Arabic writing was never confined to the page. From the first century of Islam it appears on coinage, on milestones, on the walls and domes of mosques, on textiles, on metalwork and on ceramics, and in each of those settings it did work that images did elsewhere. The reform of the coinage under the Umayyad caliph Abd al-Malik, which replaced figural types inherited from Byzantine and Sasanian models with purely epigraphic ones, is the clearest single instance: the writing became the design.
Architectural inscription developed its own conventions. Bands of angular Kufic run along cornices and around domes at a scale that makes them legible from the floor, and their letters are stretched, squared and interlaced to fill the space exactly, so that reading them requires knowing what they are likely to say as much as decoding the strokes. Later hands, thuluth above all, brought the rounded cursive on to walls at monumental size. In tile and in carved plaster the same text may be repeated as a border, mirrored, or worked into a knot in which the letters are barely separable from the ornament around them.
Two things follow for anyone using such an inscription as evidence. The first is that architectural writing is a designed object under compositional pressure, so letter forms in it are not a reliable guide to the ordinary hand of the same date. The second is that these inscriptions carry Quranic text, which is why the atlas records what a programme of inscription does and where it is placed, and describes rather than reproduces what it says. Facts about the placement, the hand, the patron and the date are for the atlas to record; the text itself belongs to the tradition that transmits it.