diff --git a/htmlminimizer.py b/htmlminimizer.py
index 670b556..4c90793 100644
--- a/htmlminimizer.py
+++ b/htmlminimizer.py
@@ -1,18 +1,31 @@
 """Quick-and-dirty HTML minimizer"""
 
-import sys
+import sys, re, html.entities
 
 input_file = sys.argv[1]
 output_file = sys.argv[2]
 in_pre = False
-out = open(output_file, 'w')
+out = open(output_file, 'w', encoding="utf-8") # encoding argument! important!
 for line in open(input_file).readlines():
+    # replace entities with Unicode characters
+    for e in re.findall('&(.+?);', line):
+        n = html.entities.name2codepoint.get(e)
+        if not n:
+            if e.count('#x'):
+                # it's late, forgive me
+                n = eval(e.replace('#', '0'))
+            elif e.count('#'):
+                n = int(e.replace('#', ''))
+            else:
+                continue
+        line = line.replace('&' + e + ';', chr(n))
+
+    # strip leading and trailing whitespace, except inside <pre> blocks
     g = line.strip()
     if g.count('<pre'):
         in_pre = True
     if g.count('</pre'):
-        # this will break if you have a </pre> then <pre>
-        # on the same line, so don't do that
+        # XXX this will break if you have </pre><pre> in one line
         in_pre = False
         g = line.rstrip()
     if in_pre:
@@ -20,365 +33,3 @@ for line in open(input_file).readlines():
     else:
         out.write(g)
 out.close()
-
-out = open(output_file)
-html = out.read()
-out.close()
-html = html.replace("&aring;", "&#229;")
-html = html.replace("&#62;", "&gt;")
-html = html.replace("&#x3e;", "&gt;")
-html = html.replace("&#8835;", "&sup;")
-html = html.replace("&#x2283;", "&sup;")
-html = html.replace("&Ntilde;", "&#209;")
-html = html.replace("&#x3d2;", "&#978;")
-html = html.replace("&upsih;", "&#978;")
-html = html.replace("&Yacute;", "&#221;")
-html = html.replace("&Atilde;", "&#195;")
-html = html.replace("&#x221a;", "&#8730;")
-html = html.replace("&#x2297;", "&#8855;")
-html = html.replace("&otimes;", "&#8855;")
-html = html.replace("&aelig;", "&#230;")
-html = html.replace("&#936;", "&Psi;")
-html = html.replace("&#x3a8;", "&Psi;")
-html = html.replace("&#x395;", "&#917;")
-html = html.replace("&Epsilon;", "&#917;")
-html = html.replace("&Icirc;", "&#206;")
-html = html.replace("&Eacute;", "&#201;")
-html = html.replace("&#x39b;", "&#923;")
-html = html.replace("&Lambda;", "&#923;")
-html = html.replace("&#x2033;", "&#8243;")
-html = html.replace("&#x39a;", "&#922;")
-html = html.replace("&Kappa;", "&#922;")
-html = html.replace("&#x3c2;", "&#962;")
-html = html.replace("&sigmaf;", "&#962;")
-html = html.replace("&#8206;", "&lrm;")
-html = html.replace("&#x200e;", "&lrm;")
-html = html.replace("&cedil;", "&#184;")
-html = html.replace("&#8194;", "&ensp;")
-html = html.replace("&#x2002;", "&ensp;")
-html = html.replace("&AElig;", "&#198;")
-html = html.replace("&#x2032;", "&#8242;")
-html = html.replace("&#932;", "&Tau;")
-html = html.replace("&#x3a4;", "&Tau;")
-html = html.replace("&#x2308;", "&#8968;")
-html = html.replace("&#8659;", "&dArr;")
-html = html.replace("&#x21d3;", "&dArr;")
-html = html.replace("&#8805;", "&ge;")
-html = html.replace("&#x2265;", "&ge;")
-html = html.replace("&#8901;", "&sdot;")
-html = html.replace("&#x22c5;", "&sdot;")
-html = html.replace("&#x230a;", "&#8970;")
-html = html.replace("&lfloor;", "&#8970;")
-html = html.replace("&#8656;", "&lArr;")
-html = html.replace("&#x21d0;", "&lArr;")
-html = html.replace("&brvbar;", "&#166;")
-html = html.replace("&Otilde;", "&#213;")
-html = html.replace("&#x398;", "&#920;")
-html = html.replace("&Theta;", "&#920;")
-html = html.replace("&#928;", "&Pi;")
-html = html.replace("&#x3a0;", "&Pi;")
-html = html.replace("&#x152;", "&#338;")
-html = html.replace("&OElig;", "&#338;")
-html = html.replace("&#x160;", "&#352;")
-html = html.replace("&Scaron;", "&#352;")
-html = html.replace("&egrave;", "&#232;")
-html = html.replace("&#8834;", "&sub;")
-html = html.replace("&#x2282;", "&sub;")
-html = html.replace("&iexcl;", "&#161;")
-html = html.replace("&#8721;", "&sum;")
-html = html.replace("&#x2211;", "&sum;")
-html = html.replace("&ntilde;", "&#241;")
-html = html.replace("&atilde;", "&#227;")
-html = html.replace("&#x3b8;", "&#952;")
-html = html.replace("&theta;", "&#952;")
-html = html.replace("&#8836;", "&nsub;")
-html = html.replace("&#x2284;", "&nsub;")
-html = html.replace("&#8660;", "&hArr;")
-html = html.replace("&#x21d4;", "&hArr;")
-html = html.replace("&Oslash;", "&#216;")
-html = html.replace("&THORN;", "&#222;")
-html = html.replace("&#924;", "&Mu;")
-html = html.replace("&#x39c;", "&Mu;")
-html = html.replace("&#x2009;", "&#8201;")
-html = html.replace("&thinsp;", "&#8201;")
-html = html.replace("&ecirc;", "&#234;")
-html = html.replace("&#x201e;", "&#8222;")
-html = html.replace("&Aring;", "&#197;")
-html = html.replace("&#x2207;", "&#8711;")
-html = html.replace("&#x2030;", "&#8240;")
-html = html.replace("&permil;", "&#8240;")
-html = html.replace("&Ugrave;", "&#217;")
-html = html.replace("&#951;", "&eta;")
-html = html.replace("&#x3b7;", "&eta;")
-html = html.replace("&Agrave;", "&#192;")
-html = html.replace("&#x2200;", "&#8704;")
-html = html.replace("&forall;", "&#8704;")
-html = html.replace("&#240;", "&eth;")
-html = html.replace("&#xf0;", "&eth;")
-html = html.replace("&#x2309;", "&#8969;")
-html = html.replace("&Egrave;", "&#200;")
-html = html.replace("&divide;", "&#247;")
-html = html.replace("&igrave;", "&#236;")
-html = html.replace("&otilde;", "&#245;")
-html = html.replace("&pound;", "&#163;")
-html = html.replace("&#x2044;", "&#8260;")
-html = html.replace("&#208;", "&ETH;")
-html = html.replace("&#xd0;", "&ETH;")
-html = html.replace("&#x2217;", "&#8727;")
-html = html.replace("&lowast;", "&#8727;")
-html = html.replace("&#967;", "&chi;")
-html = html.replace("&#x3c7;", "&chi;")
-html = html.replace("&Aacute;", "&#193;")
-html = html.replace("&#x392;", "&#914;")
-html = html.replace("&#8869;", "&perp;")
-html = html.replace("&#x22a5;", "&perp;")
-html = html.replace("&#x2234;", "&#8756;")
-html = html.replace("&there4;", "&#8756;")
-html = html.replace("&#960;", "&pi;")
-html = html.replace("&#x3c0;", "&pi;")
-html = html.replace("&#x2205;", "&#8709;")
-html = html.replace("&#x2209;", "&#8713;")
-html = html.replace("&icirc;", "&#238;")
-html = html.replace("&#8226;", "&bull;")
-html = html.replace("&#x2022;", "&bull;")
-html = html.replace("&#x3c5;", "&#965;")
-html = html.replace("&upsilon;", "&#965;")
-html = html.replace("&Oacute;", "&#211;")
-html = html.replace("&#x3ba;", "&#954;")
-html = html.replace("&kappa;", "&#954;")
-html = html.replace("&ccedil;", "&#231;")
-html = html.replace("&#8745;", "&cap;")
-html = html.replace("&#x2229;", "&cap;")
-html = html.replace("&#956;", "&mu;")
-html = html.replace("&#x3bc;", "&mu;")
-html = html.replace("&#176;", "&deg;")
-html = html.replace("&#xb0;", "&deg;")
-html = html.replace("&#964;", "&tau;")
-html = html.replace("&#x3c4;", "&tau;")
-html = html.replace("&#8195;", "&emsp;")
-html = html.replace("&#x2003;", "&emsp;")
-html = html.replace("&#x2026;", "&#8230;")
-html = html.replace("&hellip;", "&#8230;")
-html = html.replace("&ucirc;", "&#251;")
-html = html.replace("&ugrave;", "&#249;")
-html = html.replace("&#8773;", "&cong;")
-html = html.replace("&#x2245;", "&cong;")
-html = html.replace("&#x399;", "&#921;")
-html = html.replace("&#x22;", "&#34;")
-html = html.replace("&quot;", "&#34;")
-html = html.replace("&#8594;", "&rarr;")
-html = html.replace("&#x2192;", "&rarr;")
-html = html.replace("&#929;", "&Rho;")
-html = html.replace("&#x3a1;", "&Rho;")
-html = html.replace("&uacute;", "&#250;")
-html = html.replace("&acirc;", "&#226;")
-html = html.replace("&#8764;", "&sim;")
-html = html.replace("&#x223c;", "&sim;")
-html = html.replace("&#966;", "&phi;")
-html = html.replace("&#x3c6;", "&phi;")
-html = html.replace("&#x2666;", "&#9830;")
-html = html.replace("&Ccedil;", "&#199;")
-html = html.replace("&#919;", "&Eta;")
-html = html.replace("&#x397;", "&Eta;")
-html = html.replace("&#x393;", "&#915;")
-html = html.replace("&Gamma;", "&#915;")
-html = html.replace("&#8364;", "&euro;")
-html = html.replace("&#x20ac;", "&euro;")
-html = html.replace("&#x3d1;", "&#977;")
-html = html.replace("&thetasym;", "&#977;")
-html = html.replace("&#x201c;", "&#8220;")
-html = html.replace("&#x2665;", "&#9829;")
-html = html.replace("&hearts;", "&#9829;")
-html = html.replace("&oacute;", "&#243;")
-html = html.replace("&#8204;", "&zwnj;")
-html = html.replace("&#x200c;", "&zwnj;")
-html = html.replace("&#165;", "&yen;")
-html = html.replace("&#xa5;", "&yen;")
-html = html.replace("&ograve;", "&#242;")
-html = html.replace("&#935;", "&Chi;")
-html = html.replace("&#x3a7;", "&Chi;")
-html = html.replace("&#x2122;", "&#8482;")
-html = html.replace("&#958;", "&xi;")
-html = html.replace("&#x3be;", "&xi;")
-html = html.replace("&#x2dc;", "&#732;")
-html = html.replace("&tilde;", "&#732;")
-html = html.replace("&#x2039;", "&#8249;")
-html = html.replace("&lsaquo;", "&#8249;")
-html = html.replace("&#x153;", "&#339;")
-html = html.replace("&oelig;", "&#339;")
-html = html.replace("&#x2261;", "&#8801;")
-html = html.replace("&#8804;", "&le;")
-html = html.replace("&#x2264;", "&le;")
-html = html.replace("&#8746;", "&cup;")
-html = html.replace("&#x222a;", "&cup;")
-html = html.replace("&#x178;", "&#376;")
-html = html.replace("&#60;", "&lt;")
-html = html.replace("&#x3c;", "&lt;")
-html = html.replace("&#x3a5;", "&#933;")
-html = html.replace("&Upsilon;", "&#933;")
-html = html.replace("&#x2013;", "&#8211;")
-html = html.replace("&yacute;", "&#253;")
-html = html.replace("&#8476;", "&real;")
-html = html.replace("&#x211c;", "&real;")
-html = html.replace("&#968;", "&psi;")
-html = html.replace("&#x3c8;", "&psi;")
-html = html.replace("&#x203a;", "&#8250;")
-html = html.replace("&rsaquo;", "&#8250;")
-html = html.replace("&#8595;", "&darr;")
-html = html.replace("&#x2193;", "&darr;")
-html = html.replace("&#x391;", "&#913;")
-html = html.replace("&Alpha;", "&#913;")
-html = html.replace("&#172;", "&not;")
-html = html.replace("&#xac;", "&not;")
-html = html.replace("&#x26;", "&#38;")
-html = html.replace("&oslash;", "&#248;")
-html = html.replace("&acute;", "&#180;")
-html = html.replace("&#8205;", "&zwj;")
-html = html.replace("&#x200d;", "&zwj;")
-html = html.replace("&laquo;", "&#171;")
-html = html.replace("&#x201d;", "&#8221;")
-html = html.replace("&Igrave;", "&#204;")
-html = html.replace("&micro;", "&#181;")
-html = html.replace("&#173;", "&shy;")
-html = html.replace("&#xad;", "&shy;")
-html = html.replace("&#8839;", "&supe;")
-html = html.replace("&#x2287;", "&supe;")
-html = html.replace("&szlig;", "&#223;")
-html = html.replace("&#x2663;", "&#9827;")
-html = html.replace("&agrave;", "&#224;")
-html = html.replace("&Ocirc;", "&#212;")
-html = html.replace("&#8596;", "&harr;")
-html = html.replace("&#x2194;", "&harr;")
-html = html.replace("&#8592;", "&larr;")
-html = html.replace("&#x2190;", "&larr;")
-html = html.replace("&frac12;", "&#189;")
-html = html.replace("&#8733;", "&prop;")
-html = html.replace("&#x221d;", "&prop;")
-html = html.replace("&#x2c6;", "&#710;")
-html = html.replace("&ocirc;", "&#244;")
-html = html.replace("&#x2248;", "&#8776;")
-html = html.replace("&#168;", "&uml;")
-html = html.replace("&#xa8;", "&uml;")
-html = html.replace("&#8719;", "&prod;")
-html = html.replace("&#x220f;", "&prod;")
-html = html.replace("&#174;", "&reg;")
-html = html.replace("&#xae;", "&reg;")
-html = html.replace("&#8207;", "&rlm;")
-html = html.replace("&#x200f;", "&rlm;")
-html = html.replace("&#x221e;", "&#8734;")
-html = html.replace("&#x3a3;", "&#931;")
-html = html.replace("&Sigma;", "&#931;")
-html = html.replace("&#x2014;", "&#8212;")
-html = html.replace("&#8593;", "&uarr;")
-html = html.replace("&#x2191;", "&uarr;")
-html = html.replace("&times;", "&#215;")
-html = html.replace("&#8658;", "&rArr;")
-html = html.replace("&#x21d2;", "&rArr;")
-html = html.replace("&#8744;", "&or;")
-html = html.replace("&#x2228;", "&or;")
-html = html.replace("&#x3b3;", "&#947;")
-html = html.replace("&gamma;", "&#947;")
-html = html.replace("&#x3bb;", "&#955;")
-html = html.replace("&lambda;", "&#955;")
-html = html.replace("&#9002;", "&rang;")
-html = html.replace("&#x232a;", "&rang;")
-html = html.replace("&#x2020;", "&#8224;")
-html = html.replace("&dagger;", "&#8224;")
-html = html.replace("&#x2111;", "&#8465;")
-html = html.replace("&#x2135;", "&#8501;")
-html = html.replace("&alefsym;", "&#8501;")
-html = html.replace("&#8838;", "&sube;")
-html = html.replace("&#x2286;", "&sube;")
-html = html.replace("&#x3b1;", "&#945;")
-html = html.replace("&alpha;", "&#945;")
-html = html.replace("&#925;", "&Nu;")
-html = html.replace("&#x39d;", "&Nu;")
-html = html.replace("&plusmn;", "&#177;")
-html = html.replace("&frac34;", "&#190;")
-html = html.replace("&#x203e;", "&#8254;")
-html = html.replace("&#x394;", "&#916;")
-html = html.replace("&Delta;", "&#916;")
-html = html.replace("&#9674;", "&loz;")
-html = html.replace("&#x25ca;", "&loz;")
-html = html.replace("&#x3b9;", "&#953;")
-html = html.replace("&iacute;", "&#237;")
-html = html.replace("&#x3b5;", "&#949;")
-html = html.replace("&epsilon;", "&#949;")
-html = html.replace("&#x2118;", "&#8472;")
-html = html.replace("&weierp;", "&#8472;")
-html = html.replace("&#8706;", "&part;")
-html = html.replace("&#x2202;", "&part;")
-html = html.replace("&#x3b4;", "&#948;")
-html = html.replace("&delta;", "&#948;")
-html = html.replace("&#x3bf;", "&#959;")
-html = html.replace("&omicron;", "&#959;")
-html = html.replace("&#926;", "&Xi;")
-html = html.replace("&#x39e;", "&Xi;")
-html = html.replace("&#x2021;", "&#8225;")
-html = html.replace("&Dagger;", "&#8225;")
-html = html.replace("&Ograve;", "&#210;")
-html = html.replace("&Ucirc;", "&#219;")
-html = html.replace("&#x161;", "&#353;")
-html = html.replace("&scaron;", "&#353;")
-html = html.replace("&#x2018;", "&#8216;")
-html = html.replace("&#8712;", "&isin;")
-html = html.replace("&#x2208;", "&isin;")
-html = html.replace("&#x396;", "&#918;")
-html = html.replace("&#x2212;", "&#8722;")
-html = html.replace("&#8743;", "&and;")
-html = html.replace("&#x2227;", "&and;")
-html = html.replace("&#8736;", "&ang;")
-html = html.replace("&#x2220;", "&ang;")
-html = html.replace("&curren;", "&#164;")
-html = html.replace("&#8747;", "&int;")
-html = html.replace("&#x222b;", "&int;")
-html = html.replace("&#x230b;", "&#8971;")
-html = html.replace("&rfloor;", "&#8971;")
-html = html.replace("&#x21b5;", "&#8629;")
-html = html.replace("&#x2203;", "&#8707;")
-html = html.replace("&#x2295;", "&#8853;")
-html = html.replace("&Acirc;", "&#194;")
-html = html.replace("&#982;", "&piv;")
-html = html.replace("&#x3d6;", "&piv;")
-html = html.replace("&#8715;", "&ni;")
-html = html.replace("&#x220b;", "&ni;")
-html = html.replace("&#934;", "&Phi;")
-html = html.replace("&#x3a6;", "&Phi;")
-html = html.replace("&Iacute;", "&#205;")
-html = html.replace("&Uacute;", "&#218;")
-html = html.replace("&#x39f;", "&#927;")
-html = html.replace("&Omicron;", "&#927;")
-html = html.replace("&#8800;", "&ne;")
-html = html.replace("&#x2260;", "&ne;")
-html = html.replace("&iquest;", "&#191;")
-html = html.replace("&#x201a;", "&#8218;")
-html = html.replace("&Ecirc;", "&#202;")
-html = html.replace("&#x3b6;", "&#950;")
-html = html.replace("&#x3a9;", "&#937;")
-html = html.replace("&Omega;", "&#937;")
-html = html.replace("&#957;", "&nu;")
-html = html.replace("&#x3bd;", "&nu;")
-html = html.replace("&frac14;", "&#188;")
-html = html.replace("&aacute;", "&#225;")
-html = html.replace("&#8657;", "&uArr;")
-html = html.replace("&#x21d1;", "&uArr;")
-html = html.replace("&#x3b2;", "&#946;")
-html = html.replace("&#x192;", "&#402;")
-html = html.replace("&#961;", "&rho;")
-html = html.replace("&#x3c1;", "&rho;")
-html = html.replace("&eacute;", "&#233;")
-html = html.replace("&#x3c9;", "&#969;")
-html = html.replace("&omega;", "&#969;")
-html = html.replace("&middot;", "&#183;")
-html = html.replace("&#9001;", "&lang;")
-html = html.replace("&#x2329;", "&lang;")
-html = html.replace("&#x2660;", "&#9824;")
-html = html.replace("&spades;", "&#9824;")
-html = html.replace("&#x2019;", "&#8217;")
-html = html.replace("&thorn;", "&#254;")
-html = html.replace("&raquo;", "&#187;")
-html = html.replace("&#x3c3;", "&#963;")
-html = html.replace("&sigma;", "&#963;")
-out = open(output_file, 'w')
-out.write(html)
-out.close()
diff --git a/publish b/publish
index bc41444..29e2904 100755
--- a/publish
+++ b/publish
@@ -7,7 +7,7 @@ cp robots.txt *.js *.css build/
 rm -f examples/*.pyc
 cp -R examples build/
 
-# minimize HTML (note: this script is quite fragile and relies on knowledge of how I write HTML)
+# minimize HTML (XXX this script is quite fragile and relies on knowledge of how I write HTML)
 for f in *.html; do
   python htmlminimizer.py "$f" build/"$f"
 done
@@ -41,7 +41,7 @@ sed -i -e "s|<link rel=stylesheet type=text/css media='only screen and (max-devi
 sed -i -e "s|dip3\.js|http://wearehugh.com/dip3/${revision}.js|g" build/*.html
 sed -i -e "s|html5\.js|http://wearehugh.com/dip3/html5.js|g" build/*.html
 
-# images will be served from a separate domain
+# images would be served from a separate domain if we had any, which we currently don't
 #sed -i -e "s|bsb.png|http://wearehugh.com/dip3/bsb.png|g" build/*.html
 
 # minimize URLs
diff --git a/xml.html b/xml.html
index c0c9e2c..46690ff 100644
--- a/xml.html
+++ b/xml.html
@@ -244,20 +244,29 @@ mark{display:inline}
 
 <h2 id=xml-parse>Parsing XML</h2>
 
-<p>Python comes with an efficient XML parsing library called Etree.
+<p>Python can parse XML documents in several ways. It has traditional <a href=http://en.wikipedia.org/wiki/XML#DOM>DOM</a> and <a href=http://en.wikipedia.org/wiki/Simple_API_for_XML>SAX</a> parsers, but I will focus on a different library called Etree.
 
 <p class=d>[<a href=examples/feed.xml>download <code>feed.xml</code></a>]
 <pre class=screen>
->>> import xml.etree.ElementTree as etree
->>> tree = etree.parse("examples/feed.xml")
->>> root = tree.getroot()
->>> root
-&lt;Element {http://www.w3.org/2005/Atom}feed at cd1eb0>
-</pre>
+<a><samp class=p>>>> </samp><kbd>import xml.etree.ElementTree as etree</kbd>    <span>&#x2460;</span></a>
+<a><samp class=p>>>> </samp><kbd>tree = etree.parse("examples/feed.xml")</kbd>  <span>&#x2461;</span></a>
+<a><samp class=p>>>> </samp><kbd>root = tree.getroot()</kbd>                    <span>&#x2462;</span></a>
+<a><samp class=p>>>> </samp><kbd>root</kbd>                                     <span>&#x2463;</span></a>
+<samp>&lt;Element {http://www.w3.org/2005/Atom}feed at cd1eb0></samp></pre>
+<ol>
+<li>The Etree library is part of the Python standard library, in <code>xml.etree.ElementTree</code>.
+<li>The primary entry point for the Etree library is the <code>parse()</code> function, which can take a filename or a file-like object [FIXME xref]. This function parses the entire document at once. If memory is tight, there are ways to parse an XML document incrementally instead.
+<li>The <code>parse()</code> function returns an object which represents the entire document. This is <em>not</em> the root element. To get a reference to the root element, call the <code>getroot()</code> method.
+<li>As expected, the root element is the <code>feed</code> element in the <code>http://www.w3.org/2005/Atom</code> namespace. The string representation of this object reinforces an important point: an XML element is a combination of its namespace and its tag name (also called the <i>local name</i>). Every element in this document is in the Atom namespace, so the root element is represented as <code>{http://www.w3.org/2005/Atom}feed</code>.
+</ol>
+
+<blockquote class=note>
+<p><span>&#x261E;</span>Etree represents XML elements as <code>{<var>namespace</var>}<var>localname</var></code>. You&#8217;ll see and use this format in multiple places in the Etree library.
+</blockquote>
 
 <h3 id=xml-elements>Elements Are Lists</h3>
 
-<p>FIXME
+<p>In Etree, an element acts like a list. The items of the list are the element&#8217;s children.
 
 <pre class=screen>
 >>> root.tag