PEAR installer adoption rate statistics

Hi all, I've put together a simple cron job at pear.php.net which parses the apache log files to generate statistics about the adoption rates of PEAR installer and PHP version. Here is the site: http://pear.php.net/~greg/stats/ The data is indexed by the date that the cron job runs, and so will update over time. To get this information, the script greps over the httpd access log for all GET requests of a .tgz file from pear.php.net with a user agent of "PHP/*" or "PEAR/*". PEAR version 1.4.0 and earlier sent "PHP/version" as the user agent, as in "PHP/4.3.11", since PEAR 1.4.1, the user agent string is "PEAR/pearversion/PHP/phpversion" like "PEAR/1.4.5/PHP/5.1.0" This obviously only shows people who have typed "pear install blah/upgrade blah/upgrade-all" but I would imagine this is a target audience for deciding what versions of package.xml and PHP you want to support. This is because people who are downloading through distros or other systems will have the right mix of PEAR Installer/packages because this is usually done by hand - and it's also not in our hands to control this. It will be interesting to see how the versioning trend changes over time, but it looks like PEAR end-users are keeping up with their PHP version. This is serious incentive for PEAR folks to get their asses in gear and start writing PHP 5.1+ code to support the larger user group, and good news for pecl devs who are worried about supporting PEAR 1.3 and package.xml 1.0 - you don't need to do this any longer, using package.xml 2.0 as the only package.xml for new extensions is a viable option with only 3-4% of users using PEAR 1.3. Greg