LARA

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
Next revision Both sides next revision
sav08:relational_semantics_of_procedures [2009/05/26 17:45]
vkuncak
sav08:relational_semantics_of_procedures [2009/05/26 20:35]
vkuncak
Line 10: Line 10:
  
 How can we define semantics for program with recursive procedures? How can we define semantics for program with recursive procedures?
 +
 +We consider first the case of procedures without parameters; adding parameters is straightforward.
  
 ===== One Recursive Procedure ===== ===== One Recursive Procedure =====
Line 113: Line 115:
 \] \]
  
-The results extend to any number of call sites and any number of mutually recursive procedures, we just consider a function ​$G : {\cal R}^d \to {\cal R}^d$ where $dis the number of procedures; this mapping describes howgiven one approximation of procedure semantics, compute a better approximation that is correct for longer executions. We have operations $\sqsubseteq$ and $\sqcup$ analogous to $\subseteq$ and $\cup$ (indeedthe domain ${\cal R}^d$, that is $(2^{S\times S})^d$is isomorphic to $2^{S\times S \times \{1,2,\ldots,d\}}$).+We define lattice structure on ${\cal R}^2by  
 +\[ 
 +   (r_1,r_2) \sqsubseteq ​(r'​_1,​r'​_2) ​\mbox{ iff } \ (r_1 \subseteq ​r'​_1) ​\land (r_2 \subseteq r'​_2) 
 +\] 
 +\[ 
 +   (r_1,r_2) \sqcup (r'_1,r'_2) = (r_1 \cup r'_1, r_2 \cup r'​_2) 
 +\] 
 +Note that
 +\[ 
 +   G(\emptyset,​\emptyset= ([\![assume(x==0)]\!] \circ [\![wasEven=true]\!][\![assume(x==0)]\!] \circ [\![wasEven=false]\!]) 
 +\] 
 +\[ 
 +   ​G(G(\emptyset,​\emptyset)) = 
 +\begin{array}[t]{@{}l@{}} 
 +      \bigg( ([\![assume(x==0)]\!] \circ [\![wasEven=true]\!]) \cup 
 +      ([\![assume(x!=0)]\!] \circ [\![x=x-1]\!] \circ G(\emptyset,\emptyset).\_2)\ ​, \
 +      ([\![assume(x==0)]\!] \circ [\![wasEven=false]\!]) \cup 
 +      ([\![assume(x!=0)]\!] \circ [\![x=x-1]\!] \circ G(\emptyset,\emptyset).\_1) \bigg) 
 +\end{array} 
 +\] 
 +where $p.\_1$ and $p.\_2$ denote first, respectively,​ second, element of the pair $p$.
  
-The meaning ​of all procedures ​is therefore the tuple $p_* \in {\cal R}^n$ given by $\bigsqcup\limits_{\ge 0G^n(\emptyset,\ldots,\emptyset)$.+The results extend to any number ​of call sites and any number of mutually recursive ​procedures, we just consider a function ​$G : {\cal R}^\to {\cal R}^d$ where $d$ is the number of procedures; this mapping describes howgiven one approximation of procedure semanticscompute a better approximation that is correct for longer executions
  
-**Remark:​** ​$p_*is the least fixpoint of $G$, so by [[sav08:​Tarski'​s Fixpoint Theorem]], it is also the least $psuch that $G(p) \sqsubseteq p$. Thus, $G(p) \sqsubseteq pimplies ​$p_* \sqsubseteq p$.+We define ​$\sqsubseteqand $\sqcupanalogous to $\subseteqand $\cup$ (indeed, the domain ${\cal R}^d$, that is $(2^{S\times S})^d$, is isomorphic to $2^{S\times S \times \{1,​2,​\ldots,​d\}}$).
  
 +The meaning of all procedures is then the tuple $p_* \in {\cal R}^n$ given by $\bigsqcup\limits_{n \ge 0} G^n(\emptyset,​\ldots,​\emptyset)$.
 +
 +**Remark:** $p_*$ is the least fixpoint of $G$, so by [[sav08:​Tarski'​s Fixpoint Theorem]], it is also the least $p$ such that $G(p) \sqsubseteq p$. Thus, $G(p) \sqsubseteq p$ implies $p_* \sqsubseteq p$.