The term architecture as originally applied to computer systems referred to that part of the computer system which is visible to the programmer–i.e., the instruction set, interrupt characteristics, etc. As such, the computer architecture field has seen few advances in the past several years. Among computer architectures as different as the Burroughs B6700 and the IBM System/370 there is no clear advantage of one relative to another. Performance is determined much more by the circuit technology and the memory size and technology than by the peculiarities of the instruction set. The same situation appears to exist in minicomputers (e.g., the PDP-11 and HP-3000). Although no microprocessor using a stack architecture is known to this writer, it seems unlikely that such a CPU would have unique characteristics. Therefore in setting up the Lake Arrowhead Workshop session on Latency Cost and Architecture it seemed unprofitable to spend time investigating computer architecture in the classical sense. Computers will very likely "look" the same in the foreseeable future as they do now.