[jsr-314-open] [293-Ajax] REOPEN rename <f:ajax> to <f:clientBehavior>?

[Andy Schwartz]

Hi Ed -

I can see the argument that we should provide a generic 
<f:clientBehavior> that plays the same role that <f:converter> and 
<f:validator> play for Converters and Validators - ie. allows the page 
author to attach an arbitrary attached object instance by id/binding.  
(I kind of thought that we might have provided this already, but looking 
at the release candidate 2 PFD tag docs, I do not see this.)  However, I 
do not see <f:clientBehavior> as as a replacement for <f:ajax>, but an 
addition to our tag API.  Since such an addition would make the Behavior 
API more consistent with our Converter/Validator APIs - something we've 
been striving for from the start of the Behavior API design process - I 
would be happy to see this go in.

I am very much opposed, however, to removing the <f:ajax> tag in favor 
of a generic <f:clientBehavior> tag, for the same reasons that I would 
be opposed to removing, say, <f:convertNumber> in favor of the generic 
<f:converter> tag.

Andy

Ed Burns wrote On 4/23/2009 4:43 AM ET:
> I am opening the 2.0 door a crack to consider this one renaming change.
> Don't expect that I'll let other things through.
>
> Several people have suggested renaming f:ajax to f:clientBehavior.  Here
> are some arguments for and against renaming.
>
> Arguments for renaming
>
> * more correct.  The attached object is a clientBehavior
>
> * it's possible and likely to use the tag for non-ajax things
>
> * this is the 20% case
>
> Arguments against renaming
>
> * most users will be using the tag for Ajax
>
> * IDE autocomplete is important to consider.  Many new users rely
>   heavily on this feature and if we don't have an ajax tag, we force
>   them to look at the docs.  Autocomplete junkies hate that.
>
> * this is the 80% case
>
> My opinion is not to rename, but Pete, Martin and I have different
> thoughts on this so I'll bring it to the EG.
>
> Ed
>