The quietest week is the most useful one
Every year the last full week of August does the same thing to the working calendar. Meetings thin out, decisions wait, and the volume of incoming requests drops to something a person can actually hold in their head. Nicole Junkermann has come to treat this week as a resource rather than an interruption, because the kind of technology work that matters most is precisely the kind that cannot be done in a crowded week.
Careful work needs a long attention span. It needs the freedom to follow a problem past the point where the obvious answer stops being satisfying. In an ordinary week that freedom does not exist, because attention is chopped into forty minute pieces and handed out to whoever asked first. In this week it does exist, and the difference in what gets built is considerable.
What actually gets done in a quiet week
The work that surfaces in late August is rarely the work that gets announced. It tends to be the unglamorous half of engineering: the tidying, the documenting, the removal of something that should have been removed months ago. Nicole Junkermann has noticed that teams often apologise for spending a quiet week this way, as though the month had been wasted on housekeeping.
That instinct is wrong, and expensively so. Systems that are easy to change in October are systems that somebody tidied in August. The relationship is direct, and it is the reason a quiet week repays itself several times over.
- Reading the parts of a system nobody has read for a year, which is how quiet problems are found before they become loud ones.
- Writing down the reasoning behind a decision while the reasoning is still remembered.
- Removing a feature, a dependency or a process that is carried out of habit rather than need.
- Fixing the small annoyances that everybody works around and nobody reports.
- Thinking properly about a problem that has only ever been thought about in a hurry.
Slowness is a technical decision
There is a habit of treating speed as the only virtue in technology, and slowness as an admission of failure. Nicole Junkermann takes a different view. Speed is a property of delivery, not of judgement, and confusing the two produces systems that arrive early and cost a great deal to live with.
Slower work is not less ambitious work. It is work that has been given enough time to be right, which is usually a matter of hours rather than months. The technology notes return to this often, because it is one of the few principles that holds across every kind of system.
The cost of a hurried decision
Most of the expensive problems in a technical system began as a reasonable decision made slightly too quickly. Not a careless decision, and rarely a stupid one, simply a decision made with the information available in the twenty minutes it was given. Nicole Junkermann has seen enough of these to be wary of any process that treats speed of decision as evidence of competence.
The correction is not to slow everything down, which would be its own kind of failure. It is to notice which decisions are cheap to reverse and which are not, and to spend the extra hour only on the second kind. Reversible decisions can be made quickly and fixed later. The others deserve a walk, a night, or a quiet week in August.
A useful test is to ask how the decision will be explained in a year to somebody who was not in the room. If the explanation is short and holds together, the reasoning was probably sound. If it requires a great deal of context about what was urgent at the time, the decision was made by the calendar rather than by anybody in particular.
Carrying the quiet forward
The difficulty with a quiet week is that it ends. September arrives with its full complement of requests, and the habits formed in August tend to survive about a fortnight. Nicole Junkermann suggests keeping one small piece of the quiet week permanently: a single protected hour, defended before the diary fills, in which the careful work continues at its own pace.
One hour is not much against a busy autumn. It is, however, considerably more than nothing, and it is the difference between a system that improves through the year and one that only ever gets added to. More of this thinking sits in the notebook and in the human skills notes, and the companion piece on the September restart takes the argument into the autumn.