Good news! I did a bunch more work over the weekend to take the feedback from Friday's call about the triples serialization format and have now fully documented the rules for
(a name suggested by Joseph that I think works very well). Quad tree format retains all the benefits of triples format with the following additional features:
- Even easier to read due to XDI attributes having their own level
- Directly parallels XDI graph structure and addressing syntax
- Directly analogous to RDF 1.1 datasets based on RDF quads
- No separate display format needed 90% of the time (and a trivial transformation into quad tree display format for the other 10%)
- JSON quad tree format WITHOUT implied context statements
- Same graph in quad tree display format
- JSON quad tree format WITH implied context statements
- Same graph in quad tree display format
The only thing I haven't done is added another example of a typical XDI message in quad tree format; I'll try to do that before Friday's call.
Please read it over/try it out/send any thoughts or feedback to the list.
Cooking with gas!
With these things done I can get back to the How XDI Builds on RDF paper. No promises for how much I can get done before Friday but we'll see.
=Drummond