From Considerable to Executable
At 14:00 on Saturday, Gu Ning spread the pages from the first five chapters across the table. The D17 action question card stated what was being chosen; the goal hierarchy stated what was pursued and protected; the consequence table showed how resources could change; the scenario-weighting page kept probability fields open; and the survival-condition card excluded commitments that crossed the floor.
These pages made a limited clarification eligible for comparison, but no action had yet occurred. If Tang Ke sent a message from memory, she might omit a question, expand the commitment, or leave no stopping point. If Ye Cheng read “continue contact” and reserved a bay, a low-commitment candidate would be executed as another action.
Provisional adoption condenses distributed judgments into a versioned authorization. The card must tell an executor what to do, the furthest extent permitted, when to end, and which material reopens judgment. “Handle flexibly” cannot throw critical choices back into the execution scene.
As of 14:00, external facts are unchanged. The client has sent no new message; Version 2 does not exist; and the station has neither sent a clarification, reserved a bay, bought materials, nor altered the six existing jobs.
Adoption Does Not Mean the Prediction Is Complete
D17 lacks an event definition, price, scope, and outcome probabilities for a new version. If action had to wait for every number, the station could not even take the first step of asking for conditions. Actionability does not always require a complete prediction. It can arise from adequate control of an action's boundary and adverse scenario.
The purpose of clarification is to obtain missing conditions. If the message has a cost ceiling and full failure harms neither the six-slot fulfillment nor the eight-unit protection, it may be adoptable while probability remains unknown. The reason comes from an information task and limited commitment, not an assumption of positive expectation.
Gu Ning therefore adds no “success probability” to the card. She marks the field unknown and states that Part 1's 0.5 belongs only to the closed R17 Version 1. Unknown remains unknown after adoption; adoption says only that, under present material, the bounded action is worth executing.
First Eliminate Deep Commitments Without Present Grounds
Advance purchasing needs specifications or reuse conditions; neither exists. Deep preparation needs a describable version; none exists. Reserving a bay may not breach six-slot fulfillment, but it occupies part of four-slot new capacity and needs a reservation object, release time, and missed-opportunity consequence.
By contrast, one time-bounded clarification can ask for material needed to form a subsequent object without occupying a half-day slot or using working capital. No current evidence shows that deep commitment adds a goal benefit, so it does not enter the first action version.
This does not declare reservation, purchasing, or preparation forever wrong. If the client later supplies clear scope, timing, and transaction terms, a new plan can compare them again. Saturday eliminates versions lacking present grounds; it does not erase their action names from every future candidate set.
Why No Contact and Passive Waiting Are Not Adopted for Now
Ending active contact uses the least added time and preserves all resources. But Friday's proposal to talk next week did not provide the scope, window, and terms needed to form a new version. Immediate closure would surrender one low-cost, explicit inquiry.
Simply waiting for complete conditions might also obtain material, but without a time point it cannot be distinguished from forgetting. Even with a deadline, waiting leaves the content of the questions for the client to guess and may not supply the four conditions the station actually needs.
Gu Ning does not claim that active contact has a higher response rate. She compares whether each action addresses the present unknown and whether added commitment is bounded. One structured clarification directly serves the information goal without changing bay or cash states, giving it a clearer present rationale.
One Message Completes One Task
Tang Ke initially wanted the message to describe how the station could prepare and to ask whether the client would pay part of the fee in advance. Gu Ning removed both. The first might be understood as accepting the scope; the second lacks a transaction object before a version forms.
D17-A0's task is limited to obtaining conditions sufficient to decide whether to create a next-version action object, or confirming at the deadline that those conditions were not obtained. It does not win the commission, prove the relationship, or schedule production.
The narrower the task, the clearer its adjudication. An incomplete reply can identify missing items, and no reply can close the action on time. If one message combines marketing, quotation, scheduling, and relationship maintenance, every response can be called partial success and the action may never truly expire.
Four Questions Form the Minimum Interface
First: does the client wish to continue discussing a new version? This does not assume Friday's “talk next week” already formed a commission; it confirms whether the intention remains.
Second: if so, what object and scope are proposed? The station must know what is changing and cannot assemble Version 2 from Version 1 by itself. Third: what receipt or delivery window is desired? Capacity can then be checked without reserving it first.
Fourth: how are fee, payment, and acceptance conditions to be arranged? This does not require immediate acceptance of the station's quotation; it prevents scope and time from advancing while transaction conditions remain blank. Together the four questions form an interface for later modeling. Missing any one means a complete version has not yet formed.
The Message Does Not Answer for the Client
Tang Ke's draft uses conditional language: “If you wish to continue discussing a new version, please provide...” followed by the four questions. It does not say that time has been reserved, because no bay is reserved, or that materials can be prepared immediately, because neither specifications nor purchasing has been decided.
The message carries forward Friday's record and states that Version 1's original receipt period has ended and the present inquiry concerns a possible new version. This prevents the clarification from being understood as extending the original window and prevents the station from later backdating a reply as late occurrence of R17 Version 1.
The closing states that the station will arrange subsequent discussion at 12:00 Monday based on information received. This is the adjudication point for the current step, not a promise to preserve production capacity until then. A later reply can be retained but must reopen at its new time; it cannot become a complete in-window result for A0.
Translating a Communication Ceiling into Execution
The station schedules primary production in half-day slots, but one message need not be fictionalized as using a full half day. Gu Ning sets a separate communication ceiling for this step: checking the old record, finalizing, sending, and preserving receipt together may use no more than one hour.
One hour is D17-A0's action budget, not a universal resource rule or a claim that communication has no opportunity cost. If finalization requires a longer discussion, Tang Ke must stop and report rather than expand it because “only a little remains.” Exceeding the ceiling requires a new A0 version.
This step occupies none of the four available half-day slots, alters none of the six existing tasks, and uses none of the twelve working-capital units. It does not presume overtime; the 09:00 Monday period comes from Tang Ke's authorized routine communications, and Ye Cheng need not move a task.
Aligning Roles and Authority
Tang Ke is the executor. She sends the approved message once and preserves time, text, channel receipt, and original reply. She may fix an obvious typo, but cannot add a price, promise scheduling or purchasing, or record oral interest as a formed version.
Gu Ning owns adoption and reopening. She approves the message boundary on Saturday and adjudicates A0 from the record at 12:00 Monday. If complete conditions requiring immediate handling arrive earlier, she may convene a reopening, but Tang Ke cannot commit resources on the assumption that “Gu Ning will probably approve.”
Ye Cheng does not send. At reopening, he checks whether the six existing jobs or four available slots have changed. The division gives communication, resources, and adjudication separate records and keeps everything from falling into an ambiguous “we.”
The Channel Is Part of the Action Definition
The message will be sent only through the same traceable text channel already used by both parties. Tang Ke will not also call, use several accounts, or ask an intermediary, because multiple channels change the degree of disturbance and make it harder to identify which action triggered a reply.
If the original channel is unavailable at 09:00 Monday, A0 does not switch automatically. Tang Ke records the failure and sends it to Gu Ning for reopening. A backup channel may be reasonable, but requires a fresh check of authority, traceability, and the client's agreement to its use.
The restriction does not seek laboratory purity. It aligns actual action with the card. The channel affects consequences and the reply record and therefore cannot be altered freely as an irrelevant implementation detail.
What “Send Once” Means
“Once” means sending at 09:00 Monday one approved message containing all four questions. If the platform renders a long message as several consecutive bubbles, it remains one message group if sent from the original draft in the same period, and the complete text must be preserved.
If the client does not respond promptly, Tang Ke does not add “Did you see this?” or call. An automatic second follow-up would turn one bounded clarification into continued pursuit and increase time and relationship consequences.
If the client asks for explanation of a question, Tang Ke may preserve the request but cannot negotiate scope under A0's authority. Gu Ning can reopen on the concrete question and decide whether to authorize a new communication step. Incoming feedback does not grant the executor unlimited authority to respond.
When the Action Begins and Ends
A0 is planned for 09:00 Monday. When the sending receipt exists, the action moves from adopted to executed and awaiting result. Without a receipt, a finished draft cannot be recorded as successful execution.
The observation window ends at 12:00 Monday. Information received before then enters A0's result page verbatim. If no complete reply exists at noon, A0 expires and no longer waits to answer current scheduling. Late material may trigger a new judgment but does not rewrite the in-window result.
Ending does not delete the contact. After expiry, the station returns to a state with no added resource commitment and continues the six existing jobs and other arrangements. Because this step reserved nothing, bought nothing, and changed no schedule, recovery requires no bay release or refund. Communication time already used, up to one hour, remains recorded.
Four Result Classes Avoid a Simple Success/Failure Split
The first class is receipt of all four conditions within the window. It is sufficient only to reopen and decide whether to form V2, not to accept a commission. The second is a reply with one or more conditions missing. Each omission is preserved; a response does not equal completeness.
The third is an explicit decision not to continue. This closes the present proposal for a new version while completing the information task. It is adverse for cooperation but useful for avoiding unsupported commitment. The fourth is no reply within the window, recorded as material not obtained without inferring motive.
The message may also fail to send. That is an execution or channel result, not a refusal by the client. Tang Ke preserves the platform receipt and Gu Ning reopens the channel issue. Once the classes are separate, an ambiguous “there is news” cannot automatically trigger later action.
A Complete Reply Does Not Automatically Form V2
Four complete pieces of information are material from which a new object might be defined. Gu Ning still checks whether scope can be adjudicated, the time window is clear, the transaction terms have mutually recognizable meanings, and capacity remains as it was Saturday.
If these checks pass, the team may create V2 and preserve its formation time. V2's prediction question, action options, and resource state begin with the new version. It inherits neither R17 Version 1's 0.5 nor the treatment of D17-A0's information task as commission success.
If an apparently complete reply contains conflicting conditions—for example, scope and delivery cannot both be met—V2 may remain suspended. Obtaining text and forming an actionable object are separate steps. A complete reply triggers reopening; it does not replace the conclusion of that review.
A Partial Reply Does Not Automatically Trigger a Second Question
If the client says only “we want to continue,” without scope, time, or transaction conditions, Tang Ke may feel compelled to ask immediately. A0 instead requires recording omissions until noon without automatically sending another message. This preserves the boundary of one action rather than rejecting communication.
At reopening, Gu Ning may compare a supplemental question, waiting for completion, or closing this round. The new action has its own question, ceiling, and validity. If every partial reply is absorbed into old authority, one clarification can extend indefinitely and its initial cost ceiling fails.
If the client voluntarily adds information in successive messages, Tang Ke preserves it. Refraining from asking does not mean refusing to receive. Adjudication considers the complete set and timing received before noon, not whether every answer appeared in one message.
What No Reply Does Not Mean
No reply by noon Monday means only that A0 did not obtain all four items within its window. It does not prove the client will never cooperate, that the wording was ineffective, or that every earlier understanding of the relationship was wrong.
It also does not reopen Version 1. R17's original window and adjudication remain. A0 is a later information action whose result returns to its own object and deadline.
No reply is not costless. Checking and sending consumed bounded communication time and may leave an entrance for a later reply. These consequences enter the record without invented psychological causes. A next step requires new judgment, not a story about silence that automatically authorizes pursuit.
Why Expected Value Is Not Calculated Now
Calculating A0's numerical expectation requires result probabilities and comparable values. There is no transferable action-conditional probability and no reliable conversion of information, relationship, and communication time into one unit. Forced calculation would merely dress unknowns in decimals.
A0 still has comparative grounds. Against waiting, it actively asks four critical unknowns. Against reservation and purchasing, it uses fewer resources in identified scenarios while obtaining no less of that information. Against permanent closure, it preserves one explicit, low-commitment entrance.
The comparison depends on coverage of the present outcome set. If sending itself has an unlisted material consequence, the plan reopens. Gu Ning provisionally adopts A0 without claiming that it dominates in every possible world or that low cost equals zero risk.
Assumptions Included in Provisional Adoption
First, the original text channel remains usable and supplies adequate sending records. Second, the one-hour ceiling covers checking, sending, and preservation without moving existing jobs. Third, the four questions are sufficient for deciding whether a new version should be defined.
Fourth, one courteous clarification does not itself form a production commitment. Wording, absence of reservation, and authority limits partially secure this assumption, but the client's interpretation may challenge it. If the client explicitly treats the message as a commitment, the action reopens immediately rather than arguing from the card's intent.
With assumptions disclosed, adopters know which changed facts invalidate the plan. The assumptions do not prove inevitable correctness; they are connections on which the present version depends.
What Immediately Triggers Reopening
Before 12:00 Monday, any of the following reopens A0: the original channel is unavailable; the message needs more than an hour; the six existing jobs or eight protected units change; the client sends all four conditions; the client requests an immediate commitment of scope, schedule, price, or purchasing; or an important unlisted consequence appears.
Reopening neither stops everything automatically nor continues it. Current authority cannot alone handle the new situation, so the triggering material must be preserved and a next-version judgment formed. If triggered before sending, Tang Ke pauses. If triggered after sending, she adds no commitment and preserves the record.
Ordinary listed results need not cause emergency reopening. Partial or absent replies can wait for noon adjudication unless they add a resource or commitment request. Triggers identify the model boundary; they do not make action impossible in response to every detail.
Advice, Adoption, Execution, and Result Use Four Columns
Tang Ke's proposal to send a clarification is advice. Gu Ning's approval of D17-A0 from the goal and survival cards is adoption. Execution occurs only when Tang Ke sends the approved text Monday and obtains a receipt. The client's reply and actual time consumed are results.
Four columns prevent common backward flows. Approval is not sending; sending is not agreement; a favorable reply does not prove every Saturday reason correct. Conversely, refusal does not automatically prove adoption mistaken; A0 must be assessed on whether it completed the information task at bounded cost.
If execution deviates—for example, by adding a second follow-up—result review records the deviation. Consequences created by deviation cannot all be assigned to the original plan, and a good endpoint cannot erase the deviation.
The Action Card Is a Retrievable Prospective Locking
In RC, an action card is a subject's limited preconstruction of a potential reality. It does not exhaust the future. Under current material it provisionally fixes a low-loss path while keeping refusal, silence, and changed conditions available to later observation. Adoption marks where this round of convergence begins; result classes preserve interfaces for reconvergence.
This realizes processual completeness at the scale of action. A plan is complete not because it guarantees a good reply, but because expectation, execution, external feedback, and revision can remain continuous. D17-A0 is both an action and a secondary construction that a later round of observation can take over.
D17-A0 Action Card, Version 1
Object: obtain the minimum defining conditions for a possible new version. Action: at 09:00 Monday, Tang Ke sends one message group through the existing traceable text channel, asking about intention to continue, object and scope, receipt or delivery window, and fee, payment, and acceptance terms.
Resource ceiling: checking, sending, and archiving take no more than one hour; zero labor slots and zero working-capital units are committed; no reservation, purchasing, or change to the six existing jobs. Authority: Tang Ke executes approved text and preserves originals, Gu Ning owns adoption and reopening, and Ye Cheng checks capacity only upon reopening.
Validity: planned execution at 09:00 Monday and end of observation at 12:00. Complete reply, partial reply, explicit discontinuation, no reply, and channel failure are recorded separately. A complete reply triggers only a V2 definition check; a partial reply does not authorize another question; no reply carries no inferred motive; late material is handled at its new time.
Survival check: the action touches neither the eight-unit protection nor six-slot fulfillment, leaves no resources to release on exit, and consumes unrecoverable communication time only within its ceiling. No automatic upgrade: A0 authorizes no quotation, reservation, purchasing, production preparation, or acceptance.
Why A0 Is Not the Final Plan
The zero in A0 marks its position before resource commitment, where it primarily forms an action object. It is not a universal project-naming rule and does not mean the action has no cost. Up to one hour of communication, relationship effects, and attention are real consequences.
If V2 forms, later stages may arrange confirmation, sampling, or formal input according to its scope, each with its own number, resources, and exit conditions. If it does not form, A0 can expire without demanding continuation merely because a first step was taken.
The first action card remains updateable while making updates answer to reasons. Provisional adoption is not a vague “try it and see.” It means executing this version within its validity and stopping or reopening beyond it.
Completing Adoption at 14:40 on Saturday
At 14:40 on Saturday, Gu Ning signed D17-A0 Version 1. At this time, an action plan has been adopted and placed in Monday's communication schedule. The message has not been sent, the client has not replied, V2 has not formed, and bays, working capital, and existing tasks are unchanged.
Part I has now connected Part 1's closed probability report to action. The old 0.5 did not become authorization to continue, goals were not compressed into one gain, conditional consequences did not enter actual accounts, expected value did not answer for survival, and the floor did not choose a plan in place of goals.
A0 is reasonable conditionally and for a limited time, depending on the current channel, communication ceiling, resource state, and four-question interface. Monday's execution and reply will test these connections. A good result does not exempt the record from review; a bad result cannot reject the limited probe without analysis.
Part II begins at 09:00 Monday. The next chapter asks whether the message really is a minimum useful step: it must answer a clear question, stay within authorization in execution, and let its result enter another round, rather than using “small” as an excuse for missing goals and adjudication.