Point well taken. In fact, I see this as one of the areas in which ZF must be strong to achieve even its most basic goals- along with security, code quality, and extensibility. Every feature we add must not compromise any of these aspects, or we risk reducing the overall value of ZF and sowing skepticism towards all ZF components in the PHP community.
I have to be a bit careful about forward-looking statements, but I believe it's worth sharing with everyone that we have begun an initiative at Zend to quantify the performance of ZF and identify any bottlenecks. The deliverables for this exercise include benchmarks for ZF and other major PHP frameworks against use cases that reflect real-world apps, section in the reference guide (most likely as an appendix) pointing out performance best practices, profiling ZF/creating issues for each significant performance bottleneck, and some fixes for the lowest hanging fruit bottleneck-wise that maintain full backwards compatibility. Ralph is heading up this effort, and I'm sure he'd be interested in any suggestions or help in fixing what issues we may find. ,Wil No Zend_Magic yet :/ Seriously though :), one of the things I see missing from the ZF message, is Performance. If a colleague or a client asks me how well ZF performance is, I've no idea. Maybe it's there somewhere, but I missed it. I understand performance is hard to benchmark in ZF because everyone can do her or his thing in a different way. But maybe component based
