Hi Beamers, Beam model understandably gravitates more towards what makes sense to do with data in a more theoretical manner rather than exactly how runner's process it. That also sets the model free from status quo of runners(and gives them a level playing field from semantics view point). But are there any known challenges with the approach? (If yes, I'd like to think they apply to any unification layer in general)
I have one specific example which may or may not fall into above category. What if I want to tune at the level of nuts and bolts of the underlying engine? There is no one-to-one mapping from beam constructs to runner's granular execution units(say, a bolt in storm). Cheers, Eswar.
