PhD Weekly report
Submitted the survey both to journal and to ArXiv.
Corrected the paper to comply with journal’s anti-plagiarism restrictions.
(from “A deep specification for Dropbox”)
- induce operation events ( E )
- observe outcomes of events ( H=(E,op,rb,ss,ob) )
- add “invisible operation” to justify observation ( vis, ar )
- check whether the added “invisible operations” + the concrete outcomes match a predefined model (maybe defined os FSM) (validate consistency predicates)
Tool required:
- QuickCheck to generate property-based tests
- a key value store (Cassandra?)
- a theorem prover / model checker to validate predicates against executions
So either, using a top-down approach, we embed formal specification into declarative programming languages, or we can use those formal specification in a testing framework based on property-based testing.
Papers read
- review of: FIT paper by Faleiro and Abadi, 2015
- presentation on Chain Replication at Ricon 2015 by Scott Fritchie
- A Deep Specification for Dropbox - Benjamin Pierce - ClojureCon presentation
Plan for next week
Try to investigate the use of QuickCheck: write a first prototype.
Scout theorem provers and formal tools.