The assertions in this article, and in everything else I've read, indicate that the people behind JSF have no familiarity with Struts whatsoever--they pretend in their pronouncements that JSF is solving problems that have already been solved in Struts--especially the one described here, that of flexible navigation.
Having said this, I know that my assertion (that these people are not familiar with Struts) is totally untrue--Craig McLanahan, Struts lead developer, is the lead on Sun's JSF project. So why are the JSF proponents pretending that Struts does not exist when touting JSF? Struts is far from perfect--the very feature I talk about is nicely done in Struts but could be more easily configurable by application assemblers. But, in typical Sun fashion (can you say "EJB: the next generation???") they have created a system that does "more" than Struts, but at the cost that the simple things you could have done relatively easily with Struts are now ridiculously overcomplicated. Once again, Sun does not get the maxim "keep the simple things simple!" (Something many open source projects also frequently don't get, of course.) Showing off how "flexible" a framework is while noting that the simplest "Hello World" application is now a complex mess does not demonstrate how good that new framework is--instead, it demonstrates that the forces that got us to where we are today technologically do not get what is really important in the success of a framework.
On the one side, we have Microsoft with their "wizards", making simple things simple, but making anything beyond the simple impossible to implement without sophisticated tweaking. On the other side, we have Sun, demonstrating how amazingly flexible their frameworks are but simultaneously how amazingly difficult they are to use for even the simplest things.
It is possible to build frameworks that keep simple things simple AND allow maximum flexibility. The question is, where are they?