docs: update manual description of matches pane

Signed-off-by: Hiroshi Miura <[email protected]>

doc_src/en/Dialogs_ProjectProperties.xml

@@ -127,7 +127,11 @@
 		  mid-translation may force OmegaT to upgrade old translation memories
 		  that did not use sentence segmentation, but not vice versa. However,
 		  OmegaT will attempt to create fuzzy matches for paragraphs by
-		  combining existing sentence translations.</para>
+		  combining existing sentence translations. See <link
+		  linkend="dialog.preferences.tm.matches.paragraph.from.segmented.tmx"
+		  endterm="dialog.preferences.tm.matches.paragraph.from.segmented.tmx.title"/> to enable/disable
+		  the feature.
+		  </para>
 
 		  <para>If you change the segmentation while translating, you will have
 		  to reload the project for the new segmentation to take effect. This

doc_src/en/HowTo_UseTM.xml

@@ -165,6 +165,24 @@
 	  one of its subfolders and the translated data will be immediately be
 	  available for matching purposes.</para>
 	  </listitem>
+
+          <listitem>
+            <para>When you ever run projects which use <link
+			    linkend="dialogs.project.properties.options.segmentation"
+			    endterm="dialogs.project.properties.options.segmentation.title"/>,
+		    and the current project is switched to the paragraph segmentation, you can enable a preference
+		    option <link
+	    linkend="dialog.preferences.tm.matches.paragraph.from.segmented.tmx"
+	    endterm="dialog.preferences.tm.matches.paragraph.from.segmented.tmx.title"/> that allow OmegaT matches
+		    in special way. In addition to normal fuzzy matches, OmegaT try to segment source text into
+		    sentences, and conduct fuzzy match on the segmented TMX entries, then construct paragraph from
+		    the result. A suggested paragraph can be one that is not in the translation memories. It is why
+		    an origin of TM is shown as empty.
+		    Please refer <link linkend="dialog.preferences.segmentation.setup.type"
+				    endterm="dialog.preferences.segmentation.setup.type.title"/> to know pros and
+		    cons of segmentation configuration.
+            </para>
+          </listitem>
 	</itemizedlist>
 
     <section id="how.to.tm.read.and.write">

doc_src/en/OmegaT_Preferences.xml

@@ -1482,6 +1482,15 @@ ${filePath}></programlisting></para>
 		  </table>
 		</listitem>
 	  </varlistentry>
+          <varlistentry id="dialog.preferences.tm.matches.paragraph.from.segmented.tmx">
+		  <term id="dialog.preferences.tm.matches.paragraph.from.segmented.tmx.title">
+			  <option>Paragraph from segment matches of segmented TMX on a non-segment project</option></term>
+            <listitem><para>Enable matches for paragraph, which is constructed from segment matches search on segmented
+		    TMX entries searched with sentences from a segmented source text, even when running on a
+		    non-segment(paragraph segmentation) project.</para>
+            </listitem>
+
+          </varlistentry>
 	</variablelist>
   </section>