JSON vs XOXO for document serialization

[Manuel Simoni]

JSON is very natural:

{ "system:id": "foobar",
"title": "The Foo Bar",
"related": { "system:id": "quux" } }

Using XOXO, immediately issues pop up:

* using <a> hrefs: where should the quux in the above example go: into
the <a>'s content, or into the value of the href attribute, or both?
http://code.google.com/p/buckybase/issues/detail?id=2

* how to distinguish between links and strings: as long as only plain
strings are allowed, one distinguish them easy from the <a>s. But when
you want to use for example strings with XML content, you'd have to
include them in a CDATA.
http://code.google.com/p/buckybase/issues/detail?id=1