We had rudimentary handling for quotes in French and Spanish (they need different open- and close-quote characters), but this was failing for the French abstract of the latest article, where a <cit> tag containing a quote appears. I elaborated the system a little to allow for more flexibility, in both the PDF and XHTML output XSLT. Now, if you put a rend="fr" (or "es") attribute on any element, quotation marks inside that element should be rendered using the appropriate quote marks.

Warning note to self: this doesn't handle nesting. In other words, it only looks up the tree for any ancestor which is fr or es, so if you later add an attribute on a lower-level element setting it back to en, that won't make any difference. I don't anticipate this kind of embedding in ScanCan documents, though.