Shane Smith <[EMAIL PROTECTED]> wrote on 08/24/2005 03:59:08 PM:
> > I am definitely not keen on doing this for first class components i.e.
> > fsm-input all out, but it might be worthwhile to make an exception for 
the
> > rdc:template (given it is a rapid prototyping tool).
> 
> I agree as well, but if we allow it for one, we should allow it for
> all of them, at least for consistency.  As a developer, I would
> scratch my head in bewilderment if grammar-list only worked as a
> config element for one rdc.  The best we can do is encourage best
> practices when it comes to this type of development, and leave it to
> the developer to learn from his mistakes.  (some of us never learn)
<snip/>

The rdc:template was moved out of the "component" sections of the tag docs 
a couple of weeks ago (you'll notice the change on the new link I sent in 
my last email). I thought it was important to stress that a rdc:template 
instance is not a first class RDC. There truly is an important distinction 
between component tags and the rdc:template - the latter is a higher order 
function of sorts, if the atomic RDCs are a first order function. Template 
instances give us more freedom, along with handing us more responsibility. 
So in my mind, there is no breach of consistency. But I do value your 
opinion, and that is possibly how other developers and users think as 
well. In which case, from a framework perspective, we're better off doing 
nothing here.

-Rahul

Reply via email to