Level of Development, Not Detail

The word people hear is 'detail', and it sends them the wrong way — toward visual richness, toward modelling every screw. The word that matters is development: how far an object's geometry *and* its information have progressed toward being reliable enough to act on.

Detail is about appearance. Development is about *trust*. A pretty, detailed object can be low development if its geometry is approximate and its data is placeholder — you cannot yet build from it or count on it. A visually crude object can be high development if its size, position and key data are confirmed and specific — you can rely on it completely. The whole point of an LOD is to answer one question honestly: how far can I trust this object right now? Not 'how good does it look?' Confusing the two is the field's most common and most expensive mistake, because a model that looks finished but is not *developed* invites decisions the model cannot actually support.

The widely used scale — codified in the BIMForum LOD Specification, which is refreshed regularly — runs in bands, and it is worth knowing in plain language rather than by number alone.

LOD 100 is symbolic: the object is a placeholder standing for 'something is here' — a mass, an area, an early idea; take no dimensions from it. LOD 200 is generic and approximate: recognisably a wall or a pump, with approximate size and position — good enough for coordination in principle, but not for ordering. LOD 300 is specific and reliable: accurate size, shape, location and orientation, with real data — the level at which drawings and quantities can be trusted (and, tellingly, the level India's CPWD guidelines require at coordination milestones). LOD 350 adds the interfaces between systems — connections, supports, the details that make clash detection between trades meaningful. LOD 400 is fabrication-ready: detailed enough to manufacture and install from. LOD 500 is the verified as-built record — confirmed against the real, constructed thing, the basis for operation.

Notice what climbs as you go up the ladder: not prettiness, but *reliability and specificity*. Each rung is a promise about how much you can trust the object — which is why a low-rung object should look provisional, and a high-rung object should be believed.

First, geometry and information can develop at *different rates*, and pretending they move together causes real errors. An object can have precise geometry but placeholder data, or verified data on rough geometry. This is exactly why modern practice increasingly separates the two — talking about the reliability of the geometry and the reliability of the information as distinct measures, rather than one blunt 'LOD' number. When someone says 'LOD 300', the honest question is: developed to 300 in geometry, in data, or both?

Second, higher LOD is not better — it is only appropriate. Developing an object beyond what a decision needs is waste: slow, expensive, and falsely reassuring. A design-option study needs LOD 200, not 400; forcing fabrication-level development onto every object early burns time modelling detail nobody will use and can even *hide* the design intent under premature specificity. The right LOD is the one the current purpose requires — which is precisely the subject of the next lesson. LOD is a language for stating *how much to trust*, and its discipline is matching development to need, never maximising it.