refactor: Rename normalization flags to normalized

[krlmlr]

Part of #2788 (PR 4 of the stack, on top of #2861); executes the "normalized" row of the survey — nine functions already spell this flag normalized, and these are the three stragglers.

What this does

• power_centrality(): rescale → normalized (pure registry rename; the arg sits after ...).
• tk_coords(): norm → normalized (pure registry rename). Abbreviations n/no/nor now reject as ambiguous between the two spellings; the retired exact norm = is recovered with the soft-deprecation.
• hits_scores(): scale → normalized. This function uses the strict check_dots_empty() style (no registry entry), so the rename keeps a trailing scale = deprecated() sentinel with an in-body bridge: soft-deprecation, value forwarded, and an error when both spellings are supplied. The soft-deprecated hub_score()/authority_score() keep their frozen scale formals and forward to normalized; same for bonpow() (rescale) and tkplot.getcoords() (norm), each with a local deprecated-badge @param where @inheritParams no longer covers the old name.

Deliberately untouched

• eigen_centrality(scale = ), centr_eigen(scale = ), centr_eigen_tmax(scale = ) — already-neutered deprecated() formals (normalization is always on); nothing to rename.
• Coordinate-rescaling rescale in plot.igraph() and the embedding functions' scaled — different semantics.

🤖 Generated with Claude Code

https://claude.ai/code/session_01RTPj4qNv2FWui6etixZeuR

---

Generated by Claude Code

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if f27a205 is merged into claude/issue-2788-subagents-faw1t1-3:

• ✔️as_adjacency_matrix: 806ms -> 805ms [-2.7%, +2.35%]
• ✔️as_biadjacency_matrix: 815ms -> 825ms [-0.92%, +3.27%]
• ✔️as_data_frame_both: 1.69ms -> 1.7ms [-3.15%, +4.12%]
• ✔️as_long_data_frame: 4.35ms -> 4.29ms [-7.42%, +4.34%]
• ✔️es_attr_filter: 3.04ms -> 3.08ms [-5.17%, +7.94%]
• ✔️graph_from_adjacency_matrix: 131ms -> 130ms [-4.89%, +3.46%]
• ✔️graph_from_data_frame: 4.1ms -> 4.02ms [-7.16%, +3.22%]
• ✔️vs_attr_filter: 1.64ms -> 1.69ms [-1.41%, +7.38%]
• ✔️vs_by_name: 1.05ms -> 1.04ms [-4.35%, +2%]
  Further explanation regarding interpretation and methodology can be found in the documentation.