@John: it could be good to know if your reasons to object JSF. In my
understanding JSF 2 solved many of the problems, and it have some very
useful features. Maybe if some idea is enough useful, myfaces
community can push to include it on next JSF 2.2 spec. Suggestions,
critics and tomatoes are most welcome ;-). That's how innovation
works, right?

Try developing a mostly read-only/downloads application...JSF is overkill for 
that.
State your reason for using JSF for a read-only/downloads application.

I've been thinking about using Opa http://opalang.org/  Compiles to exe.
Please study Opa make Java/JSF as easy to use as that.  It would be good to see 
a head to head comparison (same apps) to see which one is easier to read, 
quicker to build etc.

Also, caching is a nightmare with JSF.  You can cache in application scope, but 
then you have to provide a refresh button.  JSF needs a way to provide caching 
so that the application programmers don't need or care to know how many times 
the getters are called.  A correct caching scheme will avoid referencing the 
database too many times.  I realize we should use JPA instead of JDBC.   There 
needs to be a refresh button (or flush cache) with JPA, and the application and 
session scope caches (they should work in concert with the refresh button on 
the browser).  Can one tell if the refresh button has been pressed by looking 
at HTTP?   I don't know.  Maybe we need an extension to HTTP, so we don't get 
those stupid popups anymore (do you want to resend?) There's a requirement to 
both fetch new data, and a requirement to cache data for speed.  A good 
application framework would support both w/ the application developer having to 
build their own cache and refresh technology.  There's a requirement to cache 
across multiple users, and many "experiments" or main objects of concern 
selected from a list (although the most recent is the highest priority).

The amount of overhead and memory required to run JSF is a concern.   It seems 
like Oracle just wants to sell servers instead of making a good technology.

John

Reply via email to