PHP documentation generation is done in two steps.  First, you need to generate the docbook file.  You somehow have to find this page http://doc.php.net/php/dochowto/chapter-configure.php that explains how.  Once the docbook file, normally called manual.xml, is generated you have to actually generate the documentation output.  I found the PhD manual here: http://doc.php.net/phd/docs.  PhD can create a number of different documentation formats from the docbook file.  I normally just use the xhtml output and serve the page up on my local web-server to verify my changes.  I found section on rendering to be a little hard to follow.  Here are the steps I used to update the content on the Memcache:getExtendedStats page:

sudo pear install doc.php.net/phd-beta
 
# use a sparse SVN checkout so all the languages and
# the tags and branches directories are skipped
svn co http://svn.php.net/repository/phpdoc --depth empty phpdoc
 
# checkout the documentation toolset
svn update phpdoc/doc-base --set-depth empty
svn update phpdoc/doc-base/trunk --set-depth infinity
 
# checkout the english version of the documentation
svn update phpdoc/en --set-depth empty
svn update phpdoc/en/trunk --set-depth infinity
 
cd phpdoc
vim en/trunk/reference/memcache/memcache/getextendedstats.xml
 
# generate the docbook file
# disable libxml check as Cent OS 5.5 does not have LibXML version 2.7.4
# put the docbook file in the root phpdoc directory
php doc-base/trunk/configure.php --output=manual.xml --disable-libxml-check
 
# generate xhtml output in the output/ directory
phd -d manual.xml -f xhtml

Note: you can safely ignore the E_USER_WARNING about "Image does not exist" when generating the output.

I was frustrated to say the least, but knew I could not have been the first one.  I found an archived email sent by Thomas Koch to the VIM mailing list describing some of the problems with the management of scripts and the improvements he would like to see made.  Thomas also created a wiki page to organize his proposed relauch.

The great thing is that the website is built using PHP and MySQL.  That means I can lend a hand to make these changes happen.  I really don't want to see this project die off, so I have emailed the vim@vim.org mailing list prompting for more action.