'trace-based just-in-time specialisation for dynamic languages'
Nice choice to demo on. That, peeps, is the skull beneath the skin.
Mozilla's come right out and said what plenty of people think about PDF-reading browser plugins: they're often woefully insecure, generally replicate things browsers can already do and rely on proprietary technology. Mozilla's answer is HTML 5 allied with the pdf.js project, which together allow a newly-released beta of …
I mean there are things you cannot do in JS. For example executing binary files on the computer, or file access on the computer, or buffer overruns. Without those most PDF applications won't work, and PDF is reduced to a simple document format.
BTW here's a nice talk from 2010 about PDF showing some of the nice features.
http://media.ccc.de/browse/congress/2010/27c3-4221-en-omg_wtf_pdf.html
.deb
packages