[adamjohnwright]

Spec 002 T019 (specs/002-analysis-report/tasks.md): the comparison's run, recorded in research R9. This PR changes documents only. T019 is ticked here, with what is left open.

The run

• T002's tokens were not kept anywhere, so I made one fresh token the way T001 made its own (R5). It is an over-representation analysis of seven gene symbols and two made-up identifiers, posted on loopback to the development host's own analysis service, whose data are release 98.
• Today's report and pdf-export's were each made at 25 and 50 pathways, from the same services and a private copy of the release 98 graph. Each pair was compared by compare-report with Spec 002 T019: the comparison reads today's report #212.
• The reports are kept outside the repository and deleted after the run.

Every part was read on both sides, and every part agreed except three:

• The order of edits within a day. A bug, fixed in Spec 002: today's order within a day is Reviewed, Authored, Created #213.
• Modified rows, for a decision. 20 of the 50 pathways have two modified edits in the release 98 graph; R2 had said there is one. Today's prints no Modified row for such a pathway, and one row where there is one edit. pdf-export prints a row for each.
• References (1 at 25 pathways, 2 at 50):
  • A Greek letter that today's typeface lacks is written as U+0000. This is now a listed difference in the spec's Assumptions.
  • Book and web-page references are printed in spec 001's form. Today's prints a web page's address and a book's publisher after a comma. This needs a decision.

The grouping by day (T007) holds. On the two pathways with two edits of one kind on one day at different times, both reports print one row.

Not run here:

• the expression, interactors and species-comparison tokens;
• each figure's overlay check (quickstart §3).

🤖 Generated with Claude Code