Hack the Planet in Exile : I took another look at Cloud Pipe since it was...

I took another look at Cloud Pipe since it was updated over the holidays. It looks like a fine way to solve the NAT problem for some realtime clients.

I have an alternate suggestion about namespace handling. A Cloud Pipe ping looks like this on the wire:

<?xml version="1.0"?>
<packets>
	<fatPing feed="http://www.nytimes.com/services/xml/rss/userland/HomePage.xml">
		<![CDATA[&lt;title&gt;Florida&apos;s Meyer Will Take Leave, Not Resign&lt;/title&gt;
&lt;link&gt;http://www.nytimes.com/2009/12/28/sports/ncaafootball/28florida.html?partner=rss&amp;amp;emc=rss&lt;/link&gt;
&lt;guid isPermaLink=&quot;true&quot;&gt;http://www.nytimes.com/2009/12/28/sports/ncaafootball/28florida.html&lt;/guid&gt;
&lt;description&gt;Florida Coach Urban Meyer told his team he would not step down but would take an indefinite leave of absence.&lt;/description&gt;
&lt;dc:creator&gt;By PETE THAMEL&lt;/dc:creator&gt;
&lt;pubDate&gt;Sun, 27 Dec 2009 19:42:18 GMT&lt;/pubDate&gt;
&lt;category domain=&quot;http://www.nytimes.com/namespaces/keywords/des&quot;&gt;College Athletics&lt;/category&gt;
&lt;category domain=&quot;http://www.nytimes.com/namespaces/keywords/nyt_org_all&quot;&gt;University of Florida&lt;/category&gt;
&lt;category domain=&quot;http://www.nytimes.com/namespaces/keywords/nyt_per&quot;&gt;Meyer, Urban&lt;/category&gt;
&lt;category domain=&quot;http://www.nytimes.com/namespaces/keywords/nyt_per&quot;&gt;Thamel, Pete&lt;/category&gt;
]]>
		</fatPing>
	<system>
		<serialnum>28440</serialnum>
		<when>Sun, 27 Dec 2009 20:23:13 GMT</when>
		<secs>58.43333333</secs>
		</system>
	</packets>

The client then decodes the CDATA section to recover the original RSS <item>:

<title>Florida's Meyer Will Take Leave, Not Resign</title>
<link>http://www.nytimes.com/2009/12/28/sports/ncaafootball/28florida.html?partner=rss&amp;emc=rss</link>
<guid isPermaLink="true">http://www.nytimes.com/2009/12/28/sports/ncaafootball/28florida.html</guid>
<description>Florida Coach Urban Meyer told his team he would not step down but would take an indefinite leave of absence.</description>
<dc:creator>By PETE THAMEL</dc:creator>
<pubDate>Sun, 27 Dec 2009 19:42:18 GMT</pubDate>
<category domain="http://www.nytimes.com/namespaces/keywords/des">College Athletics</category>
<category domain="http://www.nytimes.com/namespaces/keywords/nyt_org_all">University of Florida</category>
<category domain="http://www.nytimes.com/namespaces/keywords/nyt_per">Meyer, Urban</category>
<category domain="http://www.nytimes.com/namespaces/keywords/nyt_per">Thamel, Pete</category>

Unfortunately the dc namespace hasn’t been declared anywhere so we don’t know what it means. Looking at the original feed, we see namespace declarations on the root element:

<?xml version="1.0"?>
<rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:media="http://search.yahoo.com/mrss/" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:nyt="http://www.nytimes.com/namespaces/rss/2.0" version="2.0">
  <channel>
    <title>NYT &gt; Home Page</title>
    ...

When extracting a single <item> from the feed, we can preserve the namespace declarations by “pushing them down” onto each <item> like so: (In this case I have taken a somewhat brute force approach of copying all the namespace declarations, even though most of them aren’t needed.)

<?xml version="1.0"?>
<rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:media="http://search.yahoo.com/mrss/" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:nyt="http://www.nytimes.com/namespaces/rss/2.0" version="2.0">
  <channel>
    <title>NYT &gt; Home Page</title>
    ...
    <item xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:media="http://search.yahoo.com/mrss/" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:nyt="http://www.nytimes.com/namespaces/rss/2.0">
      <title>Tensions Mount in Devastated Capital as Nations Step Up Aid Pledges to Haiti</title>
      <link>http://www.nytimes.com/2010/01/15/world/americas/15haiti.html?partner=rss&amp;emc=rss</link>
      <guid isPermaLink="true">http://www.nytimes.com/2010/01/15/world/americas/15haiti.html</guid>
      <description>Signs of tension and urgency were growing in the devastated capital as the police and government had all but vanished.</description>
      <dc:creator>By MARC LACEY</dc:creator>
      <pubDate>Fri, 15 Jan 2010 05:28:37 GMT</pubDate>
      ...
    </item>
...

Now the item stands alone and can be embedded into a Cloud Pipe without any encoding:

<?xml version="1.0"?>
<packets>
	<fatPing feed="http://www.nytimes.com/services/xml/rss/userland/HomePage.xml">
		<item xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:media="http://search.yahoo.com/mrss/" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:nyt="http://www.nytimes.com/namespaces/rss/2.0">
		  <title>Tensions Mount in Devastated Capital as Nations Step Up Aid Pledges to Haiti</title>
		  <link>http://www.nytimes.com/2010/01/15/world/americas/15haiti.html?partner=rss&amp;emc=rss</link>
		  <guid isPermaLink="true">http://www.nytimes.com/2010/01/15/world/americas/15haiti.html</guid>
		  <description>Signs of tension and urgency were growing in the devastated capital as the police and government had all but vanished.</description>
		  <dc:creator>By MARC LACEY</dc:creator>
		  <pubDate>Fri, 15 Jan 2010 05:28:37 GMT</pubDate>
		  ...
		</item>
		</fatPing>
	<system>
		<serialnum>28440</serialnum>
		<when>Sun, 27 Dec 2009 20:23:13 GMT</when>
		<secs>58.43333333</secs>
		</system>
	</packets>

XML experts, please let me know if I’ve gotten anything wrong here.

Hack the Planet in Exile

2010-01-15