Theory of Programming Zsolt Borsi

Program constructs

1 Syntax and semantics of program constructs

In this section we introduce some keywords that are allowed to use to build programs we
want to investigate. Next, the semantics of these constructs are provided. We give the
semantics of program constructions so that they are treated as programs in the sense of
the previous definition of program. Our aim is to define the set of state-sequences that are
assigneded to an arbitrary state by the given construction.

Definition: Let A be the common base state space of the program S1 and S2 . Let σ ∈ A be
the current state.
∞
(S1 ; S2 )(σ) ::= {α | α ∈ S1 (σ) ∩ A } ∪
{α | α ∈ S1 (σ) and | α |< ∞ and α|α| = fail} ∪
∗
{α ⊗ β | α ∈ S1 (σ) ∩ A and β ∈ S2 (α|α| )}

(S1 ; S2 ) denotes the sequence of programs S1 and S2 and can be described by a structogram
in the following way:

S1
S2

Definition: Let A be the common base state space of the program S1 . . . Sn and the condi-
tions π1 . . . πn . Let σ ∈ A be the current state.
n
S
(if π1 → S1  . . .  πn → Sn fi)(σ) ::= ω(σ) ∪ Si (σ)
i=1
σ∈Dπi ∧πi (σ)

(
{< σ, fail >} if ∃i ∈ [1..n] : σ ∈
/ Dπi ∨ ∀i ∈ [1..n] : σ ∈ Dπi ∧ ¬πi (σ)
where ω(σ) =
∅ otherwise

if π1 → S1  . . .  πn → Sn fi denotes the branches (or selection) constituted from the pro-

grams S1 , . . . , Sn and the logical functions π1 , . . . , πn , and can be described by a structogram
in the following way:

A π1 · · · A πn
A A
S1 ··· Sn

1
Theory of Programming Zsolt Borsi

Definition: Let A be the common base state space of the program S0 and the condition π.
Let σ ∈ A be the current state.

(S0 ; while π do S0 od)(σ) if σ ∈ Dπ ∧ π(σ)

(while π do S0 od)(σ) ::= {< σ >} if σ ∈ Dπ ∧ ¬π(σ)

{< σ, fail >} if σ ∈
/ Dπ


while π do S0 od denotes the loop constructed by the loopbody S0 and the π loopcondition,
and can be described by a structogram in the following way:

π
S0

2 Example - program given by structogram

2.1 Question
A = (i : N, n : N)
Write down the sequences that are assigned to the states

• (1, 3) and

• (4, 6)

by the following S program.


S
i := 0
i 6= n
A
A
i=1 A
A
i62
i := 3 · i i := i + 1

2.2 Answer
Our S program is a sequence constructed from an assignment i := 0 and a loop. The loop
will stop if its loopcondition becomes f alse, that is when i and n are equal.

The semantics of sequence and loop is well known, however we define the meaning of se-
lection in a different way than it is common in programming languages like JAVA or C++.
In our mathematical model, non-determinism is allowed. Particularly, if there are more
branches of a selection where the conditions are all true for a given state in the statespace,
then we do not select one branch from them according a special rule, instead, all the execu-
tions are allowed that can be generated starting from the given state and taking any of the

2
Theory of Programming Zsolt Borsi

branches where the condition is true.

For example, when i equals to 1 then both the conditions i = 1 and i 6 2 are true, we do
not prefer any of the corresponding programs, both i := i · 3 and i := i + 1 can be executed
starting from the state where the value of variable i is 1.

Another difference in our model compared to programming languages, that if none of the
conditions of a selection is true for a given state, then the selection aborts at the given state.
For example, when i equals to 3 then neither i = 1 nor i 6 2 is true, the selection aborts
starting from a state where the value of variable i is 3.

A program is a relation that assigns sequences to every state in the statespace. All the se-
quences assigned to any state start with the same state they are assigned to. Our S program
maps sequences to every state of the statespace such that the elements of the sequences are
states of A (i.e. the elements of the sequences are pairs).

The assignment i := 0 takes as from a state (i, n) to a state where i is set to 0 and n is un-
changed, that is (0, n). There are two sequences assigned to the state (1, 3):

< (1, 3), (0, 3), (1, 3), (3, 3) > and < (1, 3), (0, 3), (1, 3), (2, 3), (3, 3) >
Notice that the reason why two sequences are assigned to (1, 3) is, that when we are in the
state (1, 3), then both the conditions i = 1 and i 6 2 are true, there are two possible ways
how we can continue the execution of the program. In the state (3, 3) the loop stops, as its
loopcondition becomes f alse.

There are two sequences assigned to the state (4, 6):

< (4, 6), (0, 6), (1, 6), (3, 6), f ail > and < (4, 6), (0, 6), (1, 6), (2, 6), (3, 6), f ail >
Notice that the reason why these sequences terminates in the special f ail state is, that in the
state (3, 6) neither the conditions i = 1 and nor i 6 2 is true, the selection (the body of the
loop) aborts.

More precisely, the following two sequences are assigned to the state {i : 1, n : 3} by the
program:
< {i : 1, n : 3}, {i : 0, n : 3}, {i : 1, n : 3}, {i : 3, n : 3} > and < {i : 1, n : 3}, {i : 0, n : 3}, {i :
1, n : 3}, {i : 2, n : 3}, {i : 3, n : 3} >
The program assigns the following two sequences to the state {i : 4, n : 6}:
< {i : 4, n : 6}, {i : 0, n : 6}, {i : 1, n : 6}, {i : 3, n : 6}, f ail > and < {i : 4, n : 6}, {i : 0, n :
6}, {i : 1, n : 6}, {i : 2, n : 6}, {i : 3, n : 6}, f ail >