Interesting: it seems that #CommonLisp does not have a PARSE-FLOAT function or something to that effect. It has PARSE-INTEGER but to parse a float, you either need to build it yourself or use an external library.
Or did I just not find it in the specs?