ECO is the opening-classification layer. It has two public halves: Code maps ECO text such as B20 or C50a1 to a compact numeric value, and Book loads a Scid ECO text file and classifies exact Position values.

Code is the small value model. It preserves Scid's extended ECO forms, provides basic and extended display strings, expands prefixes to their final subcode, and reduces full codes to database-statistics buckets.

Book is the classifier. It reads the ECO source file, replays each line from the standard starting position, stores the final compact board, and later answers exact-position lookups or prefix listings.

ECO touches Database at the metadata boundary. Database records carry a compact ECO classification for browsing, sorting, searching and statistics, while the ECO library supplies the public code vocabulary and the position classifier that can produce those classifications.