Skip to content

Applicability and roadmap

crushr is best understood as a serious preservation utility for workflows that value verifiable post-damage handling of imperfect result sets. It is not trying to be the default answer for every routine packaging task.

Fit

crushr aligns well with:

  • post-recovery preservation of damaged or partial result sets
  • archival and preservation workflows
  • corruption research
  • long-lived storage experiments
  • ransomware / DFIR aftermath where recovered outputs are mixed-quality
  • any environment where the difference between clean extraction failure and still-provable preserved output matters operationally

Current limits

The architecture is now strong enough to justify formal documentation and productization, but the project is still in 0.x.

That means:

  • the product thesis is being proven and stabilized now
  • benchmark evidence and compression/performance characterization are next
  • metadata envelope completion is still ahead
  • evidence/custody/signing features are future extension work, not present-day core-format claims

The current product should therefore be described honestly as:

a deterministic preservation format for imperfect, derived data

—not as a finished evidence/custody platform and not as a universal archive replacement.

Roadmap logic

The sequencing is deliberate.

0.x — prove and stabilize the product

  1. Finish product-surface hardening and unified CLI/operator identity.
  2. Implement the benchmark harness and generate quantitative evidence.
  3. Use measured benchmark results to guide compression and performance work.
  4. Complete the bounded metadata and verify/report envelope needed for a serious preservation product.

1.x — stabilize the preservation platform

  1. Lock stable workflow-facing contracts.
  2. Mature packaging, reproducibility, and documentation.
  3. Make crushr predictable enough for external workflows to rely on.

2.x — extend toward evidence-grade workflows

  1. Add sidecar evidence manifests, signing, and custody-event tracking.
  2. Expand internal deterministic classification only where provable.
  3. Keep evidentiary ceremony layered on top of the stabilized platform rather than prematurely baking it into the core format.

Legitimacy does not require total completeness. A format becomes legitimate when it has a defensible purpose, a coherent architecture, a documented trust boundary, measurable evaluation evidence, and a roadmap that respects what is solved now versus what belongs later.