I'm not really experienced in @font-face, but I think it's specifing a different webfont for single unicode characters (using the range attribute).
Chrome will then try to fetch the font from the many URLs, making an HTTP request for every character in the field (say "Hello" will try to fetch ?H, ?e, ?l, ?o) in the order it finds them.
If you're on the other side (ie. the server serving the font) you can rebuild the original string by seeing which requests have been made and in which order.
Haven't tried it, but I doubt it'll try to fetch the same letter twice, so most of the time the string wouldn't be complete (but again, I haven't tried so I don't know for sure).
This may not work exactly as planned if your target string is expected to be long enough. You can see from his example that the fetching takes place out of order, as expected from asynchronous calls.
If you're on the other side (ie. the server serving the font) you can rebuild the original string by seeing which requests have been made and in which order.
Haven't tried it, but I doubt it'll try to fetch the same letter twice, so most of the time the string wouldn't be complete (but again, I haven't tried so I don't know for sure).