Architects See More Dimensions

Many either-or conflicts dissolve once you add a dimension the debaters are not seeing.

Adding a dimension such as automation moves the whole trade-off curve, so you can get more of both.

To an executive

Many apparent either-or choices, such as speed or quality, are false. Looking from another angle often reveals an option that delivers more of both.

To an engineer

When a debate stalls, ask which dimension both sides are missing. Look for automation or tooling that moves the whole trade-off curve instead of picking a point on it.

The idea

Hohpe observes that many stalled debates happen because each side is looking at a different projection of the same thing. His example: one person insists an object is a rectangle, another that it is a circle. Both are right, because it is a cylinder seen from two directions. Architects add value by revealing the extra dimension in which both views make sense.

The same pattern appears in everyday choices that look like a single axis: fast versus good, centralised versus decentralised, autonomy versus alignment. Hohpe argues, for instance, that alignment is a precondition for autonomy rather than its opposite. A familiar engineering case is speed versus quality: framed on one axis it looks like a trade-off, but automation and better tooling can move the whole curve and deliver more of both.

Why it works

Arguments on one axis can only end in compromise. If the only choice is a point on a line, someone loses. Adding a dimension opens options that neither side had considered.

Models shape what people can see. Every model simplifies, and the model a team uses decides which trade-offs look fixed. Changing the model changes the conversation.

A change of perspective beats more effort. Hohpe quotes Alan Kay's line that a change in perspective is worth 80 IQ points. Finding the missing dimension often resolves what more analysis on the same axis could not.

Where it applies

Deadlocked design reviews, where two sound positions keep colliding.
Organisational debates, such as central control versus team autonomy.
Governance versus delivery speed, where automation is often the missing dimension.
Stakeholder conflicts in which each side has a valid but partial view.

How to apply

Restate both positions fairly, so each side agrees you have understood them.
Ask what each side is optimising for, and what they are assuming stays fixed.
Look for a hidden dimension: time (now versus later), scope (per use case), layer, automation, or ownership.
Draw the problem on two axes instead of one, and show where the new option sits.
Check it is real: confirm the new dimension actually delivers both outcomes, at a cost both sides accept.

Example scenario

Security wanted every deployment manually reviewed; product wanted daily releases. Each meeting ended in a compromise that satisfied neither: weekly releases with a rushed review.

The architect added automation as a dimension. Policy checks were written as code and run in the pipeline on every change, with manual review kept for a short list of high-risk changes. Security got stronger and more consistent control, and product got daily releases.

Questions to ask

What dimension are both sides of this argument leaving out?
Is this a true trade-off, or a fixed point on a curve we could move?
Could the answer differ by use case, by layer or over time?

When not to use it

Genuine trade-offs at a known limit, such as consistency against availability during a network partition, where the dimension you need does not exist.
Decisions that simply need making, where searching for a clever reframe only delays the call.

How it misleads you

Pretending every trade-off is a false dichotomy. Some trade-offs are real and simply have to be decided. A new dimension also has its own cost: automation, layering or per-case rules add complexity, so make sure the gain is worth it.