Resources and Responsibility

Computing and storage requirements should be considered alongside the task a system is intended to perform. Processing more data or repeating a calculation needs a reason.

For intelligent systems, that means asking whether the level of computation is appropriate to the problem. For storage, it means considering what needs to be retained, for what purpose and for how long.

Account systems, runtimes and storage may address requirements shared by several projects. Reusing an existing capability can avoid repeating development and maintenance work.

The assessment should include the cost of a common dependency. A shared system needs support, and its changes may affect several projects. Reuse makes sense where those obligations are proportionate to the benefit.

The resources committed to a system continue after it is introduced. Maintenance, upgrades and eventual replacement all deserve a place in the initial decision.

Our preference is for systems that can be understood and maintained without unnecessary replacement. The ability to transfer knowledge, make changes and retire a component in an orderly way matters to its lifetime cost.

Environmental assessments should state what is measured, over which period and against which baseline. A change in one part of a system may move demand elsewhere.

Claims of improvement should account for that wider context. Assumptions and limitations should be explicit enough for another reader to understand the conclusion. These are principles for evaluating resource decisions and their consequences.