$KCODE='u'\x0d $<.map{|i|s=c=0;n=1;h={'\xe5\x8d\x83',1000,'\xe7\x99\xbe',?d,'\xe5\x8d\x81',10};i.chop.chars{|c|s+=n*h[c].to_i\x0d n=%W[d \xe4\xb8\x80 \xe4\xba\x8c \xe4\xb8\x89 \xe5\x9b\x9b \xe4\xba\x94 \xe5\x85\xad \xe4\xb8\x83 \xe5\x85\xab \xe4\xb9\x9d].index(c)||1};s+=n if !h[c];p s}
Note that non-ascii characters in the above source code will be escaped (such as \x9f).