In his latest post Chad Minick shares some thoughts as to why he thinks PHP frameworks are obsolete and how, despite there still being a use for them, a lot of them are just trying to solve the same problems all over again.
Now before I start getting hate mail, let me explain that I still think there is a market for PHP frameworks. I think if you have a moderately low traffic site that's going to live on shared hosting, and is basically a CRUD application on top of MySQL, disregard this article, it probably isn't for you. However, I've seen so many projects start out with that scope. They either grow or the developer/client wants to add more fancy new features they are starting to see elsewhere on the web. Then I see all these PHP developers hacking things that the PHP stack really isn't meant to do.
He takes some of the core features of several of the popular frameworks out there and breaks them out into a list of technologies that do that work themselves (such as "Web Server", "Dependency Management", "Caching" and "Messaging"). He points to a Java/Scala framework, Play as an example of having several of these technologies built-in where you don't have to have a whole other set of skills to use them effectively. It's an interesting idea but seems a little NIH to me.