Planning for a Repair You Cannot Yet Specify

Several records place the availability of a match squarely in the design discussion. The brick slip record asks whether spare facings from the same batch will be kept, and says that decides how an isolated repair will read once the elevation has weathered; it also asks whether the original facings can still be obtained. The rainscreen record asks how a damaged panel would be removed and replaced much later, and whether a matching panel would still exist, calling that a design decision made now.

The composite panel record asks the same question about whole panels, and notes that replacement usually means disturbing neighbouring ones, so access routes and the availability of matching panels are worth establishing early. None of these is answered by a product name; they are answered by whether spares were kept and whether the batch is recorded.

A patch is compared against its surroundings rather than against the original. The rendered insulation record describes repair as matching a texture and a colour that have weathered, and notes that patches are visible in a way they are not on a jointed facade. The exterior masonry paint record makes the same observation about a coating: patch repairs on a weathered elevation rarely match, so partial work tends to remain visible until the whole wall is done again.

Metals behave the same way for different reasons. The copper record states that new material let into old work stands out bright against weathered surroundings and takes a long time to tone in, and the galvanised record notes that a repaired area on visible work is normally still identifiable afterwards. These are properties of how surfaces age rather than of workmanship.

The area that has to be made good is set by the construction rather than by the damage. The direct render record states it directly: because the surrounding face has weathered, the practical unit of repair is often a whole panel between joints, and planning where those panel boundaries fall is planning what a future repair will look like. A facade with joints has decided its own repair units; a continuous one has not.

Bonded and factory-made assemblies raise the same question with different answers. A bonded facing cannot be taken off and put back, so the unit is whatever has to be broken out and re-bedded. A composite panel is bonded through its thickness and cannot be repaired layer by layer, so the unit is the panel. A seamed metal roof is formed from long trays, so a local repair reaches along the tray rather than staying where the damage is.

Some assemblies can only be repaired within their own system, because the geometry that makes them work belongs to the manufacturer. The composite panel record states that panel systems are proprietary arrangements, and that joint geometry, fixings and closures come from the manufacturer's documentation and are not interchangeable between systems. The rendered insulation record adds that mixing components from different sources leaves no one able to describe what the assembly is or what its documentation covers.

That is why the record of what was installed is a repair-enabling document rather than an administrative one. Where the system is identified, a compatible repair is possible; where it is not, a later contractor is working out the geometry from what is in front of them, which is a different exercise entirely.

For applied and bonded layers, the constraint is chemical as well as geometric. The liquid membrane record notes that a repair can be made by preparing an area and reapplying without lifting a sheet, but that the advantage depends on knowing which product was used and whether a compatible material is still available. It also states that where a layer is taken over an older roof, a different membrane, or a metal or masonry upstand, each surface needs its own preparation.

Coatings carry the same dependency. The silicone resin record describes the finish and whatever the maker intends beneath it as documented as a set, so a topcoat sheet read on its own leaves the rest of the build-up undescribed. And a water repellent is described as terminal: nothing is intended to go over it, and it makes later painting, rendering or bonded repair difficult, which is a repair consequence chosen at the time of application.

Read across the records, the difference between assemblies is not whether they can be repaired but how a repair will read. A lapped unit covering is described as allowing individual units to be changed without disturbing the covering as a whole, provided access exists and the fixing method allows a unit to be released. A jointed facade confines a repair to a panel between joints. A continuous coated surface spreads the visible consequence across whatever has to be recoated.

None of those is better in the abstract, and the records decline to rank them. What they do is let the question be asked while there is still a choice: if a panel of this elevation had to be made good in some years' time, what would the smallest acceptable area be, and who would be able to match it.

The records name the handover items that make a later repair possible. For a rainscreen, a record of the control layer, the trays and the bracket penetrations, because cladding covers them permanently. For curtain walling, unit marks, glass types, gasket references and drainage details, without which a later repair begins by identifying what is already there. For a precast facade, panel marks, fixing types and joint details, since a facade of near-identical units is easy to misread on site.

For applied layers, the record of what was applied, over what, and in what conditions is described as the only evidence of what exists. And for concealed work, the record made before covering is what a later owner or surveyor has instead of an inspection. Asking for those at handover is cheaper than reconstructing them.