Re: a question about framing test 16

> On 2016-04-05, at 19:11, Dave Longley <dlongley@digitalbazaar.com> wrote:
> 
> On 03/30/2016 05:11 PM, james anderson wrote:
>> 
>>> On 2016-03-30, at 20:53, Dave Longley <dlongley@digitalbazaar.com
>>> <mailto:dlongley@digitalbazaar.com>> wrote:
>>> 
>>> On 03/30/2016 12:09 PM, james anderson wrote:
>>>> good afternoon;
>>>> 
>>> 
>>> Does this adequately explain how to use these flags?
>> 
>> you describe the embed flag. ok i can follow that. my question was
>> about the effects of a type constraint, whether explicit or
>> duck-types. as far as i could tell from the tests, any apparent type
>> constraint had no effect. that was the original question.
>> 
> 
> When a frame is applied to the dataset, each matching node that is found
> will be included as a root of an output tree. Subframes may specified as
> values for particular properties in the frame.

> 
> For each property of a matching node, its associated subframe will be
> used on the set of nodes that are related via that property. If no
> subframe is specified, an implicit subframe will be generated with no
> constraints but that inherits any `@` flags set in its parent frame.

this means, the @type constraint would be inherited, but the member template would not?

> […]
> 
> As mentioned above, type constraints are not "inherited" by subframes,
> so children will match without meeting that constraint.

this gets closer to the question.
does this mean “duck type” constraints are not inherited as the member template is not inherited, but an @type would be inherited, as the generated frame "inherits any `@` flags”?
this is the behavior which i believe i had not seen in the tests.
i may have misconstrued the results.
in any case, if that is the intent, it is a clear implementation requirement.

best regards, from berlin,
---
james anderson | james@dydra.com | http://dydra.com