Choosing a process before choosing a shape
How the available machine constrains the design that can be drawn.
Choosing a process before choosing a shape matters because every manufacturing process can only produce a limited family of shapes, defined by how the tool moves, what it can reach, and what the material is allowed to do while being formed, so a shape drawn without that process in mind can turn out to be something no available machine can actually make, regardless of how sensible it looks on the screen it was drawn on.
What is really going on
A designer working purely in a drawing program is free to sketch almost any shape imaginable, an enclosed cavity with no opening to it, an overhang with nothing supporting it from below, a slot narrower than any tool that could reach in to cut it. None of those constraints exist on the screen, because a drawing is just a description of a finished geometry with no reference at all to how that geometry would actually come into being. A real process, by contrast, always has to get from raw material to finished shape through some physical sequence of operations, a tool approaching from a direction, a mould opening along a particular line, a beam travelling along a path it can actually reach, and any shape that sequence cannot physically execute is not a difficult shape to make, it is not a makeable shape by that process at all.
The stencil-letter comparison
Anyone who has ever painted letters through a cut cardboard or plastic stencil has met this constraint directly, usually without realising it had a name. The letter O looks simple enough to cut as a single ring-shaped hole, but cutting it that way leaves the disc of material in the middle of the O completely unsupported, free to fall out of the stencil the instant it is lifted, taking the letter's clean shape with it. Every real O stencil solves this with a pair of small bridges, thin strips of material left uncut across the ring, holding the island in place and leaving a small, deliberate break in the finished painted letter as the price of the stencil actually working as one piece. The letter itself did not change, the process demanded a compromise the drawing alone never would have suggested, and the same demand, in a more technical form, shows up in nearly every manufacturing process that has to hold a shape together while it is being made.
Why the constraint has to come first
Discovering a process constraint after a shape has already been finalised is far more expensive than discovering it before, since reworking a design to suit a process usually means changing dimensions, adding features like the stencil's bridges, or splitting a single part into two that get joined afterward, all of which ripple into whatever else the design already depends on. Starting instead from the process, understanding what a mould can release, what a cutting head can reach, or what a mechanism can physically be assembled in, means the shape is drawn from the outset inside the boundary of what can actually be made, and the design review that would otherwise catch the problem late becomes unnecessary because the problem was never drawn into the design in the first place.
One figure worth keeping in mind
A part redesigned late to accommodate a manufacturing constraint that was missed at the sketch stage commonly costs several times more engineering time to fix than it would have taken to design correctly from the start, not because the fix itself is difficult but because every downstream decision already made around the original shape, a mating part's dimension, a fastener's location, an assembly sequence, now has to be revisited alongside it. The same escalation covered earlier in this collection, where a manufacturing problem caught during a design review costs a short conversation and the same problem caught after tooling exists costs a redesign, applies here just as directly, since a process constraint is simply one more kind of manufacturing knowledge that is cheap early and expensive late.
Where the same idea shows up in mould-based processes
Injection moulding and die casting demand exactly the same kind of accommodation the stencil does, since a mould has to open along a single parting line and release the finished part cleanly in that direction, which rules out any undercut feature that would physically snag on the tool as it withdraws. Walls in a moulded part are conventionally given a slight draft angle, a small deliberate lean away from vertical, purely so the solidified part can slide cleanly off the tool's core rather than gripping it through friction along a perfectly parallel wall. A designer who specifies dead-vertical walls purely because that is what the finished part conceptually calls for is quietly asking the mould to do something moulds cannot reliably do, and the fix, adding draft the part never really needed for its own function, exists purely to satisfy the process rather than the design intent, the same small compromise the stencil's bridges represent in a cruder, more visible form.
Why this matters in practice
Recognising process as a constraint on shape rather than a detail to be settled afterward changes the very first step of designing a part, from sketching the ideal geometry and hoping something can make it, to asking what a chosen process can actually produce and sketching inside that boundary from the first line drawn. This does not make the design worse, since almost every real process still leaves an enormous range of shapes available within its constraints, it simply means the constraints get discovered while a shape is still cheap to change rather than after it has already been committed to drawings, tooling, and everything built on top of them, and it means the choice of process, cutting, moulding or otherwise, has to be treated as one of the very first design decisions rather than a detail settled once the drawing is otherwise finished.