Generating variational query from relational queries:

[pataei]

• fix the logical issue: we cannot just push down the choices as we want, instead we need to have optimization rules which are the combination of choice calculus optimization rules and relational algebra optimization rules. I have some in my thesis but they need to be double checked! Note that not every combination of relational algebra and CC optimizaiton rules hold! There's a small list in our pvldb'19 paper, under file FCCcombinedRELALGrules.tex!
• test it with the employee database
• could we incorporate the type system with optimization rules? eg.: could we have an opt rule that does the following: pres cond of A = F, pres cond of B = True, F<prj_{A, B} R, prj_B R> == prj_{A, B} R
  Ask Eric

[pataei]

meeting result (2/20/19):
we can incorporate the vschema into the optimization rules and not the type system. This makes opt rules interesting. and instead of proving that opt rules are semantic preserving we can now prove that they're type preserving which is easier to do although it's a more relaxed (less promising) property.