An important class of expensive and lengthy R&D projects seeks to reduce the subsequent costs of operating systems. Some energy R&D projects are of this kind. Despite assertions of technologists that R&D undertaken will produce a large technological advance, it can often be economically much wiser to aim deliberately for a far lesser technological advance--if this lesser objective is attainable earlier and at lower R&D cost. This result is intuitively plausible if one accepts discounting, or the fact that there is a time preference for money. A dollar available this year has more value to us than a dollar available next year. To make this notion quantitative, and the possible trade-offs more easily discernible, generally requires that many parameters and parameter interactions be accounted for in systematic exploration of discounting questions, even in the simplest reasonable models of cost-reducing R&D. Such a very simple model is developed in this paper. Enough detail is provided so that readers can trace through for themselves examples which may particularly interest them.