A characterization of the consistent Hoare powerdomains over dcpos

Abstract

It has been shown that for a dcpo $P$, the Scott closure of $\Gamma_{c}(P)$ in $\Gamma(P)$ is a consistent Hoare powerdomain of $P$, where $\Gamma_{c}(P)$ is the family of nonempty, consistent and Scott closed subsets of $P$, and $\Gamma(P)$ is the collection of all nonempty Scott closed subsets of $P$. In this paper, by introducing the notion of a $\mathord{\bigvee}{-}$existing set, we present a direct characterization of the consistent Hoare powerdomain: the set of all $\mathord{\bigvee}{-}$existing Scott closed subsets of a dcpo $P$ is exactly the consistent Hoare powerdomain of $P$. We also introduce the concept of an $F$-Scott closed set over each dcpo-$\vee_{\mathord{\uparrow}}$-semilattice. We prove that the Scott closed set lattice of a dcpo $P$ is isomorphic to the family of all $F$-Scott closed sets of $P$’s consistent Hoare powerdomain.

1 Introduction

The Hoare powerdomain plays an important role in modeling the programming semantics of nondeterminism, which is analogous to the powerset construction (see for example [3, 4, 5, 7, 8, 9]). The Hoare power domain over a dcpo (directed complete poset) is the free inflationary semilattice where the inflationary operator is exactly the binary join operator. The standard construction of a Hoare powerdomain consists all nonempty Scott closed subsets $\Gamma(P)$ of the dcpo $P$, order given by the inclusion relation. In [10], Yuan and Kou introduced a new type of powerdomain, called the consistent Hoare powerdomain. The new powerdomain over a dcpo is a free algebra where the inflationary operator delivers joins only for consistent pairs. They provided a concrete way of constructing the consistent Hoare powerdomain over a domain (continuous dcpo): the family of all nonempty relatively consistent Scott closed subsets of the domain is such a powerdomain. Follows from the work, there is a natural problem: whether every dcpo has a consistent Hoare powerdomain. Geng and Kou [1] gave an affirmative answer to the question by showing that the Scott closure $\overline{\Gamma_{c}(P)}$ of $\Gamma_{c}(P)$ in $\Gamma(P)$ is a consistent Hoare powerdomain over the dcpo $P$, where $\Gamma_{c}(P)$ is the collection of all nonempty consistent Scott closed subsets of $P$. One question arises naturally: can we give a direct characterization of the consistent Hoare powerdomain over every dcpo $P$? That is to say, what type of Scott closed subsets of $P$ is exactly the consistent Hoare powerdomain? This paper is mainly set to answer this question.

The consistent Hoare powerdomain can be viewed as a type of completion that embeds every dcpo into a dcpo which is also complete with respect to consistent pair joins. In [11], Zhao and Fan introduced a type of dcpo-completion embedding each poset into a dcpo, called $D$-completion. They proved that the Scott closed set lattice of a poset is order isomorphic to that of the $D$-completion. It is also the case for the sobrification, i.e., the topology lattice $\mathcal{O}(X)$ of a $T_{0}$-space $X$ is isomorphic to the topology lattice $\mathcal{O}(X^{s})$ of the sobrification $X^{s}$ of $X$ (see [2]). In this paper, we introduce a type of closed sets on every dcpo-$\vee_{\mathord{\uparrow}}$-semilattice, called $F$-Scott closed sets. We prove that the Scott closed set lattice of a dcpo is isomorphic to the family of all $F$-Scott closed subsets of its consistent Hoare powerdomain. It is known that two sober dcpos are isomorphic if and only if their Scott set lattices are isomorphic (see [6, 12]). Consequently, the consistent Hoare powerdomains for sober dcpos are uniquely determined up to isomorphism.

2 Preliminaries

This section is an introduction to the concepts and results about the consistent Hoare powerdomains over domains and dcpos. For more details, refer to [1, 2, 10]. In this paper, the order relation on each family of sets is the set-theoretic inclusion relation.

A nonempty subset $D$ of a poset $P$ is called directed if every pair has an upper bound in $D$. If every directed $D\subseteq P$ has a supremum $\mathord{\bigvee}D$, then $P$ is called a dcpo. For $A\subseteq P$, let $\mathord{\downarrow}A=\{x\in P:x\leq a$ for some $a\in A\}$, $\mathord{\downarrow}x=\mathord{\downarrow}\{x\}$ where $x\in P$. If $A=\mathord{\downarrow}A$, then $A$ is called a lower set. For $x,y\in P$, we say $x$ is way below $y$, written $x\ll y$, if $y\leq\mathord{\bigvee}D$ implies $x\in\mathord{\downarrow}D$ for any directed $D\subseteq P$. Let $\mathord{\mbox{\makebox[0.0pt][l]{\raisebox{-1.72218pt}{$\downarrow$}}$\downarrow$}}x=\{y\in P:y\ll x\}$, and $\mathord{\mbox{\makebox[0.0pt][l]{\raisebox{-1.72218pt}{$\downarrow$}}$\downarrow$}}A=\bigcup\{\mathord{\mbox{\makebox[0.0pt][l]{\raisebox{-1.72218pt}{$\downarrow$}}$\downarrow$}}a:a\in A\}$. A dcpo $P$ is called a domain if each element $x$ is the directed join of $\mathord{\mbox{\makebox[0.0pt][l]{\raisebox{-1.72218pt}{$\downarrow$}}$\downarrow$}}x$. A subset $A$ of a dcpo $P$ is Scott closed if it is a lower set and closed under directed sups. Let $cl(A)$ denote the Scott closure of $A$.

Definition 2.1.

[10] A consistent inflationary semilattice is a dcpo $P$ with a Scott continuous binary partial operator $\biguplus_{\mathord{\uparrow}}$ defined only for consistent pairs of points that satisfies three equations for commutativity $x\biguplus_{\mathord{\uparrow}}y=y\biguplus_{\mathord{\uparrow}}x$, associativity $x\biguplus_{\mathord{\uparrow}}(y\biguplus_{\mathord{\uparrow}}z)=(x\biguplus_{\mathord{\uparrow}}y)\biguplus_{\mathord{\uparrow}}z$ and idempotency $x\biguplus_{\mathord{\uparrow}}x=x$ together with the inequality $x\leq x\biguplus_{\mathord{\uparrow}}y$ for $x,y,z\in P$. The free consistent inflationary semilattice over a dcpo $P$ is called the consistent Hoare powerdomain over $P$ and denoted by $H_{c}(P)$.

For a consistent inflationary semilattice, the operator $\biguplus_{\mathord{\uparrow}}$ coincides with $\vee_{\mathord{\uparrow}}$, the join operator defined only for consistent pairs. That is to say, a consistent inflationary semilattice is exactly a dcpo that is a $\vee_{\mathord{\uparrow}}$-semilattice, also called a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice. A Scott closed subset $A$ of a domain $L$ is called relatively consistent if $A$ is the Scott closure of the directed union $\bigcup^{\mathord{\uparrow}}\{\mathord{\downarrow}F:F\in\mathcal{F}_{C}(A)\}$, where

$\mathcal{F}_{C}(A)=\{F:F$ is a nonempty finite consistent subset of $L$ and $F\subseteq\mathord{\mbox{\makebox[0.0pt][l]{\raisebox{-1.72218pt}{$\downarrow$}}$\downarrow$}}A\}$.

The set of all nonempty relatively consistent Scott closed subsets of $L$ is denoted by $R\Gamma_{C}(L)$.

Theorem 2.2.

[10] Let $L$ be a domain. The consistent Hoare powerdomain $H_{c}(P)$ over $L$ can be realized as $R\Gamma_{C}(L)$ where $\biguplus_{\mathord{\uparrow}}=\bigcup$. The embedding $j$ of $L$ into $R\Gamma_{C}(L)$ is given by $j(x)=\mathord{\downarrow}x$ for $x\in L$.

Let $\Gamma_{c}(P)$ denote the family of all nonempty, consistent and Scott closed subsets of a dcpo $P$. And let $\overline{\Gamma_{c}(P)}$ be the Scott closure of $\Gamma_{c}(P)$ in $\Gamma(P)$. The family $\overline{\Gamma_{c}(P)}$ was introduced by Geng and Kou [1], in order to declare that the consistent Hoare powerdomain over every dcpo exists.

Lemma 2.3.

[1] Let $P$ be a dcpo and $L$ is a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice. If $f:P\rightarrow L$ is a Scott continuous function, then for any $A\in\overline{\Gamma_{c}(P)}$, $\mathord{\bigvee}f(A)$ exists in $L$.

Theorem 2.4.

[1] Let $P$ be a dcpo. Then the consistent Hoare powerdomain $H_{c}(P)$ over $P$ can be realized as the $\overline{\Gamma_{c}(P)}$, where $A\vee_{\mathord{\uparrow}}B=A\bigcup B$ whenever $A,B$ are consistent in $\overline{\Gamma_{c}(P)}$. The embedding $j$ of $P$ into $\overline{\Gamma_{c}(P)}$ is given by $j(x)=\mathord{\downarrow}x$ for any $x\in P$.

3 A direct characterization of the consistent Hoare powerdomain

Definition 3.1.

A subset $A$ of a dcpo $P$ is called $\mathord{\bigvee}{-}$existing if for any continuous function $f:P\rightarrow L$ mapping into a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $L$, $\mathord{\bigvee}f(A)$ always exists in $L$.

For any directed subset $D$ of $P$, $f(D)$ is directed in $L$ since continuous functions are monotone, and then $\mathord{\bigvee}f(D)$ exists in the dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $L$. Hence every directed subset is $\mathord{\bigvee}{-}$existing. Similarly, every consistent finite subset is also $\mathord{\bigvee}$-existing. Notice that an empty set is not $\mathord{\bigvee}{-}$existing because a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $L$ may not have a least element.

Proposition 3.2.

Let $f:P\rightarrow L$ be a continuous function from dcpo $L$ to dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $L$, and $A\subseteq P$. Then

(1) $\mathord{\bigvee}f(A)$ exists $\Leftrightarrow$ $\mathord{\bigvee}f(cl(A))$ exists $\Rightarrow$ $\mathord{\bigvee}f(A)=\mathord{\bigvee}f(cl(A))$.

(2) $A$ is $\mathord{\bigvee}$-existing if and only if $cl(A)$ is $\mathord{\bigvee}$-existing.

Proof.

(1) Assume that $x$ is an upper bound of $f(A)$. Then $f(A)\subseteq\mathord{\downarrow}x$ and $A\subseteq f^{-1}(\mathord{\downarrow}x)\in\Gamma(P)$. Hence $cl(A)\subseteq f^{-1}(\mathord{\downarrow}x)$, and then $f(cl(A))\subseteq\mathord{\downarrow}x$, i.e., $x$ is an upper bound of $f(cl(A))$. The converse is clearly true. Thus $\mathord{\bigvee}f(A)$ exists iff $\mathord{\bigvee}f(cl(A))$ exists, and both imply that $\mathord{\bigvee}f(A)=\mathord{\bigvee}f(cl(A))$.

(2) It is straightforward from (1) and Definition 3.1. ∎

For any dcpo $P$, we write $P^{\vee}=\{A\subseteq P:A$ is $\mathord{\bigvee}$-existing and Scott closed$\}$. By Lemma 2.3, we immediately have that $\overline{\Gamma_{c}(P)}\subseteq P^{\vee}$. We shall show that the converse inclusion is also true, and then a characterization of the consistent Hoare powerdomain is obtained.

Definition 3.3.

A subset $A$ of a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $L$ is called $F$-Scott closed if it is Scott closed and for any consistent nonempty finite set $F\subseteq A$, $\mathord{\bigvee}F\in A$. We write $\Gamma_{F}(L)$, called the $F$-Scott closure system on $L$, for the set of all $F$-Scott closed subsets of $L$, and let $cl_{F}$ denote the corresponding closure operator. A function $f:L\rightarrow M$ between dcpo-$\vee_{\mathord{\uparrow}}$-semilattices is called $F$-Scott continuous if $f$ is continuous with respect to the $F$-Scott closure systems.

Proposition 3.4.

(1) A function $f:L\rightarrow M$ between dcpo-$\vee_{\mathord{\uparrow}}$-semilattices is a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice homomorphism iff $f$ is $F$-Scott continuous.

(2) If a nonempty subset $A$ of a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $L$ is consistent, then $cl_{F}(A)=\mathord{\downarrow}\mathord{\bigvee}A$.

Proof.

(1) Suppose that $f$ is a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice homomorphism. Let $A\subseteq M$ be $F$-Scott closed. We shall show that $f^{-1}(A)\subseteq L$ is also $F$-Scott closed. For any consistent nonempty finite $F\subseteq f^{-1}(A)$, we have $f(F)$ is also nonempty finite and consistent since $f$ is monotone. Then $f(\mathord{\bigvee}F)=\mathord{\bigvee}f(F)\in A$ and then $\mathord{\bigvee}F$ is in $f^{-1}(A)$. Similarly, $f^{-1}(A)$ is also closed with respect to directed joins. Thus $f$ is $F$-Scott continuous.

Conversely, let $f$ be $F$-Scott continuous. Clearly, $f$ is monotone. Let $F$ be a consistent nonempty finite subset of $L$. Then $\mathord{\bigvee}f(F)\leq f(\mathord{\bigvee}F)$. The set $\mathord{\downarrow}\mathord{\bigvee}f(F)$ is $F$-Scott closed in $M$. Then $f^{-1}(\mathord{\downarrow}\mathord{\bigvee}f(F))$ is also $F$-Scott closed and $F\subseteq f^{-1}(\mathord{\downarrow}\mathord{\bigvee}f(F))$. Hence $\mathord{\bigvee}F\in f^{-1}(\mathord{\downarrow}\mathord{\bigvee}f(F))$ and then $f(\mathord{\bigvee}F)\leq\mathord{\bigvee}f(F)$. Thus $f(\mathord{\bigvee}F)=\mathord{\bigvee}f(F)$. Analogously, $f(\mathord{\bigvee}D)=\mathord{\bigvee}f(D)$ for all directed $D\subseteq L$. Therefore, $f$ is a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice homomorphism.

(2) Let $C\subseteq L$ be any $F$-Scott closed set with $A\subseteq C$. Notice that each subset of $A$ is also consistent. Then $D:=\{\mathord{\bigvee}F:F\subseteq A$ is nonempty and finite$\}$ is a directed subset of $C$, and hence $\mathord{\bigvee}D=\mathord{\bigvee}A\in C$. Moreover, every $F$-Scott closed set is a lower set. Then $\mathord{\downarrow}\mathord{\bigvee}A\subseteq C$, and thus $cl_{F}(A)=\mathord{\downarrow}\mathord{\bigvee}A$. ∎

Definition 3.5.

Let $L$ be a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice. A subset $A\subseteq L$ is called $F$-$\mathord{\bigvee}$-existing if for any dcpo-$\vee_{\mathord{\uparrow}}$-semilattice homomorphism $f:L\rightarrow M$ mapping into a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $M$, $\mathord{\bigvee}f(A)$ exists in $M$.

Lemma 3.6.

Let $L$ be a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice. A subset $A\subseteq L$ is $F$-$\mathord{\bigvee}$-existing iff $cl_{F}(A)$ is $F$-$\mathord{\bigvee}$-existing.

Proof.

The process is similar to that of Proposition 3.2. Notice that every principal ideal $\mathord{\downarrow}x$ is $F$-Scott closed. ∎

Lemma 3.7.

If a subset $A$ of a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $L$ is $F$-Scott closed and $F$-$\mathord{\bigvee}$-existing, then $A=\mathord{\downarrow}a$ for some $a\in L$.

Proof.

The identity function $id:L\rightarrow L$ is a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice homomorphism. We have that $\mathord{\bigvee}id(A)=\mathord{\bigvee}A$ exists in $L$. Then every nonempty finite subset $F\subseteq A$ is consistent, and hence $\mathord{\bigvee}F$ exists. Since $A$ is $F$-Scott closed, we obtain that $D=\{\mathord{\bigvee}F:F\subseteq A$ is nonempty and finite$\}$ is a directed subset of $A$, and then $\mathord{\bigvee}D=\mathord{\bigvee}A\in A$. Thus $A=\mathord{\downarrow}\mathord{\bigvee}A$, which completes the proof. ∎

Lemma 3.8.

A subset $A$ of a dcpo $P$ is $\mathord{\bigvee}$-existing iff $j(A)$ is an $F$-$\mathord{\bigvee}$-existing subset of $\overline{\Gamma_{c}(P)}$. (Notice that $j$ is the function in Theorem 2.4.)

Proof.

Suppose that $A\subseteq P$ is $\mathord{\bigvee}$-existing. Let $f:\overline{\Gamma_{c}(P)}\rightarrow L$ be any dcpo-$\vee_{\mathord{\uparrow}}$-semilattice homomorphism mapping into a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $L$. Then $f\circ j$ is a Scott continuous function from $P$ to $L$, and hence $\mathord{\bigvee}f\circ j(A)=\mathord{\bigvee}f(j(A))$ exists in $L$. Thus $j(A)$ is an $F$-$\mathord{\bigvee}$-existing subset of $\overline{\Gamma_{c}(P)}$.

Now assume that $j(A)$ is an $F$-$\mathord{\bigvee}$-existing subset of $\overline{\Gamma_{c}(P)}$. Let $f:P\rightarrow L$ be a Scott continuous mapping into a dcpo-$\vee_{\mathord{\uparrow}}$-semilattice $L$. Then by Theorem 2.4, there exists a unique dcpo-$\vee_{\mathord{\uparrow}}$-semilattice homomorphism $\tilde{f}:\overline{\Gamma_{c}(P)}\rightarrow L$ such that $f=\widetilde{f}\circ j$. Then $\mathord{\bigvee}\tilde{f}(j(A))=\mathord{\bigvee}\widetilde{f}\circ j(A)=\mathord{\bigvee}f(A)$ exists. Thus $A$ is a $\mathord{\bigvee}$-existing subset of $P$. ∎

We now come to the characterization of the consistent Hoare powerdomains by $\mathord{\bigvee}$-existing Scott closed sets:

Theorem 3.9.

Let $P$ be a dcpo and $A\subseteq P$. Then $A\in\overline{\Gamma_{c}(P)}$ iff $A$ is Scott closed and $\mathord{\bigvee}$-existing, i.e., $\overline{\Gamma_{c}(P)}=P^{\vee}$.

Proof.

By Lemma 2.3, we have that $\overline{\Gamma_{c}(P)}\subseteq P^{\vee}$. Now suppose that $A\in P^{\vee}$. Then, by Lemma 3.8, the set $j(A)$ is an $F$-$\mathord{\bigvee}$-existing subset of $\overline{\Gamma_{c}(P)}$. And by Lemma 3.6, the $F$-Scott closed set $cl_{F}(j(A))$ is also $F$-$\mathord{\bigvee}$-existing. Then $cl_{F}(j(A))$ has a supremum which is in itself by Lemma 3.7. Since the order on $\overline{\Gamma_{c}(P)}$ is the inclusion relation, we have $\mathord{\bigvee}cl_{F}(j(A))=\bigcup cl_{F}(j(A))\in\overline{\Gamma_{c}(P)}$. For each $a\in A$, $j(a)=\mathord{\downarrow}a\in cl_{F}(j(A))$, and then $A\subseteq\bigcup cl_{F}(j(A))$. Thus $A\in\overline{\Gamma_{c}(P)}$ because $\overline{\Gamma_{c}(P)}$ is a lower set of $\Gamma(P)$, which completes the proof. ∎

Theorem 3.10.

Let $P$ be a dcpo. The family of Scott closed sets $\Gamma_{0}(P)=\Gamma(P)\bigcup\{\emptyset\}$ is order isomorphic to the $F$-Scott closure system $\Gamma_{F}(H_{c}(P))$ on the consistent Hoare powerdomain $H_{c}(P)$, i.e., $\Gamma_{0}(P)\cong\Gamma_{F}(H_{c}(P))$.

Proof.

Define a function $\eta:\Gamma_{0}(P)\rightarrow\overline{\Gamma_{c}(P)}$ by

$\eta(A)=cl_{F}(j(A))$,

where $cl_{F}(j(A))$ is the $F$-Scott closure of $j(A)=\{\mathord{\downarrow}a:a\in A\}$ in $\overline{\Gamma_{c}(P)}$. We shall show that $\eta$ is an order isomorphism. Obviously, $\eta$ is monotone.

Firstly, we prove that $\eta$ is injective. Suppose that $A,B\in\Gamma_{0}(P)$ with $A\neq B$. Without loss of generality, we may assume that there is $b\in B{\setminus}A$. Let $\mathcal{C}=\{C\in\overline{\Gamma_{c}(P)}:C\subseteq A\}$. Clearly, $j(A)\subseteq\mathcal{C}$ and $\mathord{\downarrow}b\notin\mathcal{C}$. We claim that $\mathcal{C}$ is $F$-Scott closed in $\overline{\Gamma_{c}(P)}$. Then $cl_{F}(j(A))\subseteq\mathcal{C}$, and hence $\mathord{\downarrow}b\notin cl_{F}(j(A))$. But $\mathord{\downarrow}b\in j(B)\subseteq cl_{F}(j(B))$. Thus $\eta(A)\neq\eta(B)$, i.e., $\eta$ is injective. Indeed, let $\mathcal{D}\subseteq\mathcal{C}$ be directed, then $\mathord{\bigvee}\mathcal{D}$, the supremum of $\mathcal{D}$ in $\overline{\Gamma_{c}(P)}$, is the Scott closure of $\bigcup\mathcal{D}$ in $P$, and hence $\mathord{\bigvee}\mathcal{D}\subseteq A$ since each element of $\mathcal{D}$ is contained in $A$, i.e., $\mathord{\bigvee}\mathcal{D}\in\mathcal{C}$. And if $C_{1},C_{2}\in\mathcal{C}$, then $C_{1}\bigcup C_{2}\subseteq A$, thus $C_{1}\vee C_{2}\in\mathcal{C}$ (notice that $\overline{\Gamma_{c}(P)}$ is closed under finite unions), which proves the claim.

We next prove that $\eta$ is surjective. Suppose that $\mathcal{A}$ is $F$-Scott closed in $\overline{\Gamma_{c}(P)}$. We claim that $\bigcup\mathcal{A}$ is Scott closed in $P$. Indeed, if $D\subseteq\bigcup\mathcal{A}$ is directed, then $j(D)\subseteq\mathcal{A}$ is directed, and hence $\mathord{\bigvee}j(D)=\mathord{\downarrow}\mathord{\bigvee}D\in\mathcal{A}$, i.e., $\mathord{\bigvee}D\in\bigcup\mathcal{A}$. We shall show that $\eta(\bigcup\mathcal{A})=\mathcal{A}$. Since $j(\bigcup\mathcal{A})\subseteq\mathcal{A}$, we have $\eta(\bigcup\mathcal{A})=cl_{F}(j(\bigcup\mathcal{A}))\subseteq\mathcal{A}$. Conversely, assume that $A\in\mathcal{A}$. Then $j(A)$ is consistent in $\overline{\Gamma_{c}(P)}$. By Proposition 3.4(2), we have $\mathord{\bigvee}j(A)=A\in cl_{F}(j(A))\subseteq cl_{F}(j(\bigcup\mathcal{A}))$. Thus $\eta(\bigcup\mathcal{A})=cl_{F}(j(\bigcup\mathcal{A}))=\mathcal{A}$, i.e., $\eta$ is surjective, which completes the proof. ∎

A dcpo is called sober if every Scott closed irreducible set is of the form $\mathord{\downarrow}x$. It is known that for sober dcpos $P_{1}$ and $P_{2}$, $\Gamma_{0}(P_{1})\cong\Gamma_{0}(P_{2})$ iff $P_{1}\cong P_{2}$ (see [6, 12] for more results about uniqueness of dcpos based on the Scott closed set lattices). In particularly, the Scott topology of every quasicontinuous domain is sober (each domain is a quasicontinuous domain, see [2] for the detailed definition of a quasicontinuous domain). By applying the above theorem, we immediately have that for sober dcpos, the consistent Hoare powerdomains are uniquely determined up to isomorphism:

Corollary 3.11.

If dcpos $P_{1}$ and $P_{2}$ are sober, then $H_{c}(P_{1})\cong H_{c}(P_{2})$ iff $P_{1}\cong P_{2}$.

Proof.

By Theorem 3.10, $H_{c}(P_{1})\cong H_{c}(P_{2})$ implies $\Gamma_{F}(H_{c}(P_{1}))\cong\Gamma_{F}(H_{c}(P_{2}))$ iff $\Gamma_{0}(P_{1})\cong\Gamma_{0}(P_{2})$ iff $P_{1}\cong P_{2}$ implies $H_{c}(P_{1})\cong H_{c}(P_{2})$. ∎

4 Acknowledgments

This work is supported by National Natural Science Foundation of China (No.11801491,11771134) and Shandong Provincial Natural Science Foundation, China (No. ZR2018BA004).