{"id":"forward-only-is-the-architectures-load-bearing-constraint","text":"Forward-only design is both the architecture's most universally applied pattern (spanning event streams, task pipelines, and notification delivery as a single processing primitive) and its primary correctness mechanism (containing temporal gaps, preventing regression, enabling safe cost shifting despite accepted weaknesses) — making it the single most load-bearing architectural constraint across all 25 modules.","truth_value":"IN","source":"","source_url":"","source_hash":"","justifications":[{"type":"SL","antecedents":["forward-only-is-universal-processing-primitive","forward-only-preserves-correctness-despite-accepted-gaps"],"outlist":[],"label":"Forward-only is universal in scope AND primary in correctness role — uniquely load-bearing"}],"dependents":["forward-only-is-universally-load-bearing-across-all-layers","write-path-validity-spans-the-full-complexity-spectrum"],"metadata":{},"created_at":"","updated_at":"","reviewed_at":"","verified_at":"","retracted_at":"","explanation":{"steps":[{"node":"forward-only-is-the-architectures-load-bearing-constraint","truth_value":"IN","reason":"SL justification valid","antecedents":["forward-only-is-universal-processing-primitive","forward-only-preserves-correctness-despite-accepted-gaps"],"label":"Forward-only is universal in scope AND primary in correctness role — uniquely load-bearing"},{"node":"forward-only-is-universal-processing-primitive","truth_value":"IN","reason":"SL justification valid","antecedents":["event-and-task-processing-share-forward-only-correctness","notification-instantiates-forward-only-delivery"],"label":"depth-5 beliefs covering streams+tasks and delivery are both unused; combining reveals forward-only spans every processing paradigm, not just data and execution layers"},{"node":"event-and-task-processing-share-forward-only-correctness","truth_value":"IN","reason":"SL justification valid","antecedents":["forward-only-stream-processing-is-exactly-once","control-data-separation-enables-forward-progress"],"label":"two independent processing domains converge on the same no-regression mechanism"},{"node":"forward-only-stream-processing-is-exactly-once","truth_value":"IN","reason":"SL justification valid","antecedents":["dedup-and-finalization-are-coordinated","forward-only-design-prevents-regression-and-maximizes-progress"],"label":"Global dedup keying (event_id alone, not per ad_id) means two legitimate events sharing an ID across different entities are silently merged, breaking per-entity exactly-once semantics","outlist":["dedup-is-global-not-per-ad"]},{"node":"dedup-and-finalization-are-coordinated","truth_value":"IN","reason":"SL justification valid","antecedents":["dedup-outlives-aggregation-window","watermark-finalization-is-irreversible"],"label":"Irreversible finalization demands that dedup outlive the window — these two depth-1 conclusions are not independent but structurally dependent"},{"node":"dedup-outlives-aggregation-window","truth_value":"IN","reason":"SL justification valid","antecedents":["dedup-pruning-uses-2x-lateness","watermark-drives-finalization"],"label":"The 2× multiplier ensures dedup coverage extends beyond the point where corrections would be impossible"},{"node":"dedup-pruning-uses-2x-lateness","truth_value":"IN","reason":"premise"},{"node":"watermark-drives-finalization","truth_value":"IN","reason":"premise"},{"node":"watermark-finalization-is-irreversible","truth_value":"IN","reason":"SL justification valid","antecedents":["watermark-drives-finalization","window-lifecycle-one-directional","no-window-merging-or-retraction"],"label":"Three properties compose into hard irreversibility: watermark-only trigger + one-way lifecycle + no retraction"},{"node":"window-lifecycle-one-directional","truth_value":"IN","reason":"premise"},{"node":"no-window-merging-or-retraction","truth_value":"IN","reason":"premise"},{"node":"forward-only-design-prevents-regression-and-maximizes-progress","truth_value":"IN","reason":"SL justification valid","antecedents":["state-ratchets-prevent-regression-across-domains","pipeline-processing-maximizes-forward-progress"],"label":"State ratchets and forward-progress pipelines are complementary mechanisms — ratchets prevent regression in stateful systems, pipelines prevent regression in dataflow systems"},{"node":"state-ratchets-prevent-regression-across-domains","truth_value":"IN","reason":"SL justification valid","antecedents":["chat-monotonic-read-progress","watermark-finalization-is-irreversible"],"label":"chat read cursors are monotonic (never re-mark as unread), aggregation windows follow OPEN→CLOSED→FINALIZED with no reversal — both are state ratchets that make backwards movement structurally impossible"},{"node":"chat-monotonic-read-progress","truth_value":"IN","reason":"SL justification valid","antecedents":["chat-dual-ordering-sequence-and-lamport","chat-read-cursors-monotonic"],"label":"Sequence-number ordering + monotonic cursors create an irreversible read-progress guarantee"},{"node":"chat-dual-ordering-sequence-and-lamport","truth_value":"IN","reason":"premise"},{"node":"chat-read-cursors-monotonic","truth_value":"IN","reason":"premise"},{"node":"pipeline-processing-maximizes-forward-progress","truth_value":"IN","reason":"SL justification valid","antecedents":["video-pipeline-maximizes-useful-work-on-failure","watermark-finalization-is-irreversible"],"label":"Video pipeline (branch-independent failure containment) and stream processing (irreversible finalization) both embody the same forward-only processing principle"},{"node":"video-pipeline-maximizes-useful-work-on-failure","truth_value":"IN","reason":"SL justification valid","antecedents":["dag-failure-cascade","youtube-pipeline-dag-structure"],"label":"DAG topology + selective cascade means a thumbnail failure doesn't block a successful transcode"},{"node":"dag-failure-cascade","truth_value":"IN","reason":"premise"},{"node":"youtube-pipeline-dag-structure","truth_value":"IN","reason":"premise"},{"node":"control-data-separation-enables-forward-progress","truth_value":"IN","reason":"SL justification valid","antecedents":["video-pipeline-separates-control-from-data-flow","forward-only-design-prevents-regression-and-maximizes-progress"],"label":"decoupling control (DAG) from data (ctx dict) is what makes forward-only failure handling safe in practice"},{"node":"video-pipeline-separates-control-from-data-flow","truth_value":"IN","reason":"SL justification valid","antecedents":["youtube-ctx-dict-blackboard","youtube-pipeline-dag-structure"],"label":"DAG handles ordering and failure; ctx dict handles data — neither constrains the other"},{"node":"youtube-ctx-dict-blackboard","truth_value":"IN","reason":"premise"},{"node":"notification-instantiates-forward-only-delivery","truth_value":"IN","reason":"SL justification valid","antecedents":["notification-delivery-is-bounded","forward-only-extends-to-failure-handling"],"label":"Notification's bounded retry lifecycle is structurally isomorphic to forward-only failure handling"},{"node":"notification-delivery-is-bounded","truth_value":"IN","reason":"SL justification valid","antecedents":["notification-defensive-double-rate-check","retry-escalates-to-permanent-failure"],"label":"complementary boundedness — rate checks bound throughput, retry escalation bounds duration"},{"node":"notification-defensive-double-rate-check","truth_value":"IN","reason":"SL justification valid","antecedents":["notif-two-phase-rate-limit","notif-rate-limit-on-success-only"],"label":"Queue delay between the two check points means neither alone is sufficient; failed retries not consuming budget prevents limit exhaustion from transient failures"},{"node":"notif-two-phase-rate-limit","truth_value":"IN","reason":"premise"},{"node":"notif-rate-limit-on-success-only","truth_value":"IN","reason":"premise"},{"node":"retry-escalates-to-permanent-failure","truth_value":"IN","reason":"SL justification valid","antecedents":["notif-exponential-backoff-with-jitter","retry-converts-timeout-to-failure"],"label":"Backoff prevents cascading load; finite retry count ensures eventual resolution (success or permanent failure) rather than indefinite limbo"},{"node":"notif-exponential-backoff-with-jitter","truth_value":"IN","reason":"premise"},{"node":"retry-converts-timeout-to-failure","truth_value":"IN","reason":"premise"},{"node":"forward-only-extends-to-failure-handling","truth_value":"IN","reason":"SL justification valid","antecedents":["forward-only-design-prevents-regression-and-maximizes-progress","retry-escalates-to-permanent-failure"],"label":"The d3 forward-only node covers normal state progression; the d1 retry node covers failure progression — the emergent property is that both follow the same no-rollback principle"},{"node":"forward-only-preserves-correctness-despite-accepted-gaps","truth_value":"IN","reason":"SL justification valid","antecedents":["temporal-gaps-are-contained-by-forward-only-design","write-read-asymmetry-is-end-to-end-correct"],"label":"Containment strategy — correctness emerges from progress guarantees rather than exhaustive verification, making accepted gaps survivable"},{"node":"write-read-asymmetry-is-end-to-end-correct","truth_value":"IN","reason":"SL justification valid","antecedents":["writes-commit-irrevocably-reads-reconcile","correctness-by-construction-not-validation"],"label":"Writes commit and reads reconcile (depth-4) with structural construction (depth-3), but assumed invariants (depth-2, IN) break the end-to-end guarantee","outlist":["assumed-invariants-are-unenforced"]},{"node":"writes-commit-irrevocably-reads-reconcile","truth_value":"IN","reason":"SL justification valid","antecedents":["writes-are-cheap-reads-pay","forward-only-design-prevents-regression-and-maximizes-progress"],"label":"forward-only semantics explain WHY reads bear the full burden — there is no mechanism to go back"},{"node":"writes-are-cheap-reads-pay","truth_value":"IN","reason":"SL justification valid","antecedents":["write-time-decisions-are-lightweight-but-binding","read-path-absorbs-consistency-and-computation-cost"],"label":"Write-side minimalism and read-side cost absorption are two faces of the same design tradeoff, consistently applied across KV, chat, news feed, autocomplete, and payment"},{"node":"write-time-decisions-are-lightweight-but-binding","truth_value":"IN","reason":"SL justification valid","antecedents":["fan-out-write-pushes-references-not-data","write-time-routing-is-irrevocable"],"label":"Fan-out pushes references (lightweight) and routing decisions are permanent (binding) — the write path optimizes for speed at the cost of flexibility"},{"node":"fan-out-write-pushes-references-not-data","truth_value":"IN","reason":"SL justification valid","antecedents":["chat-fanout-on-write","news-feed-fan-out-write-pushes-ids"],"label":"Reference-based fanout limits write amplification to pointer-sized payloads"},{"node":"chat-fanout-on-write","truth_value":"IN","reason":"premise"},{"node":"news-feed-fan-out-write-pushes-ids","truth_value":"IN","reason":"premise"},{"node":"write-time-routing-is-irrevocable","truth_value":"IN","reason":"SL justification valid","antecedents":["news-feed-celebrity-threshold-at-write-time","chat-fanout-on-write"],"label":"news feed selects fan-out-on-write vs fan-out-on-read based on follower count at publish time; chat routes to inbox or offline queue based on presence at send time — both decisions are baked in at write time and not revisited"},{"node":"news-feed-celebrity-threshold-at-write-time","truth_value":"IN","reason":"premise"},{"node":"read-path-absorbs-consistency-and-computation-cost","truth_value":"IN","reason":"SL justification valid","antecedents":["lazy-read-time-evaluation-trades-write-simplicity-for-read-cost","kv-read-path-is-self-healing"],"label":"lazy eval (autocomplete decay, URL expiration, payment balance) defers computation to reads; KV read repair defers convergence to reads — the combined pattern reveals a systematic bias toward read-path complexity across the repo"},{"node":"lazy-read-time-evaluation-trades-write-simplicity-for-read-cost","truth_value":"IN","reason":"SL justification valid","antecedents":["autocomplete-decay-is-read-time","url-shortener-expiration-lazy","payment-balance-never-cached"],"label":"Three independent systems chose lazy evaluation, suggesting write-simplicity is the dominant concern in pedagogical implementations"},{"node":"autocomplete-decay-is-read-time","truth_value":"IN","reason":"premise"},{"node":"url-shortener-expiration-lazy","truth_value":"IN","reason":"premise"},{"node":"payment-balance-never-cached","truth_value":"IN","reason":"premise"},{"node":"kv-read-path-is-self-healing","truth_value":"IN","reason":"SL justification valid","antecedents":["kv-read-repair-on-get","kv-node-stores-sibling-versions"],"label":"Read repair + sibling detection make the read path an active consistency mechanism, not just a query"},{"node":"kv-read-repair-on-get","truth_value":"IN","reason":"premise"},{"node":"kv-node-stores-sibling-versions","truth_value":"IN","reason":"premise"},{"node":"correctness-by-construction-not-validation","truth_value":"IN","reason":"SL justification valid","antecedents":["structural-discipline-prevents-consistency-bugs","state-ratchets-prevent-regression-across-domains"],"label":"Both depth-2 conclusions prevent bugs structurally (immutability/sync vs. monotonic ratchets) rather than via checks, forming a unified construction-over-validation principle"},{"node":"structural-discipline-prevents-consistency-bugs","truth_value":"IN","reason":"SL justification valid","antecedents":["immutable-values-prevent-aliasing-bugs","multi-structure-sync-invariant"],"label":"immutability (KV vector clocks, leaderboard reinsert) prevents mutation aliasing; multi-structure sync (consistent hashing, leaderboard) prevents index divergence — leaderboard uses BOTH, showing these disciplines are complementary"},{"node":"immutable-values-prevent-aliasing-bugs","truth_value":"IN","reason":"SL justification valid","antecedents":["kv-vector-clock-immutable","leaderboard-update-by-remove-reinsert"],"label":"Immutable-value semantics eliminate shared-reference aliasing at the cost of allocation overhead"},{"node":"kv-vector-clock-immutable","truth_value":"IN","reason":"premise"},{"node":"leaderboard-update-by-remove-reinsert","truth_value":"IN","reason":"premise"},{"node":"multi-structure-sync-invariant","truth_value":"IN","reason":"SL justification valid","antecedents":["ch-triple-bookkeeping","leaderboard-dual-index-consistency"],"label":"Multi-structure sync is a recurring correctness burden where the failure mode is silent divergence"},{"node":"ch-triple-bookkeeping","truth_value":"IN","reason":"premise"},{"node":"leaderboard-dual-index-consistency","truth_value":"IN","reason":"premise"}]}}