Sort twenty requirements into the four document types
Task
Apply the three signals -- mandatory or not, names a technology or not, a sequence or not -- to twenty real requirement statements, and find the ones currently written in the wrong document.
Steps
- Extract twenty requirement sentences into
lab/doc-sort.csvwith headerid,sentence,mandatory,names_technology,is_sequence,verdict. - Fill
mandatoryfrom the modal verb,names_technologyfrom whether a version, algorithm, product or threshold appears, andis_sequencefrom whether ordered steps are described. - Derive
verdictmechanically: not mandatory -> guideline; sequence -> procedure; names technology -> standard; otherwise policy. Do not override the derivation with judgement yet. - Add a column
currentrecording which document each sentence is actually in today, and mark the rows wherecurrentandverdictdisagree. - For each disagreement, write one line in
lab/doc-sort-notes.mdsaying what it costs -- usually that a technical value now requires board-level change control.
Verify
awk -F, 'NR>1 && NF>=7 {n++} END {print n" sentence(s) classified"}' lab/doc-sort.csv
awk -F, 'NR>1 {print $6}' lab/doc-sort.csv | sort | uniq -c
awk -F, 'NR>1 && $6!=$7 {n++} END {print n" misplaced"}' lab/doc-sort.csv
grep -c . lab/doc-sort-notes.md
Twenty rows, all four verdict values present, and at least one misplacement with a note. A set with zero misplacements means the set was too small or the derivation was overridden.
This is an independent study companion for CompTIA SecurityX CAS-005 and is not produced by or endorsed by CompTIA.