Explanation about value node rules for tomorrow's TC call

On the XDI2 list this week, in response to seeing a demo of the fantastic new version of the XDI Graph Editor from Hubert and Jingning, I sent the following explanation of the rules for using value nodes. Markus added this topic to the agenda for tomorrow's TC call because we felt it was important to clarify how this works with all XDI operations.

=Drummond

On Thu, Aug 14, 2014 at 5:05 AM, =Drummond Reed <drummond@connect.me> wrote:

Hubert & Jingning,

Congrats. I saw a demo of the new version today when Brian Wu used it to show Les and PeterD and Isome XDI dictionary graphs. I loved it! It's fantastic to now be able to easily pick out root nodes, entity nodes, and attribute nodes. Like Markus, I too was looking for the value nodes (empty diamonds), but that's already been dealt with on this thread.

Actually, here's one important note about value nodes and XDI visual graph notation: the XDI TC decided that there could be a "long form" and a "short form" of the notation:

The "long form" would require showing the literal arc to the actual literal node (which as I remember was represented as with a square box). The literal value then would appear next to the square box.

The "short form" would not require showing a literal arc or a literal node, but just showing the value node and then, if it had an actual value, showing that value in quotes next to the value node. Thus in the short form the literal arc and literal node were taken as implicit.

Examples:

LONG FORM

=drummond <#email> & &

O --------------> ● -------------> ♦︎ ------------> ♢ . . . . . . . . > ☐ "dsr@example.com"

SHORT FORM

=drummond <#email> &

O --------------> ☻ -------------> ♦︎ ------------> ♢ "dsr@example.com"

I myself am perfectly happy with the short form, since it's easier. However it is true that the long form is the only one that shows the literal arc and the & predicate that corresponds with it.

Note that in either form, however, the open diamond is needed to indicate the presence of a value node. And if the value node is present, it MUST have a value (even if that value is JSON null). The only way to have an attribute with no value (not even null) is to not have a value node.

Markus responded:

I think this makes sense, and I remember the discussions we had on the TC about this.
XDI2 doesn't currently do this right, i.e. in theory it allows a value node without a literal node.

I just opened a Github issue for this:
https://github.com/projectdanube/xdi2/issues/11

xdi message