# Example 770 — The on time rate was over the parcels that arrived `the_on_time_rate_was_over_the_parcels_that_arrived.eml` - On-time delivery has been measured the same way for five years, from a scan taken at the customer's door, against a promise the customers themselves agreed. What a delivery is is computed below. ## EML ```eml # Self-authored for the EML case corpus (no external origin). On-time delivery # has been measured the same way for five years, from a scan taken at the # customer's door, against a promise the customers themselves agreed. What a # delivery is is computed below. # # The measurement is honest. The timestamp comes from the handheld scanned at # the door rather than the driver's phone at the kerb; the promise is in the # customer's local time, not the depot's; the definition was written with the # customers and has not been changed to flatter a quarter; and it is audited # quarterly against the carrier's own records. # # A delivery exists when a parcel is scanned as delivered. A parcel that is lost # is not a late delivery; it is not a delivery. 2460000 => parcels_accepted_a_month 2431000 => parcels_scanned_as_delivered 2392000 => deliveries_inside_the_promise 4100 => parcels_lost_or_destroyed 18600 => parcels_returned_to_sender 6300 => parcels_still_in_the_network_past_the_promise 5 => years_measured_the_same_way 20 => quarterly_audits_passed 0 => parcels_without_a_delivery_scan_in_the_denominator parcels_scanned_as_delivered - deliveries_inside_the_promise => deliveries_outside_the_promise parcels_accepted_a_month - parcels_scanned_as_delivered => parcels_that_never_became_a_delivery parcels_lost_or_destroyed + parcels_returned_to_sender => parcels_accounted_for_without_arriving parcels_accounted_for_without_arriving + parcels_still_in_the_network_past_the_promise => parcels_absent_from_the_denominator int(deliveries_inside_the_promise * 10000 / parcels_scanned_as_delivered) => on_time_over_deliveries_myriad int(deliveries_inside_the_promise * 10000 / parcels_accepted_a_month) => on_time_over_acceptances_myriad on_time_over_deliveries_myriad - on_time_over_acceptances_myriad => the_distance_between_the_two_myriad "parcels accepted a month : " + str(parcels_accepted_a_month) ^0 " scanned as delivered : " + str(parcels_scanned_as_delivered) ^0 " inside the promise : " + str(deliveries_inside_the_promise) ^0 " outside it : " + str(deliveries_outside_the_promise) ^0 "years measured the same way : " + str(years_measured_the_same_way) ^0 "quarterly audits passed : " + str(quarterly_audits_passed) ^0 "" ^0 "on time, over deliveries : " + str(on_time_over_deliveries_myriad) + " per ten thousand" ^0 "on time, over acceptances : " + str(on_time_over_acceptances_myriad) + " per ten thousand" ^0 " distance between the two : " + str(the_distance_between_the_two_myriad) + " per ten thousand" ^0 "" ^0 "parcels that never became a delivery : " + str(parcels_that_never_became_a_delivery) ^0 " lost or destroyed : " + str(parcels_lost_or_destroyed) ^0 " returned to sender : " + str(parcels_returned_to_sender) ^0 " still in the network : " + str(parcels_still_in_the_network_past_the_promise) ^0 " in the denominator : " + str(parcels_without_a_delivery_scan_in_the_denominator) ^0 "" ^0 # ---- what the measurement verified ---- "the on-time measurement" ^0 " timestamp : the handheld scanned at the door, not the" ^0 " driver's phone at the kerb" ^0 " clock : the customer's local time, not the depot's" ^0 " definition : written with the customers, unchanged for" ^0 " " + str(years_measured_the_same_way) + " years" ^0 " audits against the carrier's records : " + str(quarterly_audits_passed) + ", all passed" ^0 " verdict : ON TIME" ^0 "" ^0 " scanning at the door rather than the kerb is the part" ^0 " almost nobody does, and it is why the " + str(deliveries_outside_the_promise) ^0 " late deliveries are counted at all" ^0 "" ^0 # ---- why the ratio has a floor ---- "what can appear in each place" ^0 " numerator : parcels scanned as delivered, inside the" ^0 " promise" ^0 " denominator : parcels scanned as delivered" ^0 " what creates a row : the delivery scan" ^0 " what a lost parcel contributes : nothing; it is not a" ^0 " late delivery, it is not a delivery" ^0 " parcels in that state : " + str(parcels_absent_from_the_denominator) ^0 "" ^0 " the worst outcome the operation has is the one outcome" ^0 " the ratio cannot express" ^0 "" ^0 # ---- what the customer of a lost parcel sees ---- "a parcel that never arrived" ^0 " was it accepted : yes, and paid for" ^0 " was it late : the question does not apply; there is no" ^0 " delivery to be late" ^0 " does it appear in the on-time figure : no" ^0 " does it appear anywhere : in claims, " + str(parcels_lost_or_destroyed) + " of them" ^0 " do the two numbers meet on a page : no" ^0 "" ^0 # ---- null control ---- # The same operation, with the denominator changed to parcels accepted and every # undelivered parcel counted as outside the promise. 2460000 => nc_denominator_parcels 2392000 => nc_deliveries_inside_the_promise 20 => nc_quarterly_audits_passed "null control - measure against what was accepted" ^0 " audits passed : " + str(nc_quarterly_audits_passed) + ", unchanged" ^0 " denominator : " + str(nc_denominator_parcels) + " parcels accepted" ^0 " inside the promise : " + str(nc_deliveries_inside_the_promise) ^0 " on time : " + str(on_time_over_acceptances_myriad) + " per ten thousand" ^0 " nothing about the deliveries changed; the parcels that" ^0 " were never delivered were given a place to appear" ^0 "" ^0 # ---- the rule ---- "what a high on-time rate guarantees" ^0 " a parcel that arrived, arrived on time : exactly," ^0 " " + str(on_time_over_deliveries_myriad) + " per ten thousand of them, scanned at the door" ^0 " a parcel arrived : not addressed; a parcel that did" ^0 " not arrive never enters the ratio, in either place" ^0 "" ^0 "a rate whose denominator is created by the event in its" ^0 "numerator has a floor built into it; the failures that" ^0 "matter most are the ones that never make a row" ^0 "" ^0 "The timestamp is a door scan on the customer's clock, the definition was" ^0 "agreed with them and has stood for " + str(years_measured_the_same_way) + " years, and " + str(quarterly_audits_passed) + " audits have passed." ^0 "A delivery exists when a parcel is scanned, so " + str(parcels_absent_from_the_denominator) + " parcels a month are in" ^0 "neither place and the figure reads " + str(on_time_over_deliveries_myriad) + " per ten thousand against " + str(on_time_over_acceptances_myriad) ^0 "over what was accepted - " + str(the_distance_between_the_two_myriad) + " apart." ^0 ``` ## Python (deterministic transpilation) ```python parcels_accepted_a_month = 2460000 parcels_scanned_as_delivered = 2431000 deliveries_inside_the_promise = 2392000 parcels_lost_or_destroyed = 4100 parcels_returned_to_sender = 18600 parcels_still_in_the_network_past_the_promise = 6300 years_measured_the_same_way = 5 quarterly_audits_passed = 20 parcels_without_a_delivery_scan_in_the_denominator = 0 deliveries_outside_the_promise = parcels_scanned_as_delivered - deliveries_inside_the_promise parcels_that_never_became_a_delivery = parcels_accepted_a_month - parcels_scanned_as_delivered parcels_accounted_for_without_arriving = parcels_lost_or_destroyed + parcels_returned_to_sender parcels_absent_from_the_denominator = parcels_accounted_for_without_arriving + parcels_still_in_the_network_past_the_promise on_time_over_deliveries_myriad = int(deliveries_inside_the_promise * 10000 / parcels_scanned_as_delivered) on_time_over_acceptances_myriad = int(deliveries_inside_the_promise * 10000 / parcels_accepted_a_month) the_distance_between_the_two_myriad = on_time_over_deliveries_myriad - on_time_over_acceptances_myriad print("parcels accepted a month : " + str(parcels_accepted_a_month)) print(" scanned as delivered : " + str(parcels_scanned_as_delivered)) print(" inside the promise : " + str(deliveries_inside_the_promise)) print(" outside it : " + str(deliveries_outside_the_promise)) print("years measured the same way : " + str(years_measured_the_same_way)) print("quarterly audits passed : " + str(quarterly_audits_passed)) print("") print("on time, over deliveries : " + str(on_time_over_deliveries_myriad) + " per ten thousand") print("on time, over acceptances : " + str(on_time_over_acceptances_myriad) + " per ten thousand") print(" distance between the two : " + str(the_distance_between_the_two_myriad) + " per ten thousand") print("") print("parcels that never became a delivery : " + str(parcels_that_never_became_a_delivery)) print(" lost or destroyed : " + str(parcels_lost_or_destroyed)) print(" returned to sender : " + str(parcels_returned_to_sender)) print(" still in the network : " + str(parcels_still_in_the_network_past_the_promise)) print(" in the denominator : " + str(parcels_without_a_delivery_scan_in_the_denominator)) print("") print("the on-time measurement") print(" timestamp : the handheld scanned at the door, not the") print(" driver's phone at the kerb") print(" clock : the customer's local time, not the depot's") print(" definition : written with the customers, unchanged for") print(" " + str(years_measured_the_same_way) + " years") print(" audits against the carrier's records : " + str(quarterly_audits_passed) + ", all passed") print(" verdict : ON TIME") print("") print(" scanning at the door rather than the kerb is the part") print(" almost nobody does, and it is why the " + str(deliveries_outside_the_promise)) print(" late deliveries are counted at all") print("") print("what can appear in each place") print(" numerator : parcels scanned as delivered, inside the") print(" promise") print(" denominator : parcels scanned as delivered") print(" what creates a row : the delivery scan") print(" what a lost parcel contributes : nothing; it is not a") print(" late delivery, it is not a delivery") print(" parcels in that state : " + str(parcels_absent_from_the_denominator)) print("") print(" the worst outcome the operation has is the one outcome") print(" the ratio cannot express") print("") print("a parcel that never arrived") print(" was it accepted : yes, and paid for") print(" was it late : the question does not apply; there is no") print(" delivery to be late") print(" does it appear in the on-time figure : no") print(" does it appear anywhere : in claims, " + str(parcels_lost_or_destroyed) + " of them") print(" do the two numbers meet on a page : no") print("") nc_denominator_parcels = 2460000 nc_deliveries_inside_the_promise = 2392000 nc_quarterly_audits_passed = 20 print("null control - measure against what was accepted") print(" audits passed : " + str(nc_quarterly_audits_passed) + ", unchanged") print(" denominator : " + str(nc_denominator_parcels) + " parcels accepted") print(" inside the promise : " + str(nc_deliveries_inside_the_promise)) print(" on time : " + str(on_time_over_acceptances_myriad) + " per ten thousand") print(" nothing about the deliveries changed; the parcels that") print(" were never delivered were given a place to appear") print("") print("what a high on-time rate guarantees") print(" a parcel that arrived, arrived on time : exactly,") print(" " + str(on_time_over_deliveries_myriad) + " per ten thousand of them, scanned at the door") print(" a parcel arrived : not addressed; a parcel that did") print(" not arrive never enters the ratio, in either place") print("") print("a rate whose denominator is created by the event in its") print("numerator has a floor built into it; the failures that") print("matter most are the ones that never make a row") print("") print("The timestamp is a door scan on the customer's clock, the definition was") print("agreed with them and has stood for " + str(years_measured_the_same_way) + " years, and " + str(quarterly_audits_passed) + " audits have passed.") print("A delivery exists when a parcel is scanned, so " + str(parcels_absent_from_the_denominator) + " parcels a month are in") print("neither place and the figure reads " + str(on_time_over_deliveries_myriad) + " per ten thousand against " + str(on_time_over_acceptances_myriad)) print("over what was accepted - " + str(the_distance_between_the_two_myriad) + " apart.") ``` ## stdout (executed) ```text parcels accepted a month : 2460000 scanned as delivered : 2431000 inside the promise : 2392000 outside it : 39000 years measured the same way : 5 quarterly audits passed : 20 on time, over deliveries : 9839 per ten thousand on time, over acceptances : 9723 per ten thousand distance between the two : 116 per ten thousand parcels that never became a delivery : 29000 lost or destroyed : 4100 returned to sender : 18600 still in the network : 6300 in the denominator : 0 the on-time measurement timestamp : the handheld scanned at the door, not the driver's phone at the kerb clock : the customer's local time, not the depot's definition : written with the customers, unchanged for 5 years audits against the carrier's records : 20, all passed verdict : ON TIME scanning at the door rather than the kerb is the part almost nobody does, and it is why the 39000 late deliveries are counted at all what can appear in each place numerator : parcels scanned as delivered, inside the promise denominator : parcels scanned as delivered what creates a row : the delivery scan what a lost parcel contributes : nothing; it is not a late delivery, it is not a delivery parcels in that state : 29000 the worst outcome the operation has is the one outcome the ratio cannot express a parcel that never arrived was it accepted : yes, and paid for was it late : the question does not apply; there is no delivery to be late does it appear in the on-time figure : no does it appear anywhere : in claims, 4100 of them do the two numbers meet on a page : no null control - measure against what was accepted audits passed : 20, unchanged denominator : 2460000 parcels accepted inside the promise : 2392000 on time : 9723 per ten thousand nothing about the deliveries changed; the parcels that were never delivered were given a place to appear what a high on-time rate guarantees a parcel that arrived, arrived on time : exactly, 9839 per ten thousand of them, scanned at the door a parcel arrived : not addressed; a parcel that did not arrive never enters the ratio, in either place a rate whose denominator is created by the event in its numerator has a floor built into it; the failures that matter most are the ones that never make a row The timestamp is a door scan on the customer's clock, the definition was agreed with them and has stood for 5 years, and 20 audits have passed. A delivery exists when a parcel is scanned, so 29000 parcels a month are in neither place and the figure reads 9839 per ten thousand against 9723 over what was accepted - 116 apart. ``` ## Round-trip `ok: true` — round-trip fixpoint reached (python1 == python2) ## Trace event types eml:run:start · eml:assign · eml:output · eml:run:done