Re: T5: default binding is prop even in template

[Martin Strand]

Per-parameter default bindings is a great idea.
But as D&J said, this shouldn't be used everywhere but only for the most  
obvious ones such as value, source, validate, etc  And, you must always be  
able to use an explicit prefix if you want. Any parameters without a  
default prefix are still handled as before (literal: in templates and  
prop: in annotations) so informal parameters will work as expected.

Martin

On Wed, 17 Jan 2007 15:49:49 +0100, B.S.Navin <[EMAIL PROTECTED]> wrote:

I too am for default binding types for parameters - I feel, it does not  
make much difference if a developer needs to look at the documentation  
to know what's the default binding, or if he needs to explicitly specify  
the binding. In either case, he needs to know the binding type for that  
parameter.

So default bindings for parameters shouldn't affect developers much.

I also second Markus' idea of allowing the default to be overridden.  
This feature may be really helpful for projects that use custom  
bindings, whose values finally result in the same Class type.

- Navin

On 17-Jan-07, at 6:25 PM, Markus Joschko wrote:

Wouldn't it be possible to combine both approaches?
If no explicit binding type is given a default binding type for a
parameter is assumed?
The disadvantage is that it might confuse people as it allows to do
the same thing in two different ways.


On 1/17/07, D&J Gredler <[EMAIL PROTECTED]> wrote:
I like the direction Howard is taking with the bindings, but I think  
Ron has
a good point. What matters is not only what the developer has to *do*,  
but
also what the developer has to *know*. This extends even to areas like
Jesse's suggestion of making this configurable: if this is  
configurable,
that configurability is just one more thing that Tapestry developers
can/should be aware of.

A good test might be to ask: "given that the developer knows the  
standard
available binding types, will it be obvious which binding type  
corresponds
to this parameter?" Once this is done, it becomes a matter of debate  
as to
whether the increased load on the developer (having to conciously think
about the obvious default for a parameter) is worth the savings (less
typing, more concise code, DRY).

I'd say continue down the per-parameter-defaults path without even  
making it
configurable, as long as they're all obvious and it's not going to  
create
any rifts in the community...

Just my opinion, of course.


On 1/17/07, RonPiterman <[EMAIL PROTECTED]> wrote:
>
> At the late alpha / early beta days of T4, I was getting very  
confused
> with default binding, and argued strongly against it.
>
> I still think this can be very confusing - and am wondering if there  
is
> anything one could do to avoid the confusion:
>
> (Is default prefix going to affect both templates and annotations?)
>
> One thing we could do is decide as a development guideline, only very
> obvious bindings get a default prefix:
>
> most obvious for me are listener bindings, (today's) source, value  
and
> index in For, value in Form - any others com in mind?
>
>
> I must say I have my thoughts about the way the new mantra "let the
> framework do things for the developer" is being aplyied:
>
> for me, the mantra would be : "let the developer KNOW LESS" instead  
of
> DO LESS.  The default binding issue is not different. If I have to  
look
> at the docu in order to KNOW which default binding a parameter has -
> then its a bad feature. If used with great care - then maybe it can
> rock...
>
> Just my two Euro-Cent...
>
> Cheers,
> Ron
>
>
>
>
>
> Howard Lewis Ship wrote:
> > I'd really like to explore the potential for this approach (default
> > binding prefixes for parameters).
> >
> > I haven't gotten a lot of feedback ... Kent is very opposed to the
> > idea.  Jesse and a few others are for it. I think it will be neat.
> >
> > I'd don't want to ride roughshod over Kent's wishes, but I don't  
know
> > how to proceed ... we need a protocol for evaluating the results of
> > this experiment and deciding whether to commit to it or not while  
the
> > Djinn is still in the bottle.
> >
> > Ideas?
> >
> > On 1/16/07, Jesse Kuhnert <[EMAIL PROTECTED]> wrote:
> >> That should be two. ;) (maybe my previous post wasn't clear on  
this)
> >>
> >> On 1/16/07, Howard Lewis Ship <[EMAIL PROTECTED]> wrote:
> >> > Well, that's one :-)
> >> >
> >> > On 1/16/07, Nick Westgate <[EMAIL PROTECTED]> wrote:
> >> > > +1 for Howard's per parameter defaults. It looks great.
> >> > >
> >> > > I was firmly in the other camp, but the time seems right with  
T5.
> >> > > (Still pondering Kent's latest post on this.)
> >> > >
> >> > > Cheers,
> >> > > Nick.

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]