new escape(): %hh, %uhhhh for everything over unicode 256.

encodeURIComponent(): encode in UTF-8, then %hh

For chrome, which is interpreted as UTF-8, the switch from escape() to encodeURIComponent is necessary and correct. For Mozilla-specific web sites that relied on Mozilla's quirky escape() behavior, the answer is not clear. For English-only sites, escape() is better until Netscape 4 finishes dying.

Btw, encodeURIComponent rocks. I used in the "Blogidate XML Well-formedness" bookmarklet to create the body of "data:text/xml;charset=UTF-8," URLs, allowing the bookmarklet to work with XHTML that contains Japanese text.