The RDFa parser currently provided has several flaws:
-
Highly recursive - can hit OutOfMemory or StackOverflow exception for heavily nested HTML
-
Is limited to HTML only
-
Code is complicated by need to support RDFa 1.0 and RDFa 1.1 which have some significant differences
Instead a new abstract RDFa Core 1.1 parser will be developed which will allow different host languages - HTML, XML, SVG - to be plugged into it as desired. The parser should be pure RDFa 1.1 and invoke the old RDFa 1.0 parser only where the specification says it should