Elamur Groupantscomputing-io
The groupant beside models/computing-io.groupant in the Mova repository: what the code computes, what it never does, how to go back from its outputs, its invariants, cost and checks.
Facts
Implementation material. A Mova code model describes this code at one source revision: partial models, finite checks and their stated scope. It is not a reviewed general exposition, and its labels are shown as written, never raised.
Components
| component | label | where |
|---|---|---|
| approximant | partial | forward |
| ast | not materialized | |
| composition | partial | composition |
| cost | symbolic source derived | cost |
| descriptant | present | descriptant |
| invariant | partial | invariant |
| tests | see source local evidence | checks |
Obernants
The typed relations around the forward model: the behaviour complement (what the code never does), the reverse reconstruction (every input compatible with an output) and specialised views.
| id | type | label | where |
|---|---|---|---|
behavior-complement | behavior_complement | partial source derived model | behavior-complement |
reverse-reconstruction | reverse_reconstruction | partial source derived model | reverse-reconstruction |
value-transport | representation_transport | partial source derived model | value-transport |
effect-order | effect_order | partial source derived model | effect-order |
failure-precedence | failure_precedence | partial source derived model | failure-precedence |
resource-composition | resource_cost | partial source derived model | resource-composition |
Model text
descriptant
computing-io: wave30
builtins/mod registers files/io; io calls display::format_disp and exp_str; fileio uses table inference/date/name adapters and column-major Mat; main creates Interp with selected fs and selected tests/fs CLI helper observes exit/stdout/stderr/journal/files. sprintf column-major traversal and display row layout are different adapters; JSON orientation/nonfinite loss prevents identity roundtrip. Octave calls lexical filter is not parser validation, and VMM digest helper evidence is not scenario correctness.
This source-bound partial group organizes existing Mova code. Native evidence and excluded effects are recorded in source sidecars and the root wave30 report. Registration is not a universal proof or whole-system runtime claim.
model
computing-io
forward
Registry::with_defaults wires files and io builtin pointers into Interp. Files maps validated char paths to Interp.fs effects/status; io maps Value storage and fmt to String then Interp.out. fileio transports CSV/JSON to typed columns and matrices used by table builtins. CLI make_fs imports before Interp.run; finish_fs exports only on ok, then journal. Selected fs tests invoke this CLI on temporary synthetic data; Octave and VMM harnesses contribute only bounded selector/hash-helper evidence.
invariant
Memory is default CLI/interpreter backend; file builtins do not bypass it. Formatting and serialization themselves perform no filesystem writes. Failed interpreter status suppresses CLI export; successful finalization is required for exit0. Test helpers use bounded synthetic data and frozen baselines stay unchanged.
cost
Parsing has a linear input baseline; formatting adds repeated segment/flag scans up to O(format length * expanded atoms) and output allocation. JSON serialization copies each subtree at every ancestor, bounded by O(depth * output bytes); a singleton-array chain can require quadratic copying. Table key discovery/lookup can cost O(rows*keys^2). VFS/import/export and interpreter costs are delegated; external subprocess and frozen-suite costs remain unmeasured and unexecuted.
composition
builtins/mod registers files/io; io calls display::format_disp and exp_str; fileio uses table inference/date/name adapters and column-major Mat; main creates Interp with selected fs and selected tests/fs CLI helper observes exit/stdout/stderr/journal/files. sprintf column-major traversal and display row layout are different adapters; JSON orientation/nonfinite loss prevents identity roundtrip. Octave calls lexical filter is not parser validation, and VMM digest helper evidence is not scenario correctness.
domain
Valid Value/Mat/Table storage, UTF-8 input, synthetic in-memory or owned temporary filesystem fixtures, bounded output widths and available resources. CPU only.
limitations
Delegated VFS and interpreter contracts are required; allocation and host fault injection absent. No external Octave corpus, VMM400 evaluation, benchmark or Monte Carlo run.
checks
- 13 exact named tests twice debug/twice release; identical sorted outcomes
- 9 compiled assertion mutants across 9 sources; byte-exact restore and positive rerun
- 18 fresh syntax projections paired equal; source-scoped nested byte/hash bindings
catalog
behavior-complement
Interp failure excludes CLI export_dir but permits journal and previous imports; VFS error with requested status excludes throw for mkdir/rmdir/move/copy. J::Num/numeric-class nonfinite conversion excludes the original numeric literal; logical NaN instead converts to true under the nonzero predicate. No runtime scenario guarantee follows from helper checks.
reverse-reconstruction
Printed output maps to sets of Value/format antecedents, serialization admits missing/null and shape ambiguities, exit codes admit multiple failure stages. Joint journal and filesystem observations narrow compatible histories but do not reconstruct all interpreter steps.
value-transport
sprintf flattens m.re column-major, display accesses j*rows+i into row strings, JSON vectors collapse orientation. Matching scalar numbers is insufficient to prove shape-preserving adapters.
effect-order
parse FS flags -> open/import -> run -> conditional export -> journal. Export failure can suppress journal; failed script skips each export but still attempts journal. Delete loops preserve earlier removals on later error.
failure-precedence
CLI limit parsing precedes mem/import compatibility; atom validation precedes output; CSV syntax precedes table-width validation. These branches must be preserved across caller wiring.
resource-composition
Captured strings plus serialized buffers plus VFS memory add rather than cancel; repeated format/flag scans, requested precision/width, JSON subtree copying and key lookup amplify input cost. No measured latency or global cap is supplied.
This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/models/computing-io.groupant/