Identify any structural
limitations (e.g., no texts under n characters should be
included, non-English texts should be excluded, etc). Consider
whether the entire corpus or merely a subset will be hand-coded. In
the latter case, identify how the subset should be selected (randomly,
representatively, etc.)