feat(lint): a controlled topic tag only when one fits the post

_data/post_conventions.yml

@@ -48,8 +48,9 @@ math:
   defined_macros: [llbracket, rrbracket, textsc]
 
 tags:
-  # Every post carries at least one: it is what _plugins/related_posts.rb joins
-  # posts on. A tag only one paper can have connects nothing.
+  # A post carries one when one names its topic: it is what
+  # _plugins/related_posts.rb joins posts on. A tag only one paper can have
+  # connects nothing.
   controlled:
     - Agentic-AI
     - Alignment

_plugins/related_posts.rb

@@ -19,8 +19,8 @@
 # better than a wrong suggestion.
 #
 # That strictness depends on tags that can meet. Per-paper contribution tags
-# rarely do, so each post should also carry a controlled topic tag (the list is
-# `tags.controlled` in _data/post_conventions.yml) — a looser rule here is not the fix for a post with no suggestions.
+# rarely do, so a post should also carry a controlled topic tag that fits it
+# (the list is `tags.controlled` in _data/post_conventions.yml) — a looser rule here is not the fix for a post with no suggestions.
 #
 # Jekyll's built-in site.related_posts is either "the 10 most recent posts" or
 # LSI, which needs the classifier gem and a slow indexing pass. This is cheaper

docs/writing.md

@@ -155,11 +155,12 @@ Tags are free-form and hyphenated, and come in two layers:
 
 - **Specific tags** name the post's own contribution
   (`Fine-Grained-Expert-Segmentation`). Precise, but they connect nothing.
-- **At least one controlled topic tag** from `tags.controlled` in
-  `_data/post_conventions.yml`. "Related reading" needs a shared tag, so a post
-  with only specific tags ships without it.
+- **A controlled topic tag** from `tags.controlled` in
+  `_data/post_conventions.yml`, when one fits. "Related reading" needs a shared
+  tag, so a post with only specific tags ships without it; that is better than
+  a topic tag the post is not about, which suggests unrelated posts.
 
-Add the topic tag; don't swap it for the specific ones. Reuse an existing tag
+Add the topic tag that fits; don't swap it for the specific ones. Reuse an existing tag
 before coining one: three names for one idea leave every post with a tag no
 other post shares. When two tags turn out to be one, keep one and move the
 other to `tags.aliases`; the lint then rejects the retired name. Renaming a tag

script/lint-posts.rb

@@ -185,7 +185,7 @@ def self.tags(post, spec)
     rules = spec["tags"]
     tags = Array(post.data["tags"]).map(&:to_s)
     found = []
-    found << "1: no controlled topic tag (one of #{rules["controlled"].join(", ")})" if (tags & rules["controlled"]).empty?
+    found << "1: no tags" if tags.empty?
     tags.each do |tag|
       if (replacement = rules["aliases"][tag])
         found << "1: tag #{tag} is retired; use #{replacement}"

test/test_lint_posts.rb

@@ -97,9 +97,14 @@ def test_rejects_an_unknown_topic_and_a_wrong_cover
     assert_finding(/cover for Paper Reviews/, lint(front: FRONT.sub("language-models.jpg", "default.jpg")))
   end
 
-  def test_tags_need_a_controlled_tag_and_current_well_formed_names
+  def test_a_post_needs_tags_but_not_a_controlled_one
+    assert_finding(/no tags/, lint(front: FRONT.sub(/^tags:.*$/, "tags: []")))
+    found = lint(front: FRONT.sub('"Efficient-Attention"', '"Multi-Token-Prediction"'))
+    refute found.any? { |f| f.include?("tag") }, found.inspect
+  end
+
+  def test_tags_need_current_well_formed_names
     found = lint(front: FRONT.sub('"Efficient-Attention"', '"Sparse-Mixture-of-Experts", "Public-Dataset-Only Training", "Community-Detection-For-RAG"'))
-    assert_finding(/no controlled topic tag/, found)
     assert_finding(/retired; use Mixture-of-Experts/, found)
     assert_finding(/joined by hyphens/, found)
     assert_finding(/should have for/, found)