how about just manually stripping out comments
and not counting any spaces, even if in text. If you have more and a few spaces, you would probably have quite a few 'words' which would most likely push up your score anyway. 4 additional 'words' would probably have about 10-15 or more characters in it, the 3 spaces wouldn't make a big deal of difference.
The idea is efficiency... succinct(sp) output should be a requirement also
