In addition to deciding on whether JSF will be satisfactory in terms of performance, you also have to think about some other practical things. Do you already know another web application framework (Struts, Webworks, Spring?) I would definitely recommend a framework over nothing.
If you aren't really familiar with any of those frameworks, then there is no extra cost associated with getting up to speed with JSF vs. the others. In my case I already know Struts, so I'm getting up to speed with JSF on the side and I'm going to slowly start introducing JSF into the project and see how I like it. JSF and Struts can coexist nicely, and its even easier if you just use JSF for certain things and Struts for others. The best way to evaluate the technology is to dive right in and start using it (after reading some good books.) There are plenty of people here and on the Sun forum that can help with your questions. Something else to consider is how easy will it be to maintain and expand your code. Yes performance is a concern, but don't forget about how long it takes to develop the codebase and then add new features. You want to go with something that makes it easier to get the job done (all of the frameworks are better than JSP and servlet with no framework in this regard.) Finally, consider your talent pool. Are you using current employees or are you hiring new programmers? Consider what your existing programmers already know. If a bunch of them already know Spring, then consider that. There will probably be more people in the job market right now that know Struts than anything else (even that is hard to find, trust me.) So maybe that is a factor. JSF is an emerging technology but I have to say I'm very impressed so far. Eventually it will have the maturity and programmer base that Struts and other frameworks now enjoy. So definitely keep up with JSF even if you decide not to use it. sean
