This paper is a case study of the adoption of computer-assisted data handling and statistical analysis as an aid to the publication of a large quantity of Roman pottery. It describes the process of adjustment and enlightment undergone by the author through the enforced adoption of new techniques. It is hoped that is will give encouragment to others who lack training in computing and statistics to make the same transition by showing that these thchniques, once mastered, open up a subject to improved understanding and interpretation. The pottery from this military site of the 1st century A.C. is shown to bear interesting and otherwise unascertainable relationships to the excavated structures. Thus, as well as facilitating actual publication, this approach has improved the general archaeological interpretation of the site.