Sketching Is Not What You Think It Is

In one sentence: what is the core claim or question about sketching that this conversation is ultimately trying to answer or argue?

Sketching in industrial design is a way of exploring different possibilities in a rough format. For me, sketching is a plural activity — multiple outputs, a progression, an exploration, an evolution of an idea through rapid, quick, ambiguous outputs. Typically done with pen on paper, digitally, or on a napkin.

Who is the audience for this discussion, and what are the stakes? Why does it matter to get the definition of sketching right?

The audience is people involved in industrial design and product design exploration at various stages — students at the beginning, practitioners at the early stages of their experience, educators weighing different approaches, and established designers who have progressed through a variety of approaches.

And what are the stakes? What goes wrong — for a student, a practitioner, an educator — if sketching is misunderstood or the wrong tool is reached for?

The stakes for society aren't huge. The stakes are at a personal development level. There can be self-deception, or there can be self-revelation. There can be discoveries about certain approaches. There can be paths explored, mental pathways built — or neglected. And there can be a sense of worth or ability by engaging all the different modes. That's what's at stake here. Nothing huge, just practical.

What do you already believe is true about sketching — something you'd consider a prior commitment, a position you'd defend even under challenge?

Sketching is twofold. It is the hand on paper creating a physical output — that's the physical act. But sketching is also a mental act that allows exploration of ideas. The ability to come up with new things, throw them away quickly, turn the page, come up with a new one — that mental sketching is integral.

Tell me everything you know, believe, suspect, or have experienced about sketching — its nature, its role in industrial design, the tools, the tensions, whatever comes. Any order. I won't interrupt. When you've said everything you want to say, just say done.

Sketching is usually done in response to a brief — a set of problems and constraints posed, looking for a possible solution to something new. Sketching at the beginning is a kind of limbering up, like stretching before jumping into hard work. There are initial throwaway sketches — a clearing of cobwebs, getting ready to explore ideas. Then you might take some wild jumps, go back and explore the spaces between, or work out a different idea. The act of sketching is both internal and external. The hand is making a mark that is slightly disassociated — not looking to be too specific, ambiguous in general. Several marks can be laid down on a page, and mentally a mark can be chosen that just feels like it fits the solution, then further refined. The secondary output is for someone else to observe, to see possibilities in it, to see it not as a final solution but open to interpretation. Those interpretations from various people contribute to the next evolution of the product — whether further sketches or more specific refinements. In the beginning, many paths and vectors are open in the solution space. The idea is to explore without preciousness. You're not doing an oil painting — you're doing pen on paper. That throwaway character of a sketch is key to its utility.

There is a utility aspect to sketching and a presentation aspect. A napkin drawing can be as rough as possible — a stick figure — and still show a relationship. At the high end, a classic industrial design sketcher uses markers and pens to produce something stylised and beautiful. That's a little intimidating for a lot of users. Some proponents of alternatives to sketching are intimidated by the killer presentation skills on ArtStation or Instagram, or by classical marker render sketchers. But that's not the goal. It's nice for presentation, beautiful for presentation — but sketching shouldn't be thought of as ego expression or competitive styling. It should be viewed as utilitarian: a way to get to an idea. That can happen at all levels of technical prowess with pen and paper.

The designer can take a look back at a sketch — turn the page, flip back, and see that sketch anew, see new possibilities. There is a deliberate imprecision in the way marks are laid on the paper; they're not trying to define a single solution. Lines are typically multiple lines. When the designer looks back, the goal is to forget a little that you drew it — to disassociate and look at it fresh. See where the possibilities might be. Turn it around, flip it over, mirror it. See a relationship that might have been missed initially, just for what it might inspire.

Sketches generally sit in sketchbooks — in a drawer, on a shelf. They can be looked up later, even years later. There might be general inspiration found, or a sense of growth, for the designer to see this trail of physical artifacts. A sketch is not really primed to meet the needs of a client, but it is something the designer can feel good about and get a sense of self-worth from.

You described sketching as having a deliberate imprecision — multiple lines laid down, no single solution defined. When you say the marks are intentionally ambiguous, is that ambiguity a conscious choice the designer makes in the moment, or is it more of a natural byproduct of working quickly? Or both — and if both, does the distinction matter?

There are probably different styles of sketching, and designers have evolved different ways of doing the act. But the classic mark — making several sweeping marks across the page, laying down light marks and moving to darker ones, using an eraser to erase out a few directions — those are very deliberate choices to introduce ambiguity. A sketch is not something where the lines are then erased and the perfect ones penned out. That is not a sketch. A sketch has deliberate ambiguity.

You mentioned that a designer can return to a sketch later and see it fresh — almost forgetting they drew it. How does that work when the sketch is digital? Does the screen, the file, the interface change the quality of that re-encounter?

There are three levels of distinction. First, the hand sketch in a sketchbook — you can fan through hundreds of variations in a minute, six sketches on a page, and wait for the ones that catch your eye. Full retinal resolution, tactile response, 100% human. Second, digital representations — thumbnail files that have been seriously reduced in clarity. They become placeholders for a memory of what the sketch was like. The number of images that can be registered in the same amount of time is, in my feeling, less with digital than with paper. Flipping through a magazine or illustrated book exposes the viewer to far more options than scrolling a web page — eye movement, tracking, digital refresh rates, everything is a little inferior. Third, 3D modelling — which brings us to the CAD part of the discussion. Whether 3D modelling, with its required constraints, places the commitment threshold so much higher than either digital or paper methods. [unverified]

You said a 3D model has 'required limitations' or 'required constraints' — can you unpack that? What specifically does CAD force on the designer that a sketch does not?

There is no ambiguity in a CAD model. It is a geometric expression. Points are absolutely determined — a point is a point, a line is a line, a surface is a surface. When something is executed in CAD, I hesitate to even say it is 'sketched.' We have to look at the word 'sketch' here. Is it possible to sketch a part in CAD? With a pre-existing concept, something could be rendered quickly and roughly in a 3D program — it could be sketched. But can a 3D program be used for sketching — to rapidly iterate, build variations, do all the things we've talked about? My belief is that it cannot. You may be able to quickly manifest a rough representation in 3D, but there is little to no ambiguity. The geometry is defined because it's a geometric engine. The word geometry and the word sketching are diametrically opposed in my mind.

Some designers speak of blocking out a rough form in CAD, deliberately leaving things unresolved, working fast and loose. Is that activity meaningfully different from what they're calling it — or is calling it 'sketching' a category error, a misappropriation of the word?

It comes from a slightly different direction in the meaning of the word — to sketch something out in CAD, to quickly rough it out. 'Roughing out' would be a more appropriate term. There is a sense of 'to sketch something' meaning to quickly make it concrete — that's slightly different from 'sketching' as an activity, which means creating something loose and amorphous. I personally find value in roughing something out in CAD, printing faded drawings of those, and doing manual drawing on top of them. There's a sense of proportion gained, a real utility in having an underlying structure that needs to be covered or arranged, with quick possibilities rendered over it. That's a combination approach worth exploring later. Calling it 'sketching in CAD' is a slight misappropriation of the word — maybe there's a linguistic tool that captures the distinction — but that's my feeling.

Imagine a student or early practitioner who has grown up entirely in digital and 3D environments. For them, pen on paper may feel equally intimidating or foreign. Does the direction of intimidation change your argument, or is it asymmetric — is there something lost by avoiding pen and paper that isn't lost by avoiding CAD early on?

When there's a dissatisfaction with the marks of the hand — when they don't feel as stylistically expressive as someone else's in the marketplace, that designer who's legendary for incredible-looking presentation sketches with perfect composition — there's a tendency to move toward augmented approaches: the computer, the digital feel, now AI-generated renderings. These can give a sense of stylistic presentability that a beginning designer, or even an experienced one, cannot render directly. But an experienced designer may not be that kind of sketcher at all. They just sketch to get ideas on paper, to get something in front of another person and have a discussion.

You've separated utility sketching from presentation sketching — but in practice, especially for students and early practitioners, those two things often happen in the same moment, in front of a client or a critic. Is there a stage in the design process where the pressure to present legitimately changes what tool is appropriate — or does the utilitarian sketch always stay upstream of that, protected from presentation pressures?

There are different scales of presentation in the process. In the beginning, a lot of loose variations are expected — possibilities to be explored. No one is looking to lock in an idea with a photorealistic rendering or a finished glossy model. At the very first stages, that kind of resolution cuts off conversation; it limits and hinders rather than opens it up. Then, at various layers, things get tighter and tighter. A sketch is not going to be used midway through a presentation, and not at the end — unless it's one of those killer marker renderings, which is more of a rendering with elements of a sketch. Sketch is ambiguity. As the process goes further, the ambiguity reduces until there is a resolution: a machine-painted model, a clay model, a photorealistic render, a 3D print. That's the hierarchy of the process.

Before mounting the challenge — one more thing. The intimidation of a digitally native person going to a piece of paper is very real. People who are used to doing everything on a phone, or prompting a result, find that putting hand and pencil to paper is an extremely intimidating act. That factor belongs in the record.

Sketching is a dual act — physical and mental — whose defining feature is deliberate ambiguity. That ambiguity is not a limitation but the core mechanism by which the solution space stays open, for the designer and for collaborators. CAD, as a geometric engine, structurally eliminates ambiguity and therefore cannot perform the function of sketching regardless of how quickly it is used. Furthermore, CAD is linear — it progresses toward a single resolution with no natural branching — whereas sketching fans out across many paths simultaneously. And the trail sketching leaves behind is dense by default: hundreds of variations accumulate in a sketchbook almost without effort, whereas CAD requires deliberate action to preserve states, and far fewer are ever saved. Pen on paper is irreplaceable at the early stages not because of tradition or skill, but because the medium itself is ambiguity-native, branch-friendly, and archive-dense. Intimidation by high-skill presentation renders is a category error — it mistakes presentation for exploration — and that intimidation runs in both directions: analog designers fleeing toward digital polish, and digitally native designers too intimidated to put hand to paper at all. Both forms of avoidance carry real developmental cost.

Yes, that will stand for now.

Your argument rests on ambiguity being the essential and irreplaceable feature of sketching — and that CAD structurally destroys it. But consider this: ambiguity in a sketch is only productive if the designer has sufficient visual vocabulary to read it. A line that could be a shoulder curve or a chamfer is only generative if the designer's eye and mind can hold both readings simultaneously. For a digitally native student who has never built that visual vocabulary through thousands of hours of hand drawing, the ambiguous sketch may not open a solution space — it may just look like a mess. In that case, a rough CAD block-out, however linear, might actually scaffold the visual vocabulary they lack, giving them a concrete object to react against, to hate, to push away from — which is itself a form of exploration. Is the productivity of ambiguity medium-dependent, or is it person-dependent? And if it's person-dependent, does your argument assume a designer who has already developed the capacity to read sketches?

It's not just the designer reading the sketch — it's everyone being presented the sketch in a meeting. Visual literacy has a role to play. The sketch is laid out such that some things are obvious; a person with imagination can imagine something in the middle; a person in marketing might see possibilities, or might see it as a very literal representation. Let's phrase it this way: there's a possibility literacy in a good group of people looking at sketches — seeing the openness, seeing the possible variations within the ambiguity, and then pushing toward one of those directions. It's not just the designer, as you posit — it's everybody in the chain. The designer's ability to look at variations and see different things is perhaps 10–20% of the value. Using deliberately ambiguous percentages.

Does that then mean the format of the sketch — its physical presence in a room, passed around, pinned to a wall, fanned through — is not incidental but actually load-bearing to the collaborative function you're describing? And if so, what does a digital file on a screen do to that group dynamic?

Physical format is load-bearing. Paper and printouts — being able to hold something in one's hand — are superior to looking at it on a screen in a presentation room. But that's not a dealbreaker. Plenty of concepts are looked at on screens and evaluated these days. There are advantages to hard copies; studies may show that or may not. We could include citation links later. For now, that position stands. [unverified]

Fine — sketching is constraint-free and that's its value. But here's the deeper problem: a designer who spends their early career sketching almost exclusively on paper may develop a visual vocabulary that is fundamentally 2D-biased — they think in silhouettes, in profiles, in elevations. When they finally enter the 3D phase, they repeatedly discover that forms which read beautifully in sketch are spatially incoherent — they don't resolve from multiple viewpoints, they don't hold up in the round. The CAD advocate would say: starting in 3D, even slowly, even linearly, builds spatial thinking from day one. The sketcher is building fluency in a language that has to be entirely retranslated before it becomes a product. Isn't the fluency you're defending partly a fluency in illusion?

Really good designers work in both mediums from the very beginning — they sketch, then move to CAD. They understand manufacturing. They are neither CAD jockeys nor pencil jockeys. A CAD jockey sits in front of SolidWorks or CATIA or NX pushing parameters around a feature tree — it's a derogatory term. A pencil jockey, or the 'Mac guy,' just sketches in Photoshop and never touches anything real — lets others do the work, sends surfacing off to China. The truly fluent designers use both. There's a time for sketching at the beginning and a time for building into CAD. Thinking too early about how a feature might sit in a tree, how a radius might fail, how a shelling operation might fail from a tangency mismatch — those concerns take away from the process. You'll cross that bridge when you come to it.

You've built your argument on ambiguity, fluidity, and branching — but those are descriptions of process, not outcome. The outcome of early-stage design exploration is a set of directions worth pursuing. A skilled CAD user, working in a program they know deeply, may arrive at that same set of directions faster than a sketcher — because they are not managing a hand, a pen, a sketchbook, and a scanning or photography step to share work. The CAD model is already shareable, already scalable, already manufacturable-adjacent. The sketcher still has to translate their sketch into a 3D object eventually — and that translation is itself a step where meaning is lost, proportions shift, and the ambiguity that felt productive on paper turns out to have been concealing problems that only emerge in geometry. So the CAD advocate would say: the ambiguity you're defending is not a feature — it's debt. You're deferring geometric reckoning, not avoiding it. And the earlier you reckon with geometry, the less waste there is in the process. Why is deferring that reckoning valuable rather than just slow?

The 'debt' framing misunderstands what sketching is for. There is no debt, no downstream investment worry, no concern about shareability or manufacturability inherent in a sketch on paper. The purpose is to explore ideas — to be constraint-free. That's where the possibilities lie. You look at something without constraints, and then those constraints arrive later to define, balance, and refine solutions. Manufacturability comes afterwards. If those concerns are superimposed at the beginning — how do I manufacture this? How do I share this file? How do I make this surface? — they kill the idea. They kill what a sketch is about. The 'debt' you describe is not a factor at all.

Parametric and generative CAD tools — Grasshopper, Fusion 360's generative design, or AI-assisted form exploration — can produce hundreds of variations automatically, branching across a solution space far wider and faster than any human hand. They can even introduce a kind of computational ambiguity — outputs the designer didn't explicitly intend and must now read and react to. Does that class of tool challenge your claim that CAD cannot sketch, or does it represent something else entirely?

Grasshopper and parametrics are genuinely interesting technologies — ones I use myself. But building a Grasshopper definition requires a considerable programming step. The ability to sketch in Grasshopper is extremely difficult. This brings us to something we haven't discussed: barriers. There is very little mental-to-output barrier between a human, a hand, and a piece of paper. Everything above that is a barrier. CAD programs are barriers. Software is a barrier. Operating systems, input devices — all barriers. Each one requires a small mental jump, and that reduces the fluidity of the process. That fluidity of expression — the ability to get ideas out of one's head — is core to the argument. Lock that in. As for mass generation: the ability to produce hundreds or thousands of variations quickly overwhelms the observer if they aren't intentional variations. AI can generate hundreds of images and quickly overwhelm the judge, the contributor. There's a curation step that matters. A designer doesn't walk into a meeting with 200 sketches — they do 200, pick 30, refine to 15 to present. No matter the possibility literacy of the audience, they cannot sort between all those variations.

When a designer fans through 200 hand sketches and selects 15, what is actually happening in that selection? Is it a rational evaluation, an intuitive one, or something else — and does the medium of the sketch affect the quality of that curation act?

The brief defines a set of constraints at the beginning. A design brief is not an equation to be solved — it's a set of constraints to be explored with some looseness. During the curation process, the exploration of the solution space might narrow things down, go outside the constraints, or ignore some of them. In the end, when there's a culling or a curation, the brief is looked at again, and the concepts that most closely meet it — in terms of the intended audience — are selected.

You've described sketching as primarily a generative and exploratory act. But you also mentioned the sketchbook as a personal archive — a trail that gives the designer a sense of growth and self-worth, independent of any brief or client. Are those two functions — the generative tool and the personal record — in tension with each other, or do they reinforce each other?

The archive doesn't really have an effect on the process — it's a nice thing, but the footprints in the snow don't affect what's ahead. They're simply a record. Before moving forward: some of this discussion is a response to certain loud voices in the education space saying that sketching isn't all that important, that everything can be done in 3D CAD and in some cases faster. I firmly disagree with that. Push harder from that standpoint — challenge me from the position that it can all be done in CAD, faster.

Your argument rests on ambiguity being essential. But these are descriptions of process, not outcome. A skilled CAD user may arrive at the same directions faster. The CAD model is already shareable, scalable, manufacturable-adjacent. The ambiguity you're defending is not a feature — it's debt. You're deferring geometric reckoning, not avoiding it. The earlier you reckon with geometry, the less waste there is. Why is deferring that reckoning valuable rather than just slow?

Here is the broadened thesis: "Sketching is a dual act — physical and mental — whose essential feature is deliberate ambiguity. That ambiguity is not imprecision to be overcome; it is the mechanism by which a solution space is held open, explored, and shared. The sketch is a social object as much as a personal one: in the hands of a group with possibility literacy — designers, marketers, collaborators — its openness invites interpretation, generates direction, and keeps conversation alive in ways that a resolved model cannot. This social function depends on the sketch remaining a sketch: the moment its ambiguity is cleaned away, its collaborative utility collapses. Of all the tools available to a designer, the hand on paper carries the lowest barrier between mind and mark. Every layer above it — software, input device, operating system, programming step — adds friction that reduces the fluidity of exploration. Speed matters: a sketch takes thirty seconds; no other medium matches that. And the trail a sketchbook leaves behind is dense by default, a branching archive of variations that accumulates almost without effort, available to be re-encountered, re-read, and mined for overlooked possibility. CAD, as a geometric engine, resolves by nature — it is the right tool when constraints, proportion, and manufacture must be reckoned with. Neither tool replaces the other. The strong designer develops a hybrid fluency: the ability to read what a moment requires, and to move between openness and constraint as the problem demands. Intimidation — whether the analog designer fleeing toward digital polish, or the digitally native designer too intimidated to put hand to paper — is a category error in both directions. It mistakes the tool for the activity, and the activity for the output. To reach for CAD before sketching has done its work is not efficiency; it is the premature closure of possibility. To avoid CAD when geometry must be faced is equally a failure of fluency."

"penned out" replaced with "inked out"

"possibility literacy" bolded as introduced term

Removed sentence: "Using deliberately ambiguous percentages."

"At the high end" replaced with "At the other end of the sketch spectrum"

"ArtStation or Instagram" narrowed to "Instagram"; "classical marker render sketchers" revised to "classical sketch marker renderings"; "possibly" added

"That factor belongs in the record" replaced with "This will only increase as time goes on"

Proposal 2 accepted — continuous prose reorganized around linguistic distinction as entry point

"Debt" framing removed; constraints passage reorganized without AI-interview framing device