Architecture sells options

Greg Hohpe describes¹ software architecture as being a way to defer decisions. Using a financial metaphor, architecture sells options.

Couple of examples he gives: You can defer the decision of server size if you architecture is horizontally scalable because you can add additional servers later. You can defer the decision of an authentication mechanism if the architecture separates concerns well. This second example also ties in to Abstractions raise tempo.

Point to note: even if the future expense is more than the current one, the option is still valuable as it allows for deferring the decision in exchange for more information. This is useful because we mostly do not have enough information at any time. The world is complex (Complexity), and has a surprising amount of detail (Reality has a suprising amount of detail). In general, the more the uncertainty, the more valuable options are. In times of more uncertainty, the business should therefore invest more into architecture.

Architecture, then, despite the high-modernist, dictator-ish images that it evokes, can actually be agile (Agile is the steering wheel), and Increase your tempo.