Back to dashboard
`docs/design/search-quality-eval.md` §14.2 Phase 1.1 (deprecation-aware), query class E from §1.4

Search-quality baseline: deprecation-aware (Phase 1.1, v1.2.0 candidate)

This audit answers a focused question: when a developer queries a concept that exists in both modern Swift form (value type / stdlib) and the original NS-prefixed Objective-C form, does cupertino's ranker put the modern form above the legacy form? This is the most user-visible failure mode for an AI coding agent (Criterion 2: an agent told to use NSURLConnection when URLSession is the right answer in 2026 generates wrong code).

Measured 2026-05-20·Strong

Headline result
30 / 30 (100%)

Read in detail

Each card opens its own page. The headline and charts above are all you need at a glance; the cards are for the why and how.

Sources cited in this measurement

Every metric and method this audit relies on, with a link to the foundational source. Auto-collected from the audit text.

Reciprocal Rank Fusion (k=60)

Cormack, Clarke, Büttcher (2009), SIGIR

Open citation

Mean Reciprocal Rank

Voorhees (1999), TREC-8 QA Report

Open citation

P@k (Precision at k)

Manning, Raghavan, Schütze (2008) IIR §8.4

Open citation