Case 451
Neither date was set by the work
neither_date_was_set_by_the_work.eml - Each team set its date from the other team's date. Where either date came from is computed below.
ok: true — round-trip fixpoint reached (python1 == python2)updated 2026-08-19
EML
eml# Self-authored for the EML case corpus (no external origin). Each team set its
# date from the other team's date. Where either date came from is computed
# below.
#
# Aligning to the dependency is correct planning. A team that ships before the
# thing it integrates with is finished has shipped nothing, and a team that
# ships long after has wasted the gap. Reading the other side's date and
# planning against it is what coordination means.
#
# A date read from a plan is not evidence about work. When both sides do it,
# the pair holds two dates that agree with each other and neither of which was
# derived from an estimate, so the first contact with the work moves both.
#
# The work-derived dates are computed alongside so the gap is visible.
# weeks of real work remaining, per team
14 => work_a
9 => work_b
2 => integration_margin
0 => week
# each team's announced date, in weeks from now
16 => date_a
18 => date_b
"real work remaining : team A " + str(work_a) + " weeks, team B " + str(work_b) + " weeks" ^0
"announced dates : team A week " + str(date_a) + ", team B week " + str(date_b) ^0
"integration margin : " + str(integration_margin) + " weeks" ^0
"" ^0
"what each announced date was derived from" ^0
" team A : team B's date " + str(date_b) + " minus the margin " + str(integration_margin) + " = " + str(date_b - integration_margin) ^0
" team B : team A's date " + str(date_a) + " plus the margin " + str(integration_margin) + " = " + str(date_a + integration_margin) ^0
if date_a == date_b - integration_margin:
if date_b == date_a + integration_margin:
" the pair is consistent, and each date's only support is the other" ^0
"" ^0
"what the work implies" ^0
" team A : " + str(work_a) + " weeks of work, so week " + str(work_a) ^0
" team B : " + str(work_b) + " weeks of work, so week " + str(work_b) ^0
" the binding one is team A at week " + str(work_a) ^0
if date_a > work_a:
" the announced date for A is " + str(date_a - work_a) + " weeks later than its own work needs" ^0
if date_b > work_b + integration_margin:
" and B's is " + str(date_b - work_b - integration_margin) + " weeks later than B's work plus the margin" ^0
"" ^0
# ---- a discovery on team B ----
#
# Three weeks of work appear on the smaller side. Because each date is anchored
# to the other, the correction propagates in both directions.
3 => discovery
work_b + discovery => work_b2
"team B discovers " + str(discovery) + " more weeks of work, taking it to " + str(work_b2) ^0
"" ^0
"if each side re-derives from the other, one round at a time" ^0
date_a => a
date_b => b
0 => r
while r < 4:
r + 1 => r
b => old_b
a + integration_margin => b
if work_b2 + integration_margin > b:
work_b2 + integration_margin => b
b - integration_margin => a
if work_a > a:
work_a => a
" round " + str(r) + " : A week " + str(a) + ", B week " + str(b) ^0
"" ^0
if a == date_a:
if b == date_b:
" neither date moved. " + str(discovery) + " weeks of work appeared and the pair is" ^0
" unchanged, because each date is supported by the other and the" ^0
" margin between them still holds" ^0
"" ^0
"where it comes to rest" ^0
" team A : week " + str(a) + ", against " + str(work_a) + " weeks of work" ^0
" team B : week " + str(b) + ", against " + str(work_b2) + " weeks of work plus " + str(integration_margin) ^0
if a > work_a:
" A is still " + str(a - work_a) + " weeks past what its own work needs, because it is" ^0
" anchored to B and B is anchored to the margin" ^0
"" ^0
# ---- what the same discovery costs if the dates come from the work ----
work_a => wa
work_b2 + integration_margin => wb
0 => joint
if wa > wb:
wa => joint
else:
wb => joint
"if both dates were derived from estimates and reconciled once" ^0
" team A : week " + str(wa) + ", team B : week " + str(wb) ^0
" the pair ships at week " + str(joint) ^0
if b > joint:
" against week " + str(b) + " under the mutual anchoring, a difference of " + str(b - joint) ^0
"" ^0
# ---- what the announced dates are evidence of ----
"what a reader can conclude from the two announced dates" ^0
" that the teams agree : yes" ^0
" that either date is achievable : nothing" ^0
" the agreement was produced by copying, which is also what agreement" ^0
" produced by evidence looks like from outside" ^0
"" ^0
# ---- the control: one side anchored to the work ----
#
# One team deriving from its own estimate is enough to give the pair a floor
# that a discovery moves by exactly the discovery.
work_a => ca
work_b2 + integration_margin => cb
0 => cjoint
if ca > cb:
ca => cjoint
else:
cb => cjoint
"control - team A dates from its own estimate, B aligns to A" ^0
" before the discovery : A week " + str(work_a) + ", B week " + str(work_b + integration_margin) ^0
" after : A week " + str(ca) + ", B week " + str(cb) ^0
" the discovery moved the pair by " + str(cb - work_b - integration_margin) + ", which is the discovery" ^0
"" ^0
"Planning against the dependency is what coordination is, and the two dates" ^0
"are consistent to the week. Consistency between two copies is not evidence" ^0
"about the work, and the work is what the dates are about." ^0Python (deterministic transpilation)
pythonwork_a = 14
work_b = 9
integration_margin = 2
week = 0
date_a = 16
date_b = 18
print("real work remaining : team A " + str(work_a) + " weeks, team B " + str(work_b) + " weeks")
print("announced dates : team A week " + str(date_a) + ", team B week " + str(date_b))
print("integration margin : " + str(integration_margin) + " weeks")
print("")
print("what each announced date was derived from")
print(" team A : team B's date " + str(date_b) + " minus the margin " + str(integration_margin) + " = " + str(date_b - integration_margin))
print(" team B : team A's date " + str(date_a) + " plus the margin " + str(integration_margin) + " = " + str(date_a + integration_margin))
if date_a == date_b - integration_margin:
if date_b == date_a + integration_margin:
print(" the pair is consistent, and each date's only support is the other")
print("")
print("what the work implies")
print(" team A : " + str(work_a) + " weeks of work, so week " + str(work_a))
print(" team B : " + str(work_b) + " weeks of work, so week " + str(work_b))
print(" the binding one is team A at week " + str(work_a))
if date_a > work_a:
print(" the announced date for A is " + str(date_a - work_a) + " weeks later than its own work needs")
if date_b > work_b + integration_margin:
print(" and B's is " + str(date_b - work_b - integration_margin) + " weeks later than B's work plus the margin")
print("")
discovery = 3
work_b2 = work_b + discovery
print("team B discovers " + str(discovery) + " more weeks of work, taking it to " + str(work_b2))
print("")
print("if each side re-derives from the other, one round at a time")
a = date_a
b = date_b
r = 0
while r < 4:
r = r + 1
old_b = b
b = a + integration_margin
if work_b2 + integration_margin > b:
b = work_b2 + integration_margin
a = b - integration_margin
if work_a > a:
a = work_a
print(" round " + str(r) + " : A week " + str(a) + ", B week " + str(b))
print("")
if a == date_a:
if b == date_b:
print(" neither date moved. " + str(discovery) + " weeks of work appeared and the pair is")
print(" unchanged, because each date is supported by the other and the")
print(" margin between them still holds")
print("")
print("where it comes to rest")
print(" team A : week " + str(a) + ", against " + str(work_a) + " weeks of work")
print(" team B : week " + str(b) + ", against " + str(work_b2) + " weeks of work plus " + str(integration_margin))
if a > work_a:
print(" A is still " + str(a - work_a) + " weeks past what its own work needs, because it is")
print(" anchored to B and B is anchored to the margin")
print("")
wa = work_a
wb = work_b2 + integration_margin
joint = 0
if wa > wb:
joint = wa
else:
joint = wb
print("if both dates were derived from estimates and reconciled once")
print(" team A : week " + str(wa) + ", team B : week " + str(wb))
print(" the pair ships at week " + str(joint))
if b > joint:
print(" against week " + str(b) + " under the mutual anchoring, a difference of " + str(b - joint))
print("")
print("what a reader can conclude from the two announced dates")
print(" that the teams agree : yes")
print(" that either date is achievable : nothing")
print(" the agreement was produced by copying, which is also what agreement")
print(" produced by evidence looks like from outside")
print("")
ca = work_a
cb = work_b2 + integration_margin
cjoint = 0
if ca > cb:
cjoint = ca
else:
cjoint = cb
print("control - team A dates from its own estimate, B aligns to A")
print(" before the discovery : A week " + str(work_a) + ", B week " + str(work_b + integration_margin))
print(" after : A week " + str(ca) + ", B week " + str(cb))
print(" the discovery moved the pair by " + str(cb - work_b - integration_margin) + ", which is the discovery")
print("")
print("Planning against the dependency is what coordination is, and the two dates")
print("are consistent to the week. Consistency between two copies is not evidence")
print("about the work, and the work is what the dates are about.")stdout (executed)
textreal work remaining : team A 14 weeks, team B 9 weeks
announced dates : team A week 16, team B week 18
integration margin : 2 weeks
what each announced date was derived from
team A : team B's date 18 minus the margin 2 = 16
team B : team A's date 16 plus the margin 2 = 18
the pair is consistent, and each date's only support is the other
what the work implies
team A : 14 weeks of work, so week 14
team B : 9 weeks of work, so week 9
the binding one is team A at week 14
the announced date for A is 2 weeks later than its own work needs
and B's is 7 weeks later than B's work plus the margin
team B discovers 3 more weeks of work, taking it to 12
if each side re-derives from the other, one round at a time
round 1 : A week 16, B week 18
round 2 : A week 16, B week 18
round 3 : A week 16, B week 18
round 4 : A week 16, B week 18
neither date moved. 3 weeks of work appeared and the pair is
unchanged, because each date is supported by the other and the
margin between them still holds
where it comes to rest
team A : week 16, against 14 weeks of work
team B : week 18, against 12 weeks of work plus 2
A is still 2 weeks past what its own work needs, because it is
anchored to B and B is anchored to the margin
if both dates were derived from estimates and reconciled once
team A : week 14, team B : week 14
the pair ships at week 14
against week 18 under the mutual anchoring, a difference of 4
what a reader can conclude from the two announced dates
that the teams agree : yes
that either date is achievable : nothing
the agreement was produced by copying, which is also what agreement
produced by evidence looks like from outside
control - team A dates from its own estimate, B aligns to A
before the discovery : A week 14, B week 11
after : A week 14, B week 14
the discovery moved the pair by 3, which is the discovery
Planning against the dependency is what coordination is, and the two dates
are consistent to the week. Consistency between two copies is not evidence
about the work, and the work is what the dates are about.Trace event types
eml:run:starteml:assigneml:outputeml:run:done