Well, I consider myself to be somewhat the evangelist for Wicket, and though I had noticed the about.com blog post requesting which framework was the most popular, I found the method of emailing him rather cumbersome, and I prefer open forums so everybody can see who posted what. I can't remember asking anybody to respond, so it wasn't me :-)
Matt also gives some graphs on mailinglist statistics. And yes, Wicket has soared in July with mailinglist activity. More so than 'competing' frameworks such as MyFaces, Cocoon, Echo, WebWork. And he quotes me saying that huge mailinglist activity may be a bad sign. I still stand by my standpoint: it may be an AntiPattern when a mailinglist is highly active. But just looking at the number of messages is not enough: looking at the type of messages is more indicative. It is a bummer that qualitative measurements are nigh impossible to take.
So another rather cool measurement Matt likes to do is download numbers. So lets try to figure out what the stats are for our frameworks (sourceforge is not helping though, and Tapestry downloads are missing):
|Last 2 months
Because I don't have the actual data for the Apache projects, I am not able to calculate them precisely. I have tried to guess the numbers based on the graphs found here.
One thing that is striking is that a lot of people still download the old (2004) release of MyFaces from sourceforge.net. Another is that that number is much higher than the Apache numbers. Struts still dwarfs all other frameworks.
What does this tell us? Not too much. Wicket is at the moment compareable in popularity to MyFaces, and WebWork is seriously falling behind. If you want job security, then download struts now, as it will be around for FAR TOO LONG considering these numbers. When you also want to have fun in your job, download Wicket.