One cannot fault the objective of providing guidance on an appropriate price for a microcomputer needed to perform a certain task. The authors’ model predicts price based upon system characteristics (memory and disk capacity) and the time needed to run five problems on 22 microcomputers, including three “real world” problems. An equation based upon a regression model is presented. The work must have been done some time ago, as most computers had only 64k memories. The model provides additional evidence that Grosch’s Law is no longer a good predictor of price.
However, changes in the microcomputer marketplace (particularly in the very recent past) cast doubt on the value of these or similar efforts. First, the cost of add-ons (such as printers, disk storage, memory, and, most importantly, software) have made the cost of the base machine much less important. Second, development of two tiers of prices (original vendor (e.g., IBM) and clones (e.g., Leading Edge and generics)) clearly show how arbitrary retail prices are in the first place. Finally, the plummeting prices of microcomputers are reducing the differences between machines.
Microcomputers should be bought by selecting software first, and then considering hardware which can run it effectively. The industry could use a good discipline for conducting meaningful evaluations. This approach is not it.