Re: [xdi] comments on JSONSerializationRules

On Mon, Aug 27, 2012 at 8:47 AM, Drummond Reed <drummond.reed@xdi.org> wrote:

On Thu, Aug 23, 2012 at 9:08 AM, Markus Sabadello <markus.sabadello@xdi.org> wrote:
XDI TC,

Based on experience from recent implementation work, I've added two comments to the JSONSerializationRules/Discussion page.

One is about nested cross-references, e.g. a statement like the following:
@!9999!8888$($msg)$(!1234)$do/$add/(=a/+b/(=x/+y/=z))
Markus, I definitely agree with your comment there.

The other comment is about cases where only some of the relations of a given type have cross-references, e.g.

        @!9999!8888$($msg)$(!1234)$do/$add/=markus
        @!9999!8888$($msg)$(!1234)$do/$add/=!1111
        @!9999!8888$($msg)$(!1234)$do/$add/(=!1111!2222!3$!(+age)/!/(data:,33))
        @!9999!8888$($msg)$(!1234)$do/$add/(=!1111!2222!3$*(+tel)/+home/=!1111!2222!3$*(+tel)$!1)

On this one, while I agree this solution works, the comment I posted in reply was: "I'm wondering if the rule shouldn't be that the target of all '''$add''' statements should be a complete XDI statement, and therefore it must always be a cross-reference.''

Also posted a reply..
"Maybe, but this isn't really specific to $add, you could also construct more generic examples with the same issue:"

=a/+b/=c
=a/+b/(=x/+y/=z)

What do you think? We can always discuss on this Friday's telecon.

=Drummond

xdi message