Prototyping in a workshop you share

Designing around machine availability rather than around capability.

Prototyping in a shared workshop means designing a part for the machine that is actually free this afternoon rather than for the machine best suited to it in principle, because a theoretically superior process with a two-week queue in front of it answers the design question slower than a merely adequate process a person can walk up to and use right now.

What is really going on

A workshop full of capable machines is not the same thing as a workshop full of available machines, and the gap between those two ideas is exactly what a shared lab makes impossible to ignore. A single high-precision mill might be the objectively correct tool for a bracket, cutting it faster and to a tighter tolerance than any alternative, but if six other projects are already booked onto it for the coming fortnight, the mill's raw capability is worth nothing to whoever needs an answer this week. The practical question stops being which machine could make this part best and becomes which machine can make an adequate version of this part soonest, a substitution that feels like a compromise the first few times it is made and turns out, more often than expected, to cost very little in the finished part.

A private workshop with one owner never forces this reckoning, since the owner's only real constraint is their own time, and a machine sitting idle is simply available whenever it is wanted. A shared workshop turns every machine into something closer to a public resource with its own waiting line, and a design drawn without any thought for that queue is, in effect, a design that has assumed a private workshop it does not actually have.

The list of alternatives worth keeping

A useful habit, once the pattern becomes visible, is drawing up a short mental list of two or three machines that could plausibly make a given feature adequately, rather than settling on the single best one and stopping there, since that list is what actually determines how quickly a part starts moving once it needs to be made. A feature that can only be produced one way has, without anyone deciding it deliberately, tied that part's schedule to a single queue, while a feature with a genuine second and third option keeps the part moving regardless of which particular machine happens to be busy on the day it is needed.

The shared-kitchen comparison

A shared kitchen in a busy hostel usually has one good gas hob and a slower, less capable microwave and toaster sitting beside it, and a cook planning a meal around whichever appliance happens to be free when hunger actually strikes eats sooner, and often just as well, as one who insists on queuing for the hob because it is, in principle, the better tool for the job. The meal itself often changes shape around the appliance rather than staying fixed while the cook simply waits, toast and a microwaved egg standing in for a proper fried breakfast, and the substitution works because the goal was food on the table rather than proof that the best possible method had been used to make it. A prototype bracket cut on the machine that happens to be free follows the identical logic, the design bending slightly to suit an available tool rather than the schedule bending to wait for the theoretically ideal one.

Why the design has to bend rather than the schedule

Treating machine choice as fixed and the schedule as the only variable worth negotiating gets the trade backwards for most prototype work, since a design that has been drawn around one specific machine's strengths, a feature that only that machine can cut efficiently, a tolerance that only that machine can reliably hold, has quietly made itself dependent on that machine's queue. Drawing the same part with a little more slack, a feature simplified just enough that it can also be made on whichever machine happens to be open, keeps the actual bottleneck from being a single piece of equipment that thirty people are all trying to book at once. This does mean giving something up, a slightly less elegant feature, a slightly looser tolerance, in exchange for a part that can start being cut the moment a machine is free rather than the moment the ideal machine specifically becomes free.

The number that matters here

Parts designed with this kind of flexibility, able to be made acceptably well on two or three different machines rather than depending on exactly one, spent measurably less time sitting in a queue across that year than parts drawn around a single specific machine's capability, often getting cut within a day rather than waiting the better part of a week for one particular piece of equipment to come free. The saved time rarely showed up as an improvement to any single part, it showed up as an entire project finishing on schedule instead of stalling behind a machine six other people were also waiting on.

Why this matters in practice

Designing for availability rather than pure capability is, in effect, designing the part to have options, and a part with options simply moves through a shared workshop faster than one that has committed itself to a single machine before that machine's queue was ever checked. The practical habit is checking what is actually free, or likely to be free soon, before finalising a feature that only one machine in the building can produce, and treating a design that only works one way as a design that has taken on a dependency worth questioning rather than a design that has simply used the best available process.

What this does not explain

None of this argues for lowering the bar on parts where the capability genuinely matters, a bearing seat that needs a specific tolerance, a surface that needs a specific finish, since some features are worth waiting for the right machine regardless of the queue in front of it. The judgement being made is not availability over quality as a blanket rule, it is recognising which features on a given part actually require the best machine and which ones were only assigned to it out of habit, since the second category is usually larger than it first appears once someone actually checks.