[jsr-314-open] inter-component and form-level validation

[Andy Schwartz]

Pete, Dan -

Thanks for getting this discussion going.  Don't have time to tackle 
this in detail, but for the moment I just wanted to make a few high 
level comments...

Pete Muir wrote:
> I think we should also check that everyone agrees that cross-component 
> validation should provide first class support for BOTH (a) custom JSF 
> validators AND (b) model based. My vote would be to require (b) and 
> make a best effort at (a), but I suspect other disagree ;-)

I agree that we need to start with this question - ie. we need to start 
by nailing down which use cases we are trying to tackle.  I suspect that 
we'll want to both facilitate pre-update model cross-component 
valdiation (a) as well as post-update model cross-field validation (b).  
One thing that might help would be to look to the community to get a 
better understanding of which problems/use cases are most critical to 
solve.   For example, I think we should take a look back at use cases 
like this:

http://www.jroller.com/robwilliams/entry/jsf_multi_field_validation_not

And try to figure out which solution (or solutions) would have helped 
the obviously frustrated user solve the problem in the least painful 
way.  I am sure there are other use cases like this that we can pull 
into our discussion (though haven't had a chance to dig these up).

Regarding pre-update model validation (a), I noticed that our #1 
favorite issue:

https://javaserverfaces-spec-public.dev.java.net/issues/show_bug.cgi?id=1

Is now closed.  I believe the thinking was that the availability of 
post-validation system events in JSF 2.0 should make pre-update model 
cross-component validation more manageable.  So, perhaps our next 
questions are: how far do system events get us towards a pre-update 
model cross-component validation solution and whether there are aspects 
of the system event-based approach to cross-component validation that 
beg for a different/enhanced solution?

Oh, and, btw... given that our cross-component validation spec issue is 
currently marked as closed/fixed, perhaps we should open up some new 
spec issues to track requirements that have not yet been covered.

Andy