diff --git a/code/libraries/language-processing/java/nu/marginalia/language/sentence/SentenceSegmentSplitter.java b/code/libraries/language-processing/java/nu/marginalia/language/sentence/SentenceSegmentSplitter.java
index 531f5189..62f7f772 100644
--- a/code/libraries/language-processing/java/nu/marginalia/language/sentence/SentenceSegmentSplitter.java
+++ b/code/libraries/language-processing/java/nu/marginalia/language/sentence/SentenceSegmentSplitter.java
@@ -27,7 +27,7 @@ public class SentenceSegmentSplitter {
         else {
             // If we flatten unicode, we do this...
             // FIXME: This can almost definitely be cleaned up and simplified.
-            wordBreakPattern = Pattern.compile("([^/_#@.a-zA-Z'+\\-0-9\\u00C0-\\u00D6\\u00D8-\\u00f6\\u00f8-\\u00ff]+)|[|]|(\\.(\\s+|$))");
+            wordBreakPattern = Pattern.compile("([^/<>$:_#@.a-zA-Z'+\\-0-9\\u00C0-\\u00D6\\u00D8-\\u00f6\\u00f8-\\u00ff]+)|[|]|(\\.(\\s+|$))");
         }
     }
 
diff --git a/code/libraries/language-processing/test/nu/marginalia/language/sentence/SentenceExtractorTest.java b/code/libraries/language-processing/test/nu/marginalia/language/sentence/SentenceExtractorTest.java
index 38ccbe12..1fcf1015 100644
--- a/code/libraries/language-processing/test/nu/marginalia/language/sentence/SentenceExtractorTest.java
+++ b/code/libraries/language-processing/test/nu/marginalia/language/sentence/SentenceExtractorTest.java
@@ -28,6 +28,20 @@ class SentenceExtractorTest {
         System.out.println(dld);
     }
 
+    @Test
+    void testCplusplus() {
+        var dld = sentenceExtractor.extractSentence("std::vector", EnumSet.noneOf(HtmlTag.class));
+        assertEquals(1, dld.length());
+        assertEquals("std::vector", dld.wordsLowerCase[0]);
+    }
+
+    @Test
+    void testPHP() {
+        var dld = sentenceExtractor.extractSentence("$_GET", EnumSet.noneOf(HtmlTag.class));
+        assertEquals(1, dld.length());
+        assertEquals("$_get", dld.wordsLowerCase[0]);
+    }
+
     @Test
     void testPolishArtist() {
         var dld = sentenceExtractor.extractSentence("Uklański", EnumSet.noneOf(HtmlTag.class));