
• processing of contents
• determination of novelty and relevance
• excerpting
• secondary long-term storage in the case of downloads
• query of state of each item
• answering "what next" and "how much effort" types of questions

On the surface of it, one would suspect that bug and task tracking systems might be best suited to this effort; on the other hand, their notions of modules, features and OS types square badly with the requirements of the object being manipulated.

BibDesk type database applications on top of BibTeX files are helpful at the level of granularity of the individual book or article, esp. if one is willing to use the chapter approach to sub-sections in works such as Sacred Loneliness or Early Mormon Documents, to map the granularity gap between bibtex and the units of information at hand.

As with all structurally organized information, they might provide a way to transform out of a specific solution, once a better approach is identified.