Temporality

Thoughts

• maybe the triple format is not naturally up to the task of meta-modelled time

temporal-context

• calendar system
• time-slice granularity
• fluency approach?

genesis-timestamp

• has an associated store
• naturally precise, but could be fuzzed
• universal (even if not recorded)
  • e.g. every keystroke has one, with a text editor (in-memory store)
• context is real-world, digital,
• can be preserved or not across different stores
  • could be used for lineage
    • e.g. on every copy, the data adds a tx-ts+store-identifier
• in RDF, a triple isn't enough (and using the graph in a quad wasteful) for expressing a transaction timestamp
  • so need to add a transaction-timestamp statement
  • but then you might go down an infinitely recursive rabbit hole of tx-timestamping the tx-timestamp statement
    • maybe tx-ts triples are special, having a tx-ts equal to the tx-ts they encode
  • if you use Relator exclusively, you are golden
• should be automated

tombstone-timestamp

• formery "invalidation-timestamp", but needed a new name, because it treads on valid-time

issues

• should be stored outside the relator/data because who's to say when a piece of data is dead?

uses

• for immutable contexts, expresses "not current"
  • can be complementary or an alternative to "Successor"?
  • corrections, like data-entry errors or buggy pipeline

validity

• validity-start, valid-at, invalidity-start, invalid-at

  • only apply to relations

• abstractions shouldn't lose validity

• different kinds of validity

  • "in-world"/"in-application"
  • real-time
  • decision-time/assertion-time replaces transaction time because, for the most part, who cares when it was recorded in a store.
    • mostly, assertion-time would be equal to transaction time, at least in a "live" system

immutability

• might be helpful to differentiate between corrections and updates, but that line is very thin and unstable,
  • e.g. in a role-playing experience platform where in-world facts get bent/corrected and truth is more of a "what we decided worked best in the narrative"
  • i wanted to say abstractions are always true, but what about the case of corrections

Prior Art

References

• 4d Ontology

---

Children

1. Temporality Log