STRATA
Skip to study
XLVII

OPEN QUESTION / Open

Compression, reopened

Open questionR01 · 2026-11-17
Scope of this record

An XLVII research question and proposed protocol. No experiment has run.

Statement

Under fixed memory and latency budgets, where can compression-based classification remain useful? [source]

Equation

Unrecorded

Not established. A useful application would not establish a universal identity.

Observation

Begin with the linked historical application; preserve its source and stated scope.

Progressive questions

Does the result persist under a documented temporal split, distribution shift, and matched resource budgets?

Responses

No new measurements have been collected. Dataset, toolchain, and baselines remain to be pinned.

Interpretation

Possible usefulness is a hypothesis. Age, reputation, or presentation neither establishes nor rules it out.

Conclusion

Open. The missing measurement is retained as the next study, not filled with an assumed outcome.

Application

Proposed: preregister the split, byte preprocessing, update order, baselines, false-positive constraint, latency, peak memory, and uncertainty method. Publish input hashes, parameters, outputs, and limitations. Do not inspect test labels while choosing the protocol.

Work through a related example.

IV.15 / Compression-based classification and contemporary reuse

Compare two class-conditioned descriptions.

Train two illustrative text models on separate, labeled collections and score a held-out message under each. Record which model assigns the shorter effective description and whether headers or repeated templates dominate that difference. Split related messages together when forming training and test sets so that near-duplicates do not make the evaluation artificially easy.

V.12 / Recovery of abandoned, overlooked, or repurposed work

Evaluate a rediscovered method against a current baseline.

Choose an older, documented method and a small task whose requirements are explicit. Reconstruct the original input assumptions, then compare it with a simple contemporary baseline using the same data and resource accounting. Record where the older method's constraints are beneficial, irrelevant, or incompatible with the task.

Follow earlier and later work.

Before

After

No linked record.

Follow the chain
  1. Compression → classification · 1 link away · UNKNOWN
  2. What survives a replication? · 1 link away · UNKNOWN
  3. Dynamic Markov modelling · 2 links away · KNOWN
  4. Information is not meaning · 2 links away · UNKNOWN
  5. Compute has a budget · 2 links away · UNKNOWN
  6. Memory as a collective state · 3 links away · UNKNOWN
  7. Erasure has conditions · 3 links away · UNKNOWN
  8. Attention without recurrence · 3 links away · UNKNOWN
  9. Prediction and residual error · 4 links away · UNKNOWN
  10. Reversibility and the record · 4 links away · UNKNOWN
  11. The chemical compass · 5 links away · UNKNOWN

Unfinished.

Present-day utility

A versioned benchmark with specified data, budgets, baselines, and error constraints.

Change the tested conditions explicitly; retain the earlier scope.

Placement.

Technology

Reconsideration is not rehabilitation of unsupported claims.

Sources.

  1. Spam Filtering Using Statistical Data Compression Models

    Journal of Machine Learning Research · Volume 7(97), 2673–2698 (2006); abstract.

    Journal metadata and abstract read; no independent replication performed. Checked 2026-11-17.

Record history
  1. 2026-11-17 / R1 / CREATED

    Open research question created. No execution claimed.