Region and Word Boundary Calculation

Region geometry

For each IA row, EyeSort starts at the user-supplied sentence offset. Region width equals the literal string length multiplied by pixels per character. Left-to-right regions advance rightward; right-to-left regions advance leftward. Only horizontal X position is used.

Whitespace is preserved and contributes to width. EyeSort does not impose a leading-space convention: put spaces in the region cells exactly where the desired boundary calculation requires them.

Boundary ownership

In left-to-right mapping, ordinary regions use half-open intervals [start, end), so a fixation exactly on a shared edge belongs to the following region. The final region includes its outer edge. Word boundaries are also calculated from text spans and stored for metadata.

Assumptions

Geometry is deterministic only when sentence offset, fixed-width font, font size, display scaling, resolution, and uniform pixels per character reflect the actual experiment. Multiline text, proportional fonts, or variable character widths are outside the verified model.

EyeSort calculates current_word and current_word_text, but the 1.0 label GUI selects named regions rather than each word independently.

Always validate the table from Inspect Parsed Regions against stimulus records.