Decision 0010 step 3 asks for a three-build comparison and names no command #5

New issue

Closed

opened 2026-09-28 00:12:27 -03:00 by marcos-mendez · 0 comments

marcos-mendez commented

2026-09-28 00:12:27 -03:00

(Migrated from github.com)

docs/decisions/0010-composable-units.md accepts the component split and
sets the condition every extraction has to meet. Step 3 of "The order of work"
reads:

Only then move components one at a time as we touch them, re-running this
three-build comparison for each, including the control-vs-control run.
Never as a migration.

"This three-build comparison" points at the narrative of the experiment
earlier in the note, not at anything anybody can run. The note describes the
three builds, the tables and the 173 explained files, but it names no command,
and until now no command existed: the capture was a scratch file in root's
home directory on the build host and the comparison was ad hoc shell that was
never committed. Two component extractions, Keel-Linux/keel-mariadb#11 and
Keel-Linux/keel-postgresql#7, were approved on a number only their author
could produce.

Keel-Linux/buildtasks#10 fixes the missing tool. This issue is the other half:
the decision has to name it, so that step 3 is an instruction and not a
description of something that happened once.

Expected: step 3 of 0010 names the command, says which three captures it
takes and what a pull request is required to quote from its output, and the
note records the correction the tool makes to the measurement it reports, that
a path inside the noise floor is not evidence of noise on its own. Amend 0010
in place; this is not a new decision.

`docs/decisions/0010-composable-units.md` accepts the component split and sets the condition every extraction has to meet. Step 3 of "The order of work" reads: > Only then move components one at a time as we touch them, re-running this > three-build comparison for each, including the control-vs-control run. > Never as a migration. "This three-build comparison" points at the narrative of the experiment earlier in the note, not at anything anybody can run. The note describes the three builds, the tables and the 173 explained files, but it names no command, and until now no command existed: the capture was a scratch file in root's home directory on the build host and the comparison was ad hoc shell that was never committed. Two component extractions, Keel-Linux/keel-mariadb#11 and Keel-Linux/keel-postgresql#7, were approved on a number only their author could produce. Keel-Linux/buildtasks#10 fixes the missing tool. This issue is the other half: the decision has to name it, so that step 3 is an instruction and not a description of something that happened once. Expected: step 3 of 0010 names the command, says which three captures it takes and what a pull request is required to quote from its output, and the note records the correction the tool makes to the measurement it reports, that a path inside the noise floor is not evidence of noise on its own. Amend 0010 in place; this is not a new decision.