Closed lou1306 closed 6 years ago
Let us generalize this:
r = r1, ..., rn
be a sequence of (distinct) references to interface (r. stigmergic, environment) variablese1, ..., en
a sequence of expressionsThen the basic process r1, ... rn <- e1, ..., en
(r. r1, ... rn <~ e1, ..., en
, r1, ... rn <-- e1, ..., en
):
e1, ..., en
against the current state of the agent. Let v1, ..., vn
the resulting valuesv_i
to the corresponding r_i
.Notice that after the process is completed, the global clock will have been increased n
times. Currently, this has no consequences, as tupled stigmergic variables still end up with equal timestamps. However, we should correct this if we want use the global clock for other things in the future.
Suppose
<var1, var2>
is a tuple inside some stigmergy.Right now an agent can only update the variables one at a time:
In some scenarios, however, it would be useful to allow for an atomic update of both values. Proposed syntax is
which should avoid conflicts with the existing constructs