Add missing steps to Hash N-Degree Quads.

spec/latest/rdf-dataset-normalization/index.html

@@ -637,7 +637,7 @@ <h3>Algorithm</h3>
 
       <p>The inputs to this algorithm are the <tref>normalization state</tref>,
         the <i>identifier</i> for the <tref>RDF blank node</tref> to
-        recursively hash quads for, and <i>path identifier issuer</i> which is
+        recursively hash quads for, and path identifier <i>issuer</i> which is
         an <tref>identifier issuer</tref> that issues temporary
         <tref>RDF blank node identifier</tref>s. The output from this algorithm
         will be a <tref>hash</tref> and the <tref>identifier issuer</tref> used
@@ -682,10 +682,12 @@ <h3>Algorithm</h3>
             </li>
           </ol>
         </li>
+        <li>Create an empty string, <i>data to hash</i>.</li>
         <li>For each <i>related hash</i> to <i>blank node list</i> mapping in
           <i>hash to related blank nodes map</i>, sorted lexicographically
           by <i>related hash</i>:
           <ol class="algorithm">
+            <li>Append the <i>related hash</i> to the <i>data to hash</i>.</li>
             <li>Create a <tref>string</tref> <i>chosen path</i>.</li>
             <li>Create an unset <i>chosen issuer</i> variable.</li>
             <li>For each <i>permutation</i> of <i>blank node list</i>:
@@ -733,8 +735,7 @@ <h3>Algorithm</h3>
                       of <i>path</i> is greater than or equal to the length
                       of <i>chosen path</i> and <i>path</i> is lexicographically
                       greater than <i>chosen path</i>, then skip to the next
-                      <i>permutation</i>.
-                    </li>
+                      <i>permutation</i>.</li>
                   </ol>
                 </li>
                 <li>If <i>chosen path</i> is empty or <i>path</i> is
@@ -743,13 +744,14 @@ <h3>Algorithm</h3>
                   to <i>issuer copy</i>.</li>
               </ol>
             </li>
-            <li>Return the <i>chosen issuer</i> and the <tref>hash</tref> that
-              results from passing the lexicographically-sorted
-              <i>related hash</i>es in <i>hash to related blank nodes map</i>
-              and the joined <i>chosen path</i> through the
-              <tref>hash algorithm</tref>.</li>
+            <li>Append <i>chosen path</i> to <i>data to hash</i>.</li>
+            <li>Replace <i>issuer</i>, by reference, with
+              <i>chosen issuer</i>.</li>
           </ol>
         </li>
+        <li>Return <i>issuer</i> and the <tref>hash</tref> that results from
+          passing <i>data to hash</i> through the
+          <tref>hash algorithm</tref>.</li>
       </ol>
     </section>
   </section>