RE: Resolutions for features at risk [JSON-LD]

On Monday, June 03, 2013 6:52 PM, Dave Longley wrote:
> On 06/03/2013 07:35 AM, Markus Lanthaler wrote:
> > Would you support changing the conversion of JSON numbers to always
> > produce xsd:doubles?
> 
> This question was for Gregg and Sandro, but I don't think that I would
> support *always* converting JSON numbers to xsd:doubles. I'd like this
> use case to be supported:
> 
> Read some data from RDFa that has a datatype of xsd:integer. Convert
> the data to JSON-LD w/native number conversion on. Work with the data,
> possibly modifying the number. Convert the data to back to RDF (N-
> Quads) where the value (object) has an xsd:integer datatype and it
> looks like an integer (not a double).
> 
> This use case is currently supported by the spec.

But it breaks for large integers and doubles that happen to have no
fraction. That was a proposal by Sandro or Pierre-Antoine. I do not feel
strongly about this but I think it allows us to simplify the algorithms
without losing much functionality.



--
Markus Lanthaler
@markuslanthaler

Received on Monday, 3 June 2013 20:00:27 UTC