Here are performance numbers for a variety of approaches using a list of length 256. You can see that having a local reference to the global chr function (as discussed in the article) can give a slight speed boost. Using join() in conjunction with map() performs fairly well.

  python -m timeit -s 'l = range(256);' '"".join([chr(i) for i in l])'
  # 10000 loops, best of 3: 80.5 usec per loop
  python -m timeit -s 'l = range(256); lchr = chr' '"".join([lchr(i) for i in l])'
  # 10000 loops, best of 3: 75 usec per loop
  python -m timeit -s 'import itertools; l = range(256); lchr = chr' '"".join(itertools.imap(chr, l))'
  # 10000 loops, best of 3: 52.8 usec per loop
  python -m timeit -s 'l = range(256); lchr = chr' '"".join(map(chr, l))'
  # 10000 loops, best of 3: 48.9 usec per loop
  python -m timeit -s 'l = range(256); lchr = chr' '"".join(lchr(i) for i in l)'
  # 10000 loops, best of 3: 87.2 usec per loop
  python -m timeit -s 'l = range(256); lchr = chr' '
  s = ""
  for i in l:
    s += lchr(i)
  '
  # 10000 loops, best of 3: 64.2 usec per loop