These are chat archives for jdubray/sam

Jun 2017
Jean-Jacques Dubray
Jun 20 2017 01:14 UTC
I found this panel quite interesting (in the lack of direction)
I do think we need to go beyond types and think in terms of Model integrity, it's ridiculous to think that a data structure definition invented in the 70s still makes sense today.
Jean-Jacques Dubray
Jun 20 2017 01:22 UTC
Static Typing makes us lazy for the validation rules that matter
Jean-Jacques Dubray
Jun 20 2017 13:52 UTC
It was also interesting to hear that OOP and FP are pretty much two faces of the same coin.
The most interesting comment with Excel as a programming language, there are "500 Million" Excel developers and yet the software engineer community don't pay any attention to it. I am not too surprised to see that SAM fits like a glove behind Excel's programming model.
Fred Daoud
Jun 20 2017 14:10 UTC
I am not too surprised to see that SAM fits like a glove behind Excel's programming model.
I find that interesting, would you care to elaborate JJ?
devin ivy
Jun 20 2017 14:16 UTC
i think he's referencing excel's reactive interface
i recall hearing that it's based on the "reactive loop" (i.e. roughly the flux pattern)
Jean-Jacques Dubray
Jun 20 2017 14:43 UTC

@devinivy @foxdonut it goes a bit beyond reactivity. Of course Excel is 100% and possibly the best illustration of Reactive Programming, but Excel could be implemented (not saying it is the best implementation) with the SAM pattern:

  • input to new cell -> action
  • all the cell (constants, formulas...) -> model
  • UI -> state representation
  • next-action -> I think Excel has something equivalent, but I don't remember exactly

Excel does not break the semantics of a SAM pattern a bit

You can see that in Reactive Programming the graph of dependencies is too constraining to build programs (other than Excel), SAM doesn't require any of these dependencies. You can "code" your model, no need for declarative dependencies. Like MobX showed, these dependencies fit well in the State function to create the State Representation, but not to mutate the model...
Marcus Feitoza
Jun 20 2017 23:15 UTC
Hi, anyone want to play with Elm + SAM?