I agree with your point, yet we thought the data we gathered could be interesting, for instance we observe that ~60% of routes have human readable comments, but less than 10% have actual examples.