added/modified files for experimental integration.

[fractaldragonflies]

Getting up to speed on GitHub as well. Hopefully correctly.

Added my notebook that I used for experiments before running full test scripts. Changes really were experimental -- so no problem rejecting some of my experiments.

• Added my Markov code using NLTK3.5 as composition to markov.py. No problem using Tiago's ngram module, I just didn't have the background to use correctly.
• Added code for creating 'form chars' and 'sca' directly to data.py. Dropped 'loan' as I preferred to test the condition versus 'borrowed'.
• Very little change to plot.py.
• Few changes to tests/ test_data.py and test_markov.py.
• Changes to commands/plot_entropy.py. BUT I didn't know how to test it! I don't know how to invoke the CLI.
• Work around also for my lack of knowledge on getting path right, I added 'src' to path as first step in test modules. I realized there has to be better way, but already too much time fussing with it.
• Notebook was convenient way to play with individual modules while developing. Using Spyder, so maybe possible from Spyder as well.

[fractaldragonflies]

A note on markov.py. I used the post_order variable as indication of how many previous segments we are conditioning the current segment on. Not sure if this was appropriate.

[LinguList]

So to summarize: we seem to have a common understanding (my comments are not all important, some wrong, as I wasn't seeing the full picture), and we can keep on working. The important thing is the "formspec" I mention, as it spares a lot of work.