What Parametric Means

Start by separating two things people constantly confuse: a *drawing* and a *model*. A drawing is a record of decisions already made - this street here, that block there, fixed in place. A parametric model is a set of *instructions and relationships* that will produce a drawing once you supply the numbers. The word parametric simply means "controlled by parameters", and a parameter is any input you have chosen to leave adjustable: block size, street width, density, setback, height, the angle of the grid, how far density decays from a centre. Fix all the parameters and the model spits out one specific plan. Leave them adjustable and you hold a whole family of plans in your hand, indexed by those numbers.

The crucial shift is where the information lives. In a hand drawing, meaning is in the marks - erase the marks and it is gone. In a parametric model, meaning is in the *logic*: the statement that "a block is bounded by streets of width W spaced distance D apart, and each plot within it is set back S from the street and rises to height H" is true no matter what W, D, S and H happen to be right now. The visible plan is just today's readout of that logic. This is why a parametric model feels less like a picture and more like a small piece of reasoning about how a piece of city is put together.

That reframing has real consequences for practice. It means the hard, valuable work moves upstream, from drafting to *deciding what the relationships are* - which quantities matter, how they depend on one another, what is a driver and what is a consequence. A good parametric urbanist spends their thought on the structure of the model, not on redrawing. It also means the model can be honest or dishonest in a way a single drawing cannot: a drawing shows one outcome and hides its reasoning, while a parametric model exposes its assumptions as explicit relationships you can inspect, challenge and change. That exposure is the best thing about the method - and, as we will see, the assumptions you *forgot* to expose are the most dangerous thing about it.

The heart of parametric thinking is the link between a parameter and the form it governs. In a well-built model that link is *associative*: elements are defined in terms of one another rather than pinned to fixed coordinates, so a change to one input propagates through every element that depends on it. Picture a simple neighbourhood model with one exposed knob, block size. When you increase it, the model does not just enlarge one square. It respaces the street grid, so there are fewer, wider blocks; it re-subdivides each block into plots by the plot-depth rule; it recomputes how many dwellings fit, and therefore the density; it may re-route a bus line that was pinned to a certain street spacing. One number moved, and the whole plan re-formed consistently, every part still obeying the rules you wrote.

This is what people mean when they say a parametric model "re-forms". It is not animation for its own sake; it is the guarantee that the plan is always internally consistent with its own logic. You cannot end up with a plot that no longer touches a street, or a density figure that contradicts the block layout, because those relationships are enforced by construction. Compare that to editing a drawing by hand, where changing the block size means a hundred small manual edits, any of which you might forget, leaving a plan that quietly contradicts itself.

But notice what the associative structure also does: it makes your assumptions *ride along invisibly* on every variant. If your model assumes plots always face the street, then bigger blocks give you bigger street-facing plots - and never the deep courtyard housing or the informal cluster that does not face a street at all, because that possibility was never in the relationships. The knob faithfully re-forms the plan within the family, and just as faithfully refuses to leave it. This is the double edge to hold from the first model you ever build: associativity gives you consistency and speed, and in exactly the same motion it locks in everything you assumed. A parameter is a promise that this quantity can vary; it is also a quiet declaration that everything you did *not* make a parameter is fixed forever.

Once you see a parametric model as a family of plans indexed by parameters, the natural next question is: how big is the family, and where are its edges? Think of every parameter as an axis. Two parameters - say block size and density - define a plane, and every point on that plane is one specific plan the model can produce. Add setback, height, grid angle and the rest, and you have a high-dimensional *design space*: a vast but bounded region containing every plan your model can reach and no others. Parametric exploration is the act of moving through this space - sampling points, comparing the plans they produce, steering toward the ones that perform or feel right.

The reach is genuinely impressive. A handful of knobs can define millions of distinct plans, far more than any team could draw, and you can traverse them in an afternoon: sweep density from low to high and watch the trade-offs appear, or fix everything except grid angle and study how orientation changes sunlight. This is the legitimate thrill of parametric work - you stop testing three schemes and start understanding a continuum, seeing not just isolated options but the *shape* of how the design behaves as its drivers move. That understanding, the felt sense of "as I push this, that gives", is often more valuable than any single plan it produces.

The boundary is just as important, and easier to forget precisely because the interior is so rich. The design space contains only what your parameters and relationships allow. A brilliant plan that violates one of your baked-in assumptions - a fabric that mixes grid and organic growth, a block shaped by a sacred route or an existing settlement, a use you did not encode - is not a distant corner of the space; it is *outside the space entirely*, invisible and unreachable, no matter how long you search. The model gives you total freedom within the family and zero freedom beyond it. So the deepest parametric skill is not turning the knobs well; it is choosing the family wisely, and staying honest that the map is not the territory - that the real city has dimensions you did not, and perhaps could not, put on an axis.

It is tempting to treat a parametric model as a neutral tool - just geometry, just math. It is not, and pretending otherwise is where trouble starts. Every parametric model is an argument about what matters, made in the language of which quantities became parameters and which relationships were written down. Those choices are never complete and never neutral. You parameterise what you can measure and name - block size, density, height, FSI - because those are the things a number can hold. The things that make urban fabric humane and just are mostly *not* on that list: the fine-grained mixture of a bazaar street, a community's attachment to a particular route, the way an informal settlement adapts to its ground, the dignity of a home that does not fit the plot rule. A model does not exclude these out of malice; it excludes them because they resist becoming a knob, and so they quietly fall outside the family from the start.

This is the parametric face of the course's central trap. The measurable becomes adjustable and therefore central; the unmeasurable becomes invisible and therefore sacrificed. Worse, the ease and consistency of the model lend it a false air of objectivity - the plan re-forms so cleanly that it feels like discovered truth rather than the readout of one person's assumptions. And the assumptions carry power: whose priorities set the parameter ranges, whose comfort the model optimises, whose existing city fits its categories and whose is erased for not fitting. In the Indian context especially, where so much of the living city is informal and organic, a parametric family built only from formal, measurable rules can literally not represent the places where most urban life happens.

None of this makes parametric modelling wrong; it makes it a tool that must be used with its critique attached. Use the model to explore, to test what-ifs fast, to keep a plan internally consistent, to understand how drivers trade off. But hold three disciplines: name your assumptions out loud so they can be challenged; keep asking what the family cannot contain; and remember that the binding choices about a city's future - what to build, for whom, at what cost to whom - are human, political and democratic, belonging to the planning authority, the participatory process, the affected communities and the governing law, not to the knob you find most satisfying to turn.