Designation attracts two beliefs that are both wrong, and because they are wrong in opposite directions they produce opposite errors. One holds that a reviewer has broad say over whatever appears on the site. The other holds that designation attaches to an elevation and everything behind it is unconstrained. Teams carrying the first concede ground nobody asked them for. Teams carrying the second are corrected, publicly and late.

The boundary itself is narrower and more specific than either belief. Authority attaches to designated fabric and to how new work meets it: what may go, what may change, how an addition reads beside what is already there, and how the whole sits in its setting. It is not general design approval over a separate structure. Where exactly that line falls is set by each designation and by the body holding it, and reading it accurately at the start is worth more than any argument deployed later.

The reason the misreadings persist is that the boundary is not written as a boundary. It is distributed across the designation instrument, the standards the body applies, and a history of decisions that is precedent in practice without being precedent in law. Establishing it is genuine research, and the research is cheap relative to being wrong about it.

Now the other half of this subject, which is reuse, and where the deciding question is almost always the same one: what was this building built to hold. Not what it looked like, and not what it was called. What it was structurally designed to carry, and for what pattern of occupation.

That question decides feasibility faster than anything else, because structural capacity built for one purpose is either generous or inadequate for another and rarely neutral. A building made to carry equipment, machinery or stored goods was given capacity a residential use will never approach, and that surplus is what makes an addition worth investigating rather than dismissing. A building made for people carrying people is a much tighter proposition.

It also shapes the plan, and the plan is where these projects get hard. I have set out what that involves on a specific building elsewhere and will not repeat it here; the short version is that a building organized around reaching machinery is not organized around daylight, and reconciling the two is the whole design problem.

What belongs on this page instead is where those two subjects meet, because they do. Making an interior habitable usually means changing openings. The exterior is usually why the building carries a designation in the first place. So the reuse question arrives back at the boundary question: not whether openings may change, which they often may, but whether a given change sits inside the original compositional logic or reads as an injury to it. That is argued in front of a body, with a consultant, against precedent, and it is not resolved by pointing at a rule.

My arguable position is that designation improves most of the buildings it touches. Not all, and the exceptions are real. But removing the easy options early forces a team to solve the specific site rather than apply a template to it, and the projects I find least interesting are the ones where nothing in the ground pushed back.

The fair counterargument is that review consumes time that cannot be recovered and that work on designated fabric moves at its own pace regardless of what the rest of the project is ready for. On a marginal project that delay is what ends it, and a building that never gets built has not been improved by anything.

The honest limit is that none of the specifics here travel. What is designated, who reviews it, against what standard, and what has been permitted before, are all local and they change. What travels is the pair of questions: where does the authority actually stop, and what was this building built to hold.

The two cases I know best are The Brooklyn Tower, for building above a designated structure, and Walker Tower, for converting one. On either kind of project a preservation consultant belongs in the work before the strategy is set rather than after it.