Re: [apps-discuss] Question about URI template and expansion of an empty list

[Erik Wilde <[e-mail]>]

hello james.

On 2013-04-08 07:50 , James M Snell wrote:
> "empty_keys" is not defined explicitly as an empty associative array.
> The only definition given for "empty_keys" in section 3.2 is
> "empty_keys := []" which is semantically and structurally identical to
> an empty list. So to amend your rules, we have:
> - an undefined value produces "X"
> - a defined-but-empty string produces "X."
> - a defined-but-empty list/associative-array produces "X"

i agree that since JSON doesn't have associative arrays, [] as a value 
has to be read as just being an empty array/list. but i am still 
wondering about francis' "why" question: if we did not have that test 
case and simply had to reason based on the spec test, how would i end up 
with "X" instead of "X."? when i am just reading the text as somebody 
trying to write code implementing it, i read the spec as telling me "X." 
is what should be the result of expansion.

cheers,

dret.

-- 
erik wilde | mailto:[e-mail]  -  tel:[phone] |
            | UC Berkeley  -  School of Information (ISchool) |
            | http://dret.net/netdret http://twitter.com/dret |