The 'wc' (word count) program that comes with most Unix/Linux systems is inexplicably slow. I mention this because people like to implement it using other languages to demonstrate why they are just as fast as C.
news.ycombinator.com/item?id=215597…
news.ycombinator.com/item?id=215597…
It's a simple program. It just counts the number of words in a program, where a "word" is anything separated by space.
In other words, the previous tweet has 1 lines, 22 words, and 117 characters, according to 'wc': 
So 'wc' is trivial. I write very complex parsers. So I often test my parsers against 'wc' for laughs. They always end up being at least ten times (10x) the speed of 'wc'. For example, I wrote a DNS server that needs to parse the "zonefile", the text file holding all the records.
This parsing task can be slow. The BIND DNS server can take 20 minutes to parse the 8-gigabte .com zonefile. My DNS server does this in under 20 seconds, so it's fast. Running 'wc' on the file takes a couple minutes.
I haven't really looked into 'wc' to see why it's slow. I suspect part of the answer is that the problem is more complex than is it seems at first blush. It has a lot of options, and can parse Unicode.
I suspect the other part of the answer is that it's dumb. The way people are taught to parse files/packet in college is dumb, helladumb, by dumb professors. So it tends to be slow, so anybody doing not-dumb stuff is sure to be an improvement.
I suppose I need to simply write a 'wc' program in C in order to prove that code written in C can be faster than C.
