Roadmap 2020

[szarnyasg]

• introduce valid delete operations – done, see GRADES-NDA paper
• migrate Datagen to Spark
• tune the distribution of deletes (ongoing on the tuning branch, Tuning #235)
  • once this is done, tune the scale factors
• rework paramgen (Rewrite parameter generator in Julia #83)
  • get it scale to SF1k+ for Interactive. As the parameter curation involves a lot of computation (estimating the sizes of intermediate join results), we need to move away from a naïve Python code.
  • add new BI queries (Q16-Q20)
• support generating microbatches of inserts/deletes
• add reaction type (enum) to likes edges WONTFIX

See also proposed refactoring and optimizations in #227

[szarnyasg]

This issue is done except for parameter generation. To be continued in #236