Arabic mathematics is more than translation: preserving an expression’s meaning

The problem begins before font selection

Consider a fraction embedded in an Arabic explanation, followed by substitution and a variable. Each element may look reasonable alone, yet punctuation and surrounding direction can change how a learner reads the whole. When the line wraps on a phone, it may become unclear which terms belong to a denominator or exponent.

We therefore approach mathematics as structured content. The task is not simply to make symbols attractive, but to preserve relationships between their parts through rendering. A new font can change a glyph’s appearance; it cannot repair grouping that the representation has already lost. The question, expression and explanatory context must be considered together.

Two layers with different responsibilities

The representation describing an expression’s meaning should be separate from its presentation. Textbooks may use different notation conventions without changing the mathematical concept. Conversely, similar-looking strings do not guarantee that a system retained the same underlying meaning.

Warda separates mathematics rendering into its own path, with presentation rules connected to educational context. This separation lets us review display without reinterpreting content in every screen. It also helps locate a defect: source content, conversion or rendering. It does not mean every possible expression is supported.

Page direction does not determine every expression

MathML Core defines direction for mathematics and discusses Arabic contexts. Paragraph direction alone therefore cannot describe an expression’s layout; implementation still needs visual checking on target devices. Original specification.

Our product decision is to let expression meaning and textbook conventions guide presentation, rather than applying a blanket reversal. Arabic text inside an expression, variables, fractions and groupings need distinct treatment. An inline expression also has different space and reading requirements from a long transformation displayed on its own line.

A visual patch versus retained structure

Separation does not remove content review. An incorrectly transcribed source expression remains incorrect even when rendered beautifully. Source integrity precedes presentation quality. When conversion is unsuccessful, an understandable failure state is preferable to displaying an expression whose meaning has not been preserved.

Representation serving explanation and verification

Separating representation and display makes an expression something that can be referenced and examined across a question, an explanation and practice. Presentation rules remain reviewable without distributing symbol handling across screens. Textbook notation and visual review belong to content quality itself.

Faithfully displaying an expression differs from judging a step within a solution. Warda’s vision connects the two without mixing their responsibilities: representation preserves meaning, while verification interprets it under task conditions. Our mathematical verification article explores those connected boundaries.