--- a/bestpractices/BestPractices.html	Tue Jan 31 22:07:01 2012 +0000
+++ b/bestpractices/BestPractices.html	Thu Feb 02 12:38:16 2012 +0000
@@ -151,8 +151,77 @@
 </form> 
 </div> 
 
+    <section id="serialization">
+       <h2>Serialization of provenance accounts</h2>
+       <p>
+        A representation of a [[PROV-DM]] Account SHOULD be in RDF
+        [[RDF-CONCEPTS]] using the [[PROV-O]] ontology in OWL2
+        [[OWL2-PRIMER]], serialised as Turtle [[TURTLE]] 
+        or any W3C-standardized RDF serialization such as RDF/XML
+        [[RDF-SYNTAX]].
+        </p>
+        <p>
+        <h3>Accounts</h3>
+        <p>
+        Each PROV Account SHOULD be a separate RDF resource, unless the
+        representation allows multiple, named RDF graphs. 
+        </p>
+        <div class="note">TODO: How to express the asserter? Within or
+        outside the account graph?</div>
+        <h3>Media type</h3>
+        <p>
+        No dedicated media type has been allocated for PROV accounts.
+        Resources should use the media types assigned for the 
+        particular RDF representation used, for instance 
+        <code>application/rdf+xml</code> for RDF/XML and
+        <code>text/turtle</code> for Turtle.
+        </p>
+        <div class="note">TODO: Links for media type</div>
+        <p>
+        A resource can be identified as a <code>prov:Account</code> if
+        it assert itself as such. In most cases this can be done using a
+        relative URI reference to the requested resource.  For instance,
+        dereferencing a URI over HTTP with content negotiation:</p>
+        <pre class="example">
+            GET http://www.example.com/prov-account1 HTTP/1.1
+            Accept: text/turtle, application/rdf+xml
 
-	<section d="specializing-provenance-ontology-for-domain-specific-provenance-applications">
+            200 OK
+            Content-Type: text/turtle
+
+            @prefix prov: &lt;http://www.w3.org/ns/prov-o/&gt; .
+            &lt;&gt; a prov:Account .
+            # ...
+        </pre>
+        <p>
+        In the above example,
+        <code>&lt;/home/stain/src/provenance-wg/prov&gt;</code> is asserted to
+        be a PROV Account. Applications should take care to process HTTP
+        redirections, HTTP headers like <code>Content-Entity</code> to
+        relate the provenance resource with any modifications to
+        relative URIs in RDF representations, such as using
+        <code>xml:base</code> and <code>@base</code>.
+        </p>
+
+
+        <div class="note">
+        TODO: Add <code>prov:Account</code> to PROV-O for this purpose.
+        </div>
+       </p>
+    </section>
+
+    <section id="existing-vocabularies">
+        <h2>Relation to existing vocabularies (DC, OpenID, Creative Commons, FOAF)</h2>
+        <div class="note">
+        To be written by Kai, Daniel, Simon, Graham.
+        </div>
+        <section id="dublincore">
+            <h3>Dublin core</h3>
+        </section>
+        
+    </section>
+
+	<section id="specializing-provenance-ontology-for-domain-specific-provenance-applications">
       <h2>Specializing the PROV Ontology for Domain-specific Provenance Applications</h2>
 	    <p>The PROV Ontology [[PROV-O]] is the normative representation of the PROV Data Model (PROV-DM) [[PROV-DM]] using the Web Ontology Language (OWL2) [[!OWL2-PRIMER]]. The PROV Ontology is conceived as a reference ontology that can be extended by various domain-specific applications to model the required set of provenance terms. The PROV Ontology classes and properties can be specialized using the following two RDFS properties:</p>
    		<ul>