Using the format inventory - finding your format and your path

1 · Learn the move · Tool choice + routing

Every format question you will ever have has the same three-part answer, and routing is the skill of getting to it fast. Can the model read this directly? What converts it? What do I hand the model afterward? The routes sort into four families. Direct-readable: csv, tsv, plain text - go straight to the slice. Structure-only: open standards like mzML, where the container parses but the payload needs tooling - the model writes the extractor. Vendor binary: .raw, .czi, .wiff, .fcs - Move 1, always; the model authors the converter, msconvert and Bio-Formats do the work. User-configured exports: the treacherous family, because two labs on the same software produce different column sets - never assume a schema, always read one real file first. Answer the three questions once per format you touch, write the answer down, and the wall you hit in week one becomes a lookup table by week four.

Route these formats for me: [list your formats with producing instrument]. For each, answer exactly three questions: (1) DIRECT - readable by a model as-is: yes / structure-only / no; (2) CONVERT - the standard open tool that produces an open format, named tools only, and say if you are unsure about my specific variant; (3) HAND - the reasoning slice for this format: shape, schema, sample, control, specific to what this data type is. Output one block per format, so I can paste it into my lab's format lookup doc.