http://arxiv.org/pdf/1304.2290v7.pdf
Interaction system of MLC implements token-passing nets and embeds read-back mechanism.
Specifically, starting from initial configuration that encodes a λ-term
<x | r[ ](x) = y, Γ(M•, y)>,
the r(ead) agent will trigger duplication and application:
s[c(x, rC[ ](y)), x] >< rC[ ][y];
@[λ(x, rC[ ](y)), x] >< rC[ ][y];
and will construct normal form of the encoded λ-term:
aM >< λ[rM [ ](x), x];
aM >< rC[ ][aC[M]];
λ[ay, rC[λy.[ ]](x)] >< rC[ ][x], where y ∈ Λ is a new variable.
If normal form N of the λ-term exists, interaction results in configuration
<aN | ∅>
which consists of only one a(tom) agent in its interface.
Interaction system of MLC implements token-passing nets and embeds read-back mechanism.
Specifically, starting from initial configuration that encodes a λ-term
<x | r[ ](x) = y, Γ(M•, y)>,
the r(ead) agent will trigger duplication and application:
s[c(x, rC[ ](y)), x] >< rC[ ][y];
@[λ(x, rC[ ](y)), x] >< rC[ ][y];
and will construct normal form of the encoded λ-term:
aM >< λ[rM [ ](x), x];
aM >< rC[ ][aC[M]];
λ[ay, rC[λy.[ ]](x)] >< rC[ ][x], where y ∈ Λ is a new variable.
If normal form N of the λ-term exists, interaction results in configuration
<aN | ∅>
which consists of only one a(tom) agent in its interface.