← Back to Archive

The drawing that came back with six questions, and what each one meant

Reading a manufacturer's queries as a list of things the drawing failed to say.

A drawing that comes back from a workshop with a list of questions has not been misunderstood, it has been read correctly and found incomplete, since every genuine question is really pointing at a place where the drawing left a decision unmade that the workshop cannot safely make on its own behalf.

A question marks the spot where the instruction ran out

A workshop reading a drawing has to be able to make every decision about the part from what is written on the page alone, and a question sent back is what happens when that workshop reaches a point where the drawing simply does not say, whether that is a dimension left off, a surface finish left unstated, or a tolerance never given for a feature that clearly needs one. Each question points precisely at the spot where the drawing stopped being an instruction and started requiring a guess. A careful workshop asks at that spot, because a guess made on its own behalf is a decision the designer never got to make.

A workshop that asks nothing at all, faced with a drawing that has real gaps, may simply have filled them with its own assumptions and told nobody. That is a worse outcome than a returned question, since the resulting part can look entirely correct while being built to a guess. Nor can a workshop sensibly guess at one gap and treat the rest of the drawing as binding, because a gap treated as optional in one place invites the reading that other instructions were only approximate too, and a shop that lets one gap slide has in effect decided that every unstated assumption on the page is fair game.

A shopping list that just says cheese

Handing someone a shopping list that says only "cheese", and sending them to the shop, produces exactly this kind of question, since that one word has no single reasonable interpretation among the dozens of cheeses on the shelf. A shopper who cares about getting it right comes back to ask instead of picking one at random and hoping. The question is the shopper correctly finding the one place the list failed to instruct them, in the same way a workshop's question finds the one place a drawing failed to instruct it. A shopper who guesses might still bring back the right cheese, but the list never made that certain, and a drawing with the equivalent gap makes nothing certain either.

Six separate questions on one drawing usually mean six places where something felt too obvious to the writer to spell out, and reading them as a group often shows the same habit repeating across the whole page: every hole given a size and no depth, perhaps, or every face given a finish except the ones that were machined last.

Answering on the drawing instead of by email

Most returned questions fall into a few familiar kinds, and each kind translates directly into a missing piece of the drawing. "How deep is this hole?" means a depth was never stated. "Which face do we measure from?" means no datum was named. "What finish do you need here?" means a surface was left to the machinist's taste, and "Does this edge need breaking?" means nobody said what to do with the sharp edges left by cutting. Read this way, the list of questions is a list of repairs.

Answering a question properly means adding whatever was missing to the master drawing itself, since the same gap will produce the same question the next time the part is made unless the drawing is corrected. Answering by email is faster in the moment and leaves the gap where it was, so the next batch, or the next shop, reaches the same ambiguity and asks the same thing again. That is how a drawing acquires a reputation for being difficult without anyone ever writing down what was missing from it. Treating each question as a defect in the document costs a little more on the day and closes the gap permanently.

The habit worth keeping is to log what was asked. A set of questions gathered across many parts starts to show which decisions a designer habitually leaves off a drawing, and that pattern is more useful than any single answer, because it fixes the next drawing before anyone has had to ask.

More on Drawings