User interfaces are opaque and routinely leave existing useful functionality unexposed. When using applications that do not entirely meet your needs, rarely will any amount of computational literacy on your part allow you to bridge those utility gaps at runtime. In addition, the flat text authored or consumed within those UIs is unstructured, inert, and brittle. When working with human-readable representations, no amount of literacy on your part will allow you to express detail or context beyond what has been provisioned by its formal grammar. This thesis identifies a systemic cause, inherited from the origins of human-computer interaction, for the divide between user interfaces, human-readable formats, and the application models themselves. I will propose a building block that has the potential to act as all three, and explore the rippling implications it could have on computing - making it more adaptable (receptive to existing literacy), making it more transparent (providing more natural avenues to acquire that literacy), and making it more layered (allow more degrees of literacy to be independently relevant).