Talk:Use Case Linked Data Timeliness

From XG Provenance Wiki
Jump to: navigation, search

1. Should we talk in the context of the defined problem about “currency”, rather than “timeliness” (according to Bouzeghoub and Peralta “A Framework for Analysis of Data Freshness”, IQIS, 2004)? --User:Aleksey

  • In the use case description it says "... Alice's application assumes b is more recent than c ...". This is indeed more in line with the definition of currency as provided by the cited paper. However, I wonder whether we should better go for the more general term freshness, leaving the actual assessment out of the use case description. It is more important that there were creation times and access times and that they could be taken into account, whatever assessment or estimation method used. --User:OlafHartig

2. In this scenario, what is the mechanism of data delivery: is it push (a source sends data to user) or pull (user queries a source)? In the case of push sources we can assess the timeliness using statistical analysis of updates (and thus, it’s not mandatory for the source to provide provenance-related metadata for such an assessment). Or do we talk about pull sources only in this use case? --User:Aleksey

  • The idea was that Bob and Carol (or their software that creates b and c) retrieve the current version of file x whenever they create b and c. Hence, it is pull. --User:OlafHartig