Inasmuch as I was talking at Balisage about adapting the Saxon-CE event model to XSL-FO, I ended up talking with Lauren Wood and Matt Patterson about the general case for a XSLT 2.0 processor, or a processor for a subset of XSLT 2.0, that could be used in language bindings and web browsers as a replacement for the stuck-at-XSLT-1.0 libXSLT processor.
Since we were going to talk about it more at the XML Summer School anyway, I did a five-minute lightning talk (slides) at the “unconference sessions”. Happily, it got more people at the Summer School trying things. Michael Kay of Saxonica was at the presentation, and not only did he welcome the idea of compiling Saxon to C, Saxonica now state that they’re looking at the possibility.