Code for doing research; includes monitoring of emacs user behaviour for empirical research, and organization of research papers.
This package keeps track of which commands the user has used interactively.
written by John C G Sturdy, October 2004
IPR declaration: all my original work
Acknowledgement: Part of the B4STEP project of the SFI (Science Foundation Ireland) at University of Limerick, written while on a SOCRATES exchange to Umeå University.
It is meant for empirical studies in software engineering -- i.e. looking at how programmers do their thing. Although other parts of the programmers' environment have been studied by HCI specialists, editors seem to have received relatively little attention (something which I am trying to rectify) despite their being the kind of program which probably receives most of the users' input.
You could put it in new users' startup files (with their consent), with command-log-save-to-file in their exit sequence, to find how fast people increase the number of commands they use as they learn emacs, and to see how wide the spread of command set sizes is among a class.
You could also use it to spot which commands deserve to have easy key-bindings. I might add some code to do analysis of this, although it's probably too much trouble to record the route by which each command was reached, and so the analysis would just use all the keymaps which were around at the time.
With suitable log analysis software, you could spot patterns of command use that suggest people learning new commands from each other, for example, as well as looking at the learning curves and the command set sizes.
Last modified: Mon Dec 05 16:03:05 GMT Standard Time 2005