Minimal Workspace Principles
Not aesthetics. Decisions.
See Affiliate Disclosure for our policy on links in Clearings.
This is not a guide to making your desk look good in photographs. Minimalism as an aesthetic is a different project entirely, and not one this collection is interested in. These are principles about cognitive load, decision overhead, and the relationship between your environment and the quality of your thinking. The desk can look however it looks.
A desk used for focused work should not also be the place where administrative tasks accumulate, physical objects pile up, or ambient reading happens. Each type of work activates a different cognitive mode; the physical environment that surrounds each type shapes which mode you enter.
Why this matters more than tidiness: A tidy but multipurpose surface still carries cognitive residue from its various uses. Purpose-specificity does more work than neatness.
Applications, browser tabs, and documents not relevant to the current task should be closed rather than minimised. Minimised windows remain in working memory. Closed windows do not. The difference in cognitive overhead is small per instance and significant in aggregate.
The objection: Some people find reopening applications disruptive. The relevant question is whether the disruption of reopening is larger than the overhead of keeping things open. In most cases, it isn't.
A reading list with more than approximately thirty items is not a reading list — it is an aspirational archive. The distinction matters because a reading list implies future action, and each item on it carries a small ongoing cost as an open loop. A settled list, with honest culling, reduces this overhead without requiring you to actually read everything.
The source: The principle draws on David Allen's work on open loops, applied specifically to saved content rather than tasks. The mechanism is similar; the remedy is the same: decide, then close.
The question of which notifications to allow is less useful than the question of when to allow any notifications at all. Defining periods during which no notifications are received — and protecting those periods — is a more robust approach than attempting to filter by source or priority.
The related reading: The research basis for this is reasonably well-established in the cognitive science literature on task-switching costs. Gloria Mark's work on recovery time after interruption is the most cited source and worth reading directly rather than through summaries.
Deciding what the working day contains before the working day begins reduces decision overhead during the day itself. The plan does not need to be followed precisely; its value is in establishing a default ordering so that the question of what to do next is pre-answered most of the time.
The distinction: This is different from time-blocking or rigid scheduling. It is a loose ordering of intentions, not a calendar. The goal is reduced decision fatigue, not perfect adherence.
Tools accumulate. Applications that were useful at one point in a workflow remain installed and occasionally visible long after the workflow they served has changed. An annual review that asks, for each tool, whether it was used in the past three months and whether the work it supports could be done without it, tends to reduce both subscription costs and the low-level cognitive presence of unused software.
The framing: The question is not whether a tool is good. It is whether it is currently earning its place in your environment. These are different questions.
Capture and processing are cognitively distinct activities. Attempting to capture, evaluate, organise, and act on information simultaneously produces neither good capture nor good processing. A single, low-friction capture location — one place where anything can go — enables genuine processing to happen at scheduled intervals rather than continuously and inadequately.
The risk: The capture location becomes a permanent holding pen rather than a transit point. The principle requires scheduled processing — capture without processing is just a different kind of accumulation.