F

thanks for your usual courteous response. Ross McKitrick has pointed out to me that the Journal of Applied Econometrics (their archive is at http://qed.econ.queensu.ca/jae/ and their policies are stated at http://www3.interscience.wiley.com/cgi-bin/jabout/4079/ProductInformation.html) strongly encourages authors to submit their command files, especially if replication requires use of a particular program. The example at their data archive includes a copy of such a computer program. This is a case where the policies extend further than merely providing the command file to an interested reviewer, but extend to the archiving of the command file with the data. This seems to me to be an effective methodology and a policy which you might well consider, if you are reflecting on archiving practices.

Verbal descriptions of methodology are often insufficiently precise and the archiving of computer scripts at the time of publication seems like a pretty good way of finalizing documentation. One would think that authors would do so for their own purposes, aside from the exigencies of publication.