Print

Print


On Sun, May 22, 2011 at 3:03 PM, Saašha Metsärantala <[log in to unmask]> wrote:

....

> The following JSON object would also be an extension of a standard XSD
> date-time object representation:
>
> {
>
>  "year": 2009,
>
>  "month": 3,
>
>  "day": 4,
>
>  "questionable": ["year"]
>
> }
>
> Likewise, the following date
>
> (2009)?-03-(04)?
>
> could be represented as
>
> {
>
>  "year": 2009,
>
>  "month": 3,
>
>  "day": 4,
>
>  "questionable": ["year", "day"]
>
> }

True.

So long as we're considering these practical implementations details,
and no trying to boil oceans (e.g. trying to cover every possible
corner case to such an extent that this spec will fall under its own
weight), then that's fine.

But I still think this is too much. If others don't agree with me, so be it.

Bruce