Anti-Patterns and Review Interventions

Page Maps

graph LR
  family["Reproducible Research"]
  program["Deep Dive DVC"]
  section["Migration Governance Dvc Boundaries"]
  page["Anti-Patterns and Review Interventions"]
  capstone["Capstone evidence"]

  family --> program --> section --> page
  page -.applies in.-> capstone

flowchart LR
  orient["Orient on the page map"] --> read["Read the main claim and examples"]
  read --> inspect["Inspect the related code, proof, or capstone surface"]
  inspect --> verify["Run or review the verification path"]
  verify --> apply["Apply the idea back to the module and capstone"]

Anti-patterns rarely announce themselves as bad ideas.

They usually begin as reasonable shortcuts:

• "just this one local file"
• "copy the latest model"
• "we can document the parameter later"
• "the remote probably has it"
• "the experiment result is obvious"

Module 10 asks you to stop those shortcuts early.

Common DVC anti-patterns

Anti-pattern	Why it spreads	Review intervention
path as identity	filenames feel concrete	ask for content identity or DVC evidence
hidden parameter	quick local tuning	move meaningful controls into reviewed params
metadata-only merge	Git diff looks complete	require remote-backed verification
copy-latest promotion	easy release movement	require versioned bundle and manifest
experiment as branch substitute	convenient exploration	require intent, comparison, and disposition
cleanup by storage pressure	cost feels urgent	require retention and dry-run review

The intervention should name the missing contract.

Intervene with a repair path

Weak:

This is an anti-pattern.

Stronger:

This promotes outputs/latest/model.json, which gives consumers a moving target. Please publish a versioned bundle with manifest, params, metrics, and review note.

The stronger comment teaches the desired state.

Avoid anti-pattern inflation

Not every imperfection is an anti-pattern.

Reserve the term for habits that damage state identity, reproducibility, comparability, promotion trust, or recovery. Otherwise the review language becomes noise.

Ask:

• would this make a future result harder to explain?
• would this make recovery harder?
• would this make consumers depend on unsupported internals?
• would this hide a meaningful control change?

If yes, intervene.

Review checkpoint

You understand this core when you can:

• identify a shortcut that damages a state contract
• explain why the shortcut is tempting
• write a review intervention with a repair path
• avoid labeling harmless style differences as anti-patterns
• connect anti-patterns to course contracts

Anti-pattern review is not scolding. It is contract preservation.