These are chat archives for jdubray/sam

Aug 2017
Jean-Jacques Dubray
Aug 02 2017 01:34
Not sure which genius created the concept of template, but Angular for sure brought it to the next level: [attr.href]="'#' + aType.Name"
Janne Siera
Aug 02 2017 06:57
Isn't that because they're trying to be a platform agnostic framework? I.e. they try to abstract away the whole dom?
Antanas A.
Aug 02 2017 07:05
[attr.href]="'#' + aType.Name" <- is it string literal?
template in template
Aug 02 2017 09:21
I actually saw what that does immediately, maybe I spent too much time in Angular templates :P
Jean-Jacques Dubray
Aug 02 2017 12:48
whatever the reason, nobody can convince me that it requires less skill and it's faster than ES6 template literal/functional HTML. I'd be curious to know how they came up with this design decision for such a simple capability. Is there some grand order that I am missing? am I completely clueless? Even if that could be easier for a designer to implement (which I seriously doubt), it would require a software engineer to learn that syntax to debug and maintain that part of the "code".
Jean-Jacques Dubray
Aug 02 2017 17:48
@antanas-arvasevicius I still don't understand what's different between ASM and FSM/NDFSM, but when you look, say at the counter example, clear actions mutate the application state, which would be expected in a classical state machine.
Jean-Jacques Dubray
Aug 02 2017 17:53
From what I can tell, ASM provide a "pseudo-code" implementation of classical actions, that's what is typically allowed in an action:
I believe like everyone else they realized how hard it was to code with classical state machines, so they provided some constraints to make it more manageable.
IMHO, the problem is not how you code the actions, it is the overall semantics of the classical state machines that is wrong.
Jean-Jacques Dubray
Aug 02 2017 17:59
Intuitively, I think they understand that "assignments" are the problem, so they introduce the notion of "consistent assignment"