Affiliations: [a] Alphametrics Corporation, PO Box 2566, Bala Cynwyd,
PA 19004-6566, USA, Tel.: +1 610 664-0386; Fax: +1 610 664-0271; Email:
[email protected]
Correspondence:
[*]
I am grateful to Ronald Bodkin, Paul Coomes, Ray Fair, Ric Herbert, Guy
Judge, David Kendrick, Michael Lovell, James MacKinnon, Michael McCarthy,
Timothy Taylor, Volker Wieland, and several anonymous readers for having
taken the time to read and comment on this paper. I alone am responsible for
the opinions stated and any errors or omissions that remain.
Abstract: This paper, originally presented at the 1997 annual meeting of the Society for
Computation in Economics, offers a perspective on the creation of econometric
software for general use. To date there have been a number of attempts on the
part of economist-developers to describe their work, but often these have been
restricted to a presentation of specific algorithms, rather than taking the form
of presenting an overall view of the process. While it stops short of proposing
a formal evaluative methodology the paper attempts to establish criteria for the
evaluation of econometric software that include the man-machine interface,
database management capabilities, specific econometric functionality, and the
general environment in which the software is used. The point of reference for
the paper is the author's experience as it relates to the development of the
MODLER BLUE econometric software package and related systems.
