Designing to a date rather than to a specification

How a fixed deadline changes which designs are even admissible.

Designing to a date means treating the calendar as a hard constraint on equal footing with any performance target, so that a design only counts as admissible if it can be finished, checked and trusted by the fixed day, whatever its theoretical performance might have been.

Time as a constraint alongside weight and cost

Design work is usually framed as a search for the solution that best satisfies a specification, with the time spent treated as elastic: refinement continues until the result is good enough, and only then is the project called finished. A fixed date removes that elasticity. Time becomes another constraint sitting alongside weight, stiffness or cost, and every candidate has to pass an extra test unrelated to how well it performs on paper: can it be designed, made, verified and, if something goes wrong, corrected within what remains on the calendar. That test rules out whole categories of otherwise excellent solution, simply because there is no realistic path from where the project stands to a finished, trusted version of them before the date. A part that takes three weeks to prove reliable may be every bit as good as one that takes three days, yet with three days left only the second is available. A specification asks whether a design meets a target; a deadline asks whether that design, starting from today, can still become a checked, trustworthy object in time, and a design can pass the first question comfortably while failing the second completely.

Cooking from the cupboard

Someone asked to cook a proper meal for a guest arriving in twenty minutes does not choose the best recipe in the book and go out for its ideal ingredients. They open the cupboard and cook from what is there, because the better dish that needs shopping for is off the menu, however good it looks on the page. Given until the weekend, the same cook might well buy exactly what the ideal recipe calls for. Both choices meet the standard of the question being asked, and the question changes once the available time changes. A team a few days from a fixed date is answering the twenty-minute version every time it selects a part, and choosing a familiar bracket over an elegant new one is the same kind of engineering, done against an equally real constraint.

The writer and former physicist Eliyahu Goldratt described something close to this in his 1984 book The Goal, where a single fixed resource, the one that cannot be sped up or worked around, ends up shaping decisions across an entire operation. On a project working to a date, the calendar is that resource. Known processes and parts that can be built and proven quickly get chosen ahead of theoretically superior but unfamiliar ones, and a working baseline stays on the car until its replacement has earned its place by demonstration. "We ran out of time" then stops being an excuse and becomes a description of a constraint that was present, and shaping decisions, from the start.

Where the calendar gives no licence

A deadline can reasonably push a team toward a heavier bracket, a less elegant assembly, or a part borrowed from an earlier design, because those are trades against performance and polish, and performance and polish can be given up without anyone getting hurt. It cannot push a team toward skipping verification of anything where failure is dangerous, a brake mounting or a steering component being the obvious examples. There the missing time would be traded against safety, a shortage no calendar can excuse, so those checks keep their full length however close the date.

My key error with this

My instinct on that chassis was to work the way I had been taught to draw, which is to take one region and finish it properly before moving to the next, so that the parts I had done were correct in full detail and the parts I had not done did not exist yet. An engine change late in the first year caught that approach exactly where it is weakest, because the rear was the region I had not reached, and what I held at that moment was a beautifully resolved front and nothing at all that connected the wheels to anything. What replaced the belief is that a design should be valid at every stage of its own completion rather than only at the end, so I began working the whole structure at once and refining it in passes, on the rule that if somebody took the drawings away on any given morning I should still be holding something that joins the wheels and carries the engine. Working that way is slower to feel finished and very much faster to recover.