All initialisation constraints can now be described in the assume section
assume also allows for relational constraints (including with the agents' id)
The parser for initializers was never that good to begin with
I am also growing weary of undef. It was probably a bad decision to add it in the first place, its semantics are extremely punishing, etc.
My suggestion is:
Variable declarations (environment, interface, stigmergy tuples) become mere lists of names
If a variable is unconstrained then its initial value is any integer
I would just do away with undef
About stigmergy syntax:
Right now variables within the same tuple are separated by commas, while tuples are separated by semicolons, and they have no "title". E.g., a,b: 0, 0; c; 1
I suggest to add a variables title and to use parentheses to group tutples, i.e., variables = (a, b), c
Rationale:
assume
sectionassume
also allows for relational constraints (including with the agents' id)I am also growing weary of
undef
. It was probably a bad decision to add it in the first place, its semantics are extremely punishing, etc.My suggestion is:
environment
,interface
, stigmergy tuples) become mere lists of namesundef
About
stigmergy
syntax:a,b: 0, 0; c; 1
variables
title and to use parentheses to group tutples, i.e.,variables = (a, b), c