Pushing the Raman principle
Clayton Lewis · 2013 · Proceedings of the 10th International Cross-Disciplinary Conference on Web Accessibility (W4A) · doi:10.1145/2461121.2461147
Summary
This paper explores the implications of what the author calls the "Raman Principle," attributed to T.V. Raman: "The way to think about the visual system is as a way to answer queries against a spatial database. If you have an alternate way to ask the queries and get the answers, you don't need the visual system." Lewis demonstrates this principle through Noodle, a visual programming system that supports non-visual operation. The paper was inspired by a panel exchange where a blind computer scientist (Jamal Mazrui) called for technology enabling people with disabilities to create their own applications, while another panelist (Gregg Vanderheiden) advocated visual programming — neither noticing the apparent contradiction. Lewis argues that the Raman Principle resolves this: rather than trying to make a visual interface accessible through screen reader rendering, one should identify what questions users need answered and provide alternative ways to ask and answer them. Noodle uses a dataflow programming model where programs are represented as interconnected blocks, with a declarative JSON representation that is independent of visual layout. The non-visual interface uses single-key keyboard commands ("b" for blocks, "c" for connectors, "e" for edges, "s" for select, "d" for delete) with spoken output to navigate and manipulate the program structure.
Key findings
The non-visual interface for Noodle was actually easier to implement than the visual one, because it avoided the substantial detail work needed for visual layout — positioning blocks, drawing data paths between them, and computing geometric shapes. This counterintuitive finding supports the Raman Principle's implication that visual rendering often adds complexity without being essential to the underlying task. Lewis notes that the Raman approach differs fundamentally from the screen reader approach: rather than attempting to render visual representations in audio or tactile form, it provides non-visual support for the operations (navigation, selection, manipulation) that the visual representation supports. The paper also observes that the Raman Principle aligns with good instructional design — both require explicitly identifying what information users need to extract from an interface, rather than assuming visual presentation inherently communicates. At the time of writing, neither interface had been tested with users beyond the author.
Relevance
The Raman Principle offers a powerful conceptual framework for accessibility practitioners working beyond traditional text-based web content. As interactive content proliferates — data visualisations, maps, educational simulations, games — conventional accessibility approaches based on screen readers and alternative text are increasingly insufficient. The principle reframes the challenge: instead of asking "how do we make this visual thing describable?" we should ask "what information does the user need, and how can we provide it?" This distinction is particularly relevant for complex interactive applications where a screen reader linearisation of the visual interface loses the structural relationships that make the interface useful. The paper also raises an important equity issue: if programming tools are visual, blind developers are excluded not just from using software but from creating it — a deeper form of digital exclusion that limits self-determination.
Tags: visual programming · blindness · non-visual interaction · screen readers · programming accessibility · inclusive design