docs: add ROR integration assessment (not recommended) (#1039)

Comprehensive evaluation of ROR (Research Organization Registry) integration
based on PR #1034 implementation and independent analysis.

Key findings:
- Fundamental mismatch: ROR identifies institutions, not journals
- Catastrophically low coverage: <10% of journals (mostly university presses)
- Weak signal: institutional affiliation ≠ journal quality
- Massive complexity: 2,437 LOC, 7 DB tables, 140MB+ data
- Better alternatives exist (enhance OpenAlex/Crossref analyzers)

Recommendation: DO NOT IMPLEMENT

PR #1034 demonstrated full implementation with real-world testing.
Cost-benefit ratio extremely unfavorable. This assessment documents
the decision to close PR #1034 without merging, preserving the branch
as reference implementation if conditions change.

Related: #1034

Co-authored-by: florath-ai-assistant[bot] <Andreas.Florath@telekom.de>

Author: coding-ai-assistant[bot]