the imported wc from sources is arround 8 times slower than our old one. it is common to run wc on large log files to count lines. so i think the implementation complexity is justified. (just like with grep)