Re: an alternative for microformat-like simplicity -- the security angle

Manu Sporny wrote:
> Ben's proposal is concerned about not having to dereference anything
> during the RDFa processor stage because it leads to complexity and the
> possibility of not generating triples at parse-time.

During the telecon today, I brought up a point that hints at the
architectural complication that results from
"dereferencing-in-the-parser": security.

Specifically, a pure JavaScript parser (i.e. bookmarklet) would be
unable to produce any output, since it cannot dereference a third-party
resource. Of course, a plugin, or JavaScript with elevated privileges,
could, but the core issue remains that the parser now needs to consider
a number of security issues related to third-party content, just to
produce the first batch of triples.

-Ben