Implementation: A version of Gini/Airtime based on Timestamps rather than Content

[xehu]

Currently, our measures of airtime equality are based on the message content --- e.g., the number of words, the number of characters, and the number of messages. Another method of quantifying 'airtime' might be the actual time used for speaking; e.g., using the start/stop times of each speaker, compute how much time each person is actively speaking for. Then, compute the balance in speaking time(s).

[xehu]

This is especially something that we can tackle once #330 is fully done :)