--- a/spec/latest/json-ld-syntax/index.html	Mon Nov 07 11:50:52 2011 +0800
+++ b/spec/latest/json-ld-syntax/index.html	Mon Nov 07 13:10:51 2011 +0800
@@ -878,7 +878,7 @@
 -->
 </pre>
 
-<p><tref>Prefix</tref>es are expanded when if the form of the value is <code>prefix:suffix</code>, and the
+<p><tref>Prefix</tref>es are expanded when the form of the value is <code>prefix:suffix</code>, and the
   prefix matches a <tref>term</tref> defined within the <tref>active context</tref>:</p>
 
 <pre class="example" data-transform="updateExample">
@@ -893,7 +893,7 @@
 </pre>
 
 <p><code>foaf:name</code> above will automatically expand out to the IRI
-<code>http://xmlns.com/foaf/0.1/name</code>.</p>
+<code>http://xmlns.com/foaf/0.1/name</code>. See <a href="#prefixes">Prefixes</a> for more details.</p>
 
 <p>An <tref>IRI</tref> is generated when a value is associated with a key using
 the <code>@iri</code> keyword:</p>
@@ -1407,7 +1407,7 @@
 -->
 </pre>
 <section>
-  <h2>Vocabulary Prefixes</h2>
+  <h2>Prefixes</h2>
   <p>
     Vocabulary terms in <tref>Linked Data</tref> documents may draw from a number of
     different Web vocabularies. At times, declaring every single term that
@@ -1423,8 +1423,8 @@
     <tref>IRI</tref> to a <tref>Web Vocabulary</tref>.
     Generally, these prefixes are used by concatenating the <em>prefix</em> and
     a <em>suffix</em> separated by a colon (<code>:</code>).
-    The <tref>prefix</tref> is a <tref>term</tref> taken from the <tref>active context</tref>,
-    a short string that identifies a particular Web vocabulary.
+    The <tref>prefix</tref> is a <tref>term</tref> taken from the <tref>active context</tref>
+    and is a short string identifying a particular <tref>IRI</tref> in a JSON-LD document.
     For example, the prefix <code>foaf</code> may be used as a short
     hand for the Friend-of-a-Friend Web Vocabulary, which is identified using
     the IRI <code>http://xmlns.com/foaf/0.1/</code>. A developer may append any of
@@ -1435,6 +1435,17 @@
     and type out the entire IRI, the developer can instead use the prefix in
     their JSON-LD markup.
   </p>
+  <p>To generate an <tref>IRI</tref> out of a <code>prefix:suffix</code> construct,
+    the value is first split into a <em>prefix</em> and <em>suffix</em> at the first
+    occurrence of a colon (<code>:</code>). If the <tref>active context</tref> contains a term
+    mapping for <em>prefix</em>, an IRI is generated by prepending the mapped
+    <em>prefix</em> to the (possibly empty) <em>suffix</em> using textual concatenation.
+    If no prefix mapping is defined, the value is used directly as an IRI. If the
+    prefix is an underscore (<code>_</code>), the IRI remains unchanged. If the
+    IRI being processed has an empty prefix, the <em>suffix</em> is processed like a
+    relative IRI (see <a href="#base-iri">Base IRI</a> and
+    <a href="#default-vocabulary">Default Vocabulary</a> for details).
+  </p>
   <p>
     The ability to use <tref>prefix</tref>es reduces the need for developers
     to declare every vocabulary term that they intend to use in