MONK TeksTale demo
October 1, 2008

Present: Amit Kumar, Duane Searsmith, Martin Mueller, Stan Ruecker, Catherine Plaisant

Duane provided a demo of the updated version of TeksTale, and a list of his current tasks. He'll send around a URL soon to a short list, since some of the code is still proprietary.

TeksTale currently uses the MONK db but a different proxy.

The system gets faster as you work, since it is caching quite a lot.

It clusters at two levels.

Martin points out we will want to be able to filter out character names. Duane says this is straightforward to do.

Duane showed some visualizations of the clustering results:

• Wordles. Since they are proprietary, we'll need to use our own Wordcloud format.
• Dendograms
• Topic maps

How much do we want to integrate tools with the MONK workbench? Is it sufficient to have different tools that can be hooked to the datastore through the same middleware?

Martin: I'd rather have two interfaces that work.
Catherine: we do want results from users, so any strategy that gives us more of those results sooner.

We could then report to Mellon that we have integration at different levels.

Could we export collection selection from the MONK workbench for loading in another tool like TeksTale, or vice versa? This seems possible.