Ecole Normale Supérieure de Cachan 61, avenue du Président Wilson Cachan Cedex France

Size: px
Start display at page:

Download "Ecole Normale Supérieure de Cachan 61, avenue du Président Wilson Cachan Cedex France"

Transcription

1 R. Hennicker and M. Bidoit Observational Logic Research Report LSV 98 6, June 1998 Ecole Normale Supérieure de Cachan 61, avenue du Président Wilson Cachan Cedex France

2 Observational Logic Rolf Hennicker*, Michel Bidoit** *Institut für Informatik, Ludwig-Maximilians-Universität München Oettingenstr. 67, D München, GERMANY **Laboratoire Specification et Verification, CNRS & Ecole Normale Supérieure de Cachan 61, Avenue du President Wilson, Cachan Cedex, FRANCE Abstract. We present an institution of observational logic which generalizes earlier approaches to observational systems specification in various ways. First, we introduce a notion of an observational signature which incorporates the declaration of a distinguished set of observers. Then, we define observational algebras whose operations are required to be compatible with the indistinguishability relation determined by the observers of an observational signature. In particular, we introduce a homomorphism concept for observational algebras which adequately expresses observational relationships between algebras. Then we consider a flexible notion of observational signature morphism which guarantees the satisfaction condition of institutions w.r.t. observational satisfaction for arbitrary first-order sentences. From the proof theoretical point of view we construct a sound and complete proof system for the observational consequence relation. Then we consider structured observational specifications (which, in contrast to previous approaches of the authors, have a built in observational semantics) and we provide a sound and complete proof system for such specifications by using a general, institution independent result of [B 98]. 1 Introduction In this paper we study a logical framework for the specification of the observable behaviour of software systems which is particularly suited for state based systems but may also be used for specifying infinite data and behavioural properties of abstract data types. Formally, we introduce an institution of observational logic and we study proof methods for first-order observational properties of structured specifications built over this institution. Although our approach is novel, it is influenced by previous behavioural approaches, in particular of [BHW 95], [P 96], [GM 97], [R 95] and [JR 97]. The important difference to [BHW 95] is that in the present approach we use a built in observational semantics which previously led to problems w.r.t. the encapsulation of observational properties of parts of a system specification (cf. [BB 91], [HN 93]). In the hidden sorted algebra approach (cf. e.g. [GM 97]) encapsulation is achieved at the cost of a rather restrictive notion of signature morphism (see the discussion in Section 5). Another popular formalism which deals with a state based view of systems is provided in the framework of coalgebras (cf. e.g. [R 95], [JR 97]). These approaches, however, have problems to deal with n-ary operations working on several non observable (hidden) argument sorts which frequently occurs in practice (see also Section 5). Moreover, coalgebraic approaches are based on terminal semantics while we are interested in a loose semantics in order to achieve sufficient freedom for the choice of implementations. The starting point of our approach is a methodological consideration: A well-known method for specifying abstract data types (or, more concretely, for writing functional programs) is to determine first a set of constructor symbols which describe how the 1

3 elements of the data type are constructed and then to define functions which can be applied to the data (usually by case distinction over the given constructors). An analogous method can be used for specifying state based systems. First, a set of observer symbols is declared which determines an indistinguishability relation (also called observational equality) for the non observable elements (i.e. for the states). Then the operations are specified, usually by describing their effects w.r.t. the given observers. Formally, these considerations lead to our notion of an observational signature which contains a distinguished set of observer symbols. A similar idea was presented in [P 96] (and recently in [DF 98]). 1 Based on the notion of an observational signature we define observational algebras as those structures whose operations are compatible with the observational equality (determined by the observers of the signature). In this way we obtain, in Section 3, a category of observational algebras with a notion of observational homomorphism that is suited to express observational relationships between algebras. Moreover, we establish a full and faithful functor from the category of observational algebras to the category of standard algebras which is compatible with the observational satisfaction relation defined in Section 4. In Section 5, we introduce the institution of observational logic. It turns out that our general notion of an observer (used in observational signatures) allows us to define a powerful notion of observational signature morphism which guarantees, nevertheless, that the (observational) satisfaction condition of the institution is valid. Then, in Section 6, we define a sound and complete proof system for observational logic. The results obtained so far allow us, first, to apply a generic construction of structured specifications over an arbitrary institution (cf. [ST 88]) thus obtaining a basic language of structured observational specifications. Secondly, we can also apply a generic construction of a sound and complete proof system for structured specifications (cf. [B 98]) which leads to a corresponding proof system for structured observational specifications. For this purpose we check that observational logic has pushouts and amalgamations and satisfies the interpolation property. 2 Algebraic Preliminaries We assume the reader to be familiar with the basic notions of algebraic specifications (cf. e.g. [LEW 96]), like the notions of (many sorted) signature Σ = (S, OP) with a set S of sorts and a set OP of operation symbols op: s 1,,s n s, total Σ-algebra A = ((A s ) s S, (f A ) f F ), Σ-term algebra T(Σ, X), valuation α : X A and interpretation I α : T(Σ, X) A. Throughout this paper we assume that the carrier sets A s of a Σ-algebra are not empty and that X = (X s ) s S is a family of countably infinite sets X s of variables of sort s S. A Σ-congruence on a Σ-algebra A is (as usual) a family A = ( A,s ) s S of equivalence relations A,s on A s compatible with the operations of Σ. The class of all Σ-algebras is denoted by Alg(Σ). Together with Σ-homomorphisms this class forms a category, for simplicity also denoted by Alg(Σ). A signature morphism σ: Σ Σ between two signatures Σ = (S, OP) and Σ = (S, OP ) is a pair (σ sorts, σ opns ) of mappings σ sorts : S S, σ opns : OP OP such that for all (op: s 1,,s n s) OP, σ opns (op) has the functionality σ sorts (s 1 ),,σ sorts (s n ) σ sorts (s). In the following we omit the indices of σ. 1 Indeed our notion of an "observer" is a generalization of an "action" in the sense of [P 96]. 2

4 For any signature morphism σ: Σ Σ the reduct functor _ σ : Alg(Σ ) Alg(Σ) is defined as usual, i.e. for any Σ -algebra A, its reduct A σ is defined by (A σ ) s = def A σ(s) for all s S and op A σ = def σ(op) A for all op OP. The reduct of a relation ϕ A x B w.r.t. σ: Σ Σ is denoted by ϕ σ where ϕ σ A σ x B σ is defined by (ϕ σ ) s = def ϕ σ(s) for all s S. The set of first-order Σ-formulas is defined as usual from equations t = r (with t, r T(Σ, X)), the logical connectives,... and the quantifiers,. We will also use infinitary Σ-formulas of ^, the form φ i or φ i where φ i is a countable set of Σ-formulas. A finitary Σ-formula is a Σ-formula which i I i I contains no infinitary disjunction (conjunction resp.). A Σ-sentence is a Σ-formula which contains no free variables. The (standard) satisfaction of a Σ-formula φ by a Σ-algebra A, denoted by A = φ, is defined as usual in the first-order predicate calculus (with a straightforward extension to infinitary conjunctions and disjunctions, cf. e.g. [K 71]). The notation A = φ is extended in the usual way to classes of algebras and sets of formulas. A Σ-sentence φ is a semantic consequence of a set Φ of Σ-sentences, also denoted by Φ = φ, if for any Σ-algebra A with A = Φ we have A = φ. 3 The Category of Observational Algebras An observational signature is a generalization of a standard algebraic signature with a distinguished set of observable sorts (determining the carrier sets of the observable values) and with a distinguished set of observer operations (determining the experiments that can be used to distinguish non observable elements, often called "states"). An n-ary operation op: s 1,, s n s with several non observable argument sorts may also be used as an observer. In this case op is equipped with a "position number" 1 i n which indicates the argument sort of the states to be observed by op. For instance, if op: s 1, s 2 s is a binary operation then we can declare either (op, 1) or (op, 2) or both, (op, 1) and (op, 2), as observer thus obtaining as much flexibility as needed in practical examples. Definition 3.1 (Observational signature) Let Σ = (S, OP) be a signature and S Obs S be a set of observable sorts. An observer is a pair (op, i) where (op: s 1,, s n s) OP is an operation symbol such that 1 i n and s i S Obs. (op, i) is a direct observer of s i if s S Obs ; otherwise it is an indirect observer. If op: s 1 s is a unary observer we will simply write op instead of (op, 1). An observational signature Σ Obs = (Σ, S Obs, OP Obs ) consists of a signature Σ = (S, OP), a set S Obs S of observable sorts and a set OP Obs of observers (op, i) with op OP. Convention We implicitly assume in the following (if not stated otherwise) that whenever we consider an observational signature Σ Obs, then Σ Obs = (Σ, S Obs, OP Obs ) with Σ = (S, OP) and similarly for Σ Obs etc. Example 3.2 The following is a simple observational signature for bank accounts with observer "bal" determining the balance of an account. Here and in the following examples we use postfix notation for unary operations and infix notation for binary operations. sorts {account, int} observable sorts {int} observers { _.bal: account int} operations {new: account, _.change_ : account, int account} 3

5 A more advanced signature for bank accounts may be obtained by introducing also an indirect observer "_.redo: account account" intended to reconstruct the previous state of an account after having performed an action. Any observational signature determines a set of observable contexts which represent those experiments which allow us to distinguish elements by the given observers. In the above example the only observable context is "z account.bal". If we additionally use the indirect observer "redo" then there are infinitely many observable contexts of the form "z account.redo.redo.bal". Definition 3.3 (Σ Obs -context) Let Σ Obs be an observational signature, let X = (X s ) s S be the generally assumed family of countably infinite sets X s of variables of sort s and let Z = ({z s }) s S be a disjoint S-sorted family of singleton sets. For all s, s S the set C(Σ Obs ) s s of Σ Obs -contexts with "application sort" s and "result sort" s is inductively defined as follows: (1) For each s S, z s C(Σ Obs ) s s. (2) For each (op, i) OP Obs with op: s 1,, s n s, for each c C(Σ Obs ) s si and pairwise disjoint variables x 1,, x n of sort s 1,, s n not occurring in c, op(x 1,, x i-1, c, x i+1,, x n ) C(Σ Obs ) s s. Each context c C(Σ Obs ) s s contains, besides variables in X, exactly one occurrence of the "context variable" z s and this is the only variable of Z occurring in c. The application of a context c C(Σ Obs ) s s to a term t of sort s, denoted by c[t], is the term obtained by substituting the term t for z s. An observable Σ Obs -context is a Σ Obs -context with observable result sort s S Obs. We denote by C(Σ Obs ) s SObs the set of observable contexts with application sort s. Elements which cannot be distinguished by the experiments of an observational signature are considered to be observationally equal. Formally, this is expressed as follows. Definition 3.4 (Σ Obs -equality) Let Σ Obs be an observational signature. For any Σ- algebra A Alg(Σ) the observational Σ Obs -equality on A is denoted by Σ Obs,A and defined by: For all s S, two elements a, b A s are observationally equal w.r.t. Σ Obs, i.e. a Σ Obs,A b, if and only if for all observable contexts c C(Σ Obs) s SObs and for all valuations α, β: X {z s } A with α(x) = β(x) if x X, α(z s ) = a, β(z s ) = b, we have I α (c) = I β (c). Obviously, if s is an observable sort, then for all a, b A, a Σ Obs,A b is equivalent to a = b (since for any observable sort s, the context variable z s is a (trivial) observable context). For any Σ-algebra A, Σ Obs,A is an equivalence relation on A but, in general, not a Σ- congruence on A. Let us illustrate this by an example. 4

6 Example 3.5 Consider the following algebra over the signature of bank accounts with observer "bal": A account = INT x INT, new A = (0,0), (n,m).bal A = n, (n,m).change A (x) = (n+m+x,m). Obviously, the operation change A is not compatible with the observational equality determined by bal A since, for any x INT, change A (x) may transform indistinguishable states like e.g. (4,5) and (4,6) into distinguishable states (9+x,5) and (10+x,6). Of course, an "observationally correct" definition of change A would be given by (n,m).change A (x) = (n+x,m). In this paper we follow the loose semantics approach to algebraic specifications where a specification can be considered as a description of all admissible implementations (represented by the models of the specification). The basic assumption of the present approach is that, having given a set of observers, an implementation can only be admissible if it respects the observational equality determined by the observers. Formally, this is expressed by the following notion of an observational algebra. Definition 3.6 (Observational algebra) Let Σ Obs be an observational signature. An observational Σ Obs -algebra is a Σ-algebra A such that Σ Obs,A is a Σ-congruence on A.2 The class of all observational Σ Obs -algebras is denoted by Alg Obs (Σ Obs ). Remark 3.7 (1) Let us consider the following special case of a Σ Obs -equality. Assume given a signature Σ = (S, OP) and a set S Obs S of observable sorts. Then we can choose each (non constant) operation symbol op OP as an observer for any non observable argument sort of op, i.e. OP Obs = def {(op, i) (op: s 1,, s n s) OP, 1 i n and s i S Obs }. We denote the resulting observational signature simply by (Σ, S Obs ) (because in this case the set of observers OP Obs is uniquely determined by Σ). For any Σ-algebra A, the observational equality induced by (Σ, S Obs ) will be denoted by S Obs,A. Obviously, this observational equality is a Σ-congruence on A. Hence in this case any Σ-algebra is an observational algebra. In particular, S Obs,A coincides with the (total) observational equality of elements defined in [BHW 95] and similarly in other approaches in the literature like, for instance, in [R 87], [GM 97]. (2) Let us now point out the relationship to the coalgebraic framework (cf. e.g. [R 95], [JR 97]). For this purpose assume that Σ Obs = (Σ, S Obs, OP Obs ) is an observational signature such that for any observer (op, i) OP Obs the operation symbol op: s 1,, s n s has exactly one non observable argument sort and this is the sort s i. Moreover, assume that any observable sort s S Obs is interpreted in any observational Σ Obs -algebra by the same fixed set of observable values (e.g. integers, booleans etc.). Then a (polynomial) functor T: Set Set can be associated to OP Obs which captures the functionality of the observer symbols. 3 Any observational Σ Obs -algebra A is an extension of a T-coalgebra C which has the same carrier sets as C and defines the non observer operations op OP\OP Obs on top of C. The fact that the extension A is an observational algebra is equivalent to the fact that each operation of OP preserves bisimilarity of elements (cf. e.g. [JR 97]). 2 Obviously, for this it is sufficient that all non observer operations are compatible with Σ Obs,A. 3 If there is more than one non observable sort in S\SObs then a set of functors has to be associated to Σ Obs. 5

7 The following lemma is a direct consequence of Lemma 15 in [BH 94]. Lemma 3.8 Let Σ Obs be an observational signature. For any Σ-algebra A Alg(Σ), A Alg Obs (Σ Obs ) if and only if Σ Obs,A coincides with the observational equality SObs,A (cf. Remark 3.7 (1) above). In order to obtain a category of observational algebras we still need an appropriate morphism notion. Of course, since any observational algebra is a Σ-algebra, one could simply use standard homomorphisms between Σ-algebras. But this does not reflect any relationships between the observable behaviour of algebras. Therefore, we have chosen another definition where an observational homomorphism is defined as a relation which is compatible with observational behaviours (cf. Theorem 3.13, Corollary 3.14). Moreover, we will also see that the associated isomorphism notion allows us to extend Scott s theorem to observational algebras and observational satisfaction (cf. Corollary 4.5). Definition 3.9 (Observational homomorphism) Let A, B Alg Obs (Σ Obs ) be two observational Σ Obs -algebras. An observational Σ Obs -homomorphism ϕ: A B is an S- sorted family (ϕ s ) s S of relations ϕ s A s x B s with the following properties for all s S: (1) For all a A s there exists b B s such that a ϕ s b. (2) For all a A s, b, b B s, if a ϕ s b then (a ϕ s b if and only if b Σ Obs,B b ). (3) For all a, a A s, b B s, if a ϕ s b and a Σ Obs,A a then a ϕ s b. (4) For all (op: s 1,, s n s) OP and a i A si, b i B si, if a i ϕ si b i for i = 1,, n then op A (a 1,, a n ) ϕ s op B (b 1,, b n ). Remark 3.10 (1) Let Σ Obs be an observational signature such that S Obs = S, i.e. all sorts are observable. Then each Σ-algebra A is an observational algebra and Σ Obs,A is the set-theoretic equality of elements. Hence for any two Σ-algebras A and B, ϕ: A B is an observational homomorphism if and only if ϕ is a (standard) Σ-homomorphism. For arbitrary observational signatures Σ Obs any standard homorphism h: A B induces also an observational homomorphism ϕ h : A B defined by ϕ h = def {(a, b) a Σ Obs,A a, h(a ) = b and b Σ Obs,A b for some a A, b B}. (2) For any two observational Σ Obs -algebras A and B, ϕ: A B is an observational homomorphism if and only if ϕ is representable by a span g: C A, h: C B (cf. [BT 95]) in the category of Σ-algebras such that g is surjective, h(c) is closed under Σ Obs,B, Ker(π B o h) = Ker(g) and Ker(π A o g) Ker(h) where π A : A A/ Σ Obs,A and π B: B B/ Σ Obs,B are the canonical epimorphisms. Theorem 3.11 (The category of observational algebras) For each observational signature Σ Obs, the class Alg Obs (Σ Obs ) together with the observational Σ Obs -homomorphisms is a category which, by abuse of notation, will also be denoted by Alg Obs (Σ Obs ). 6

8 Thereby the composition of observational homomorphisms is the usual composition of relations, i.e. for any two observational Σ Obs -homomorphisms ϕ: A B and ψ: B C, the composition ψ o ϕ is the family (ψ s o ϕ s ) s S where ψ s o ϕ s = def {(a, c) a A s, c C s, there exists b Β s with a ϕ s b and b ψ s c}. For each observational Σ Obs -algebra A, the identity id A : A A in the category Alg Obs (Σ Obs ) is the observational equality Σ Obs,A. (For the proof see Appendix.) For any observational Σ Obs -algebra A the observational equality Σ Obs,A is a Σ- congruence. Hence we can construct the quotient algebra A/ Σ Obs,A which identifies all elements of A which are "indistinguishable" by the experiments built with the observers of Σ Obs. A/ Σ Obs,A can be considered as the "black box view" of A thus representing the "observable behaviour" of A w.r.t. Σ Obs. Using this behaviour construction one can relate, for any observational signature Σ Obs, the category Alg Obs (Σ Obs ) and the category Alg(Σ) of Σ-algebras by a functor which associates to any observational algebra its observational behaviour. For morphisms this functor establishes even a one to one correspondence between observational homomorphisms ϕ: A B and standard homomorphisms h: A/ Σ Obs,A B/ ΣObs,B between the observable behaviours of A and B, i.e. the functor is full and faithful. Definition 3.12 (Behaviour functor) For any observational signature Σ Obs, F Σ Obs : Alg Obs(Σ Obs ) Alg(Σ) is defined by: For each A Alg Obs (Σ Obs ), F Σ Obs (A) = def A/ Σ Obs,A, for each observational Σ Obs -homomorphism ϕ: A B, F Σ Obs (ϕ): A/ ΣObs,A B/ ΣObs,B is defined by F ΣObs (ϕ)([a]) = def [b] if a ϕ b. For each A Alg Obs (Σ Obs ), F Σ Obs (A) is called the observational behaviour of A. Theorem 3.13 For any observational signature Σ Obs, F Σ Obs : Alg Obs(Σ Obs ) Alg(Σ) is a full and faithful functor. (For the proof see Appendix.) As a useful application of Theorem 3.13 which also points out the adequacy of our morphism notion for observational algebras we obtain that two observational algebras are observationally isomorphic if and only if they have isomorphic behaviours. The proof of this fact is standard since F Σ Obs is full and faithful. Corollary 3.14 (Observational isomorphism) Let A, B Alg Obs (Σ Obs ) be two observational Σ Obs -algebras. There exists an observational Σ Obs -isomorphism ϕ: A B if and only if F Σ Obs (A) and F ΣObs (B) are isomorphic Σ-algebras. 7

9 Remark 3.15 (1) According to Corollary 3.14 two observationally isomorphic algebras can be considered to be observationally equivalent since they have the same observable behaviour (up to standard isomorphism). In the literature explicit definitions of observational equivalences of algebras have been studied, for instance, in [R 87]. As a consequence of a result in [BHW 95] 4 these equivalence notions coincide with our notion of observational isomorphism for observational algebras. (2) In [Sch 90] a characterization of observational equivalence of algebras was given by means of correspondence relations. Indeed any observational isomorphism (being the identity on observable sorts) is a strong V-correspondence in the sense of [Sch 90] and, conversely, the closure of a strong V-correspondence under observational equality of elements is an observational isomorphism. (3) A similar relationship holds between observational isomorphisms and bisimulations of the coalgebraic framework if we assume the conditions described in Remark 3.7 (2). In particular, an observational isomorphism between two observational algebras A1 and A2 (being the identity on observable sorts) is just the greatest bisimulation between the underlying coalgebras C1 and C2; cf. Remark 3.7 (2). Analogously to the notion of an abstract data type in [LEW 96] abstract behaviour types are defined as follows. Definition 3.16 A class C Alg Obs (Σ Obs ) which is closed under observational Σ Obs -isomorphism is called an abstract behaviour type. 4 Observational Satisfaction The underlying idea of the observational satisfaction relation is to interpret the equality symbol "=" occurring in a first-order formula φ not by the set-theoretic equality but by the observational equality of elements. Hence the following definition is quite similar to the definition of the standard satisfaction relation. The only difference concerns (1) where "I α (t) = I α (r)" is replaced by "I α (t) Σ Obs,A I α(r)". Definition 4.1 (Observational satisfaction relation) The observational satisfaction relation between observational Σ Obs -algebras and Σ-formulas is denoted by = Σ Obs and defined as follows: Let A Alg Obs (Σ Obs ). (1) For any two terms t, r T(Σ, X) s of the same sort s and for any valuation α: X A, A, α = Σ Obs t = r holds if I α(t) Σ Obs,A I α(r). (2) For any arbitrary Σ-formula φ and for any valuation α: X A, A, α = Σ Obs φ is defined by induction over the structure of the formula φ in the usual way. (3) For any arbitrary Σ-formula φ, A = Σ Obs φ holds if for all valuations α: X A, A, α = Σ Obs φ holds. 4 It is shown in [BHW 95] that observational equivalences of algebras are factorizable by observational equalities (of elements). 8

10 The notation A = Σ Obs φ is extended in the usual way to classes of observational algebras and sets of formulas. Remark 4.2 Technically the observational satisfaction relation could be defined in the same way for arbitrary Σ-algebras which do not necessarilly belong to Alg Obs (Σ Obs ). But then the observational satisfaction of equations may not be compatible with (some) operations of the signature, i.e. A = Σ Obs t = r may not imply A = ΣObs op(t) = op(r) for some operation op. Obviously this cannot happen in our approach. In [DF 98] this problem is tackled by restricting the set of formulas for which the observational satisfaction relation is defined. But then it is not possible to study obervational properties which involve non observer operations. 5 Definition 4.3 (Observational consequence) A Σ-sentence φ is an observational consequence of a set Φ of Σ-sentences, also denoted by Φ = Σ Obs φ, if for any observational Σ Obs -algebra A, A = Σ Obs Φ implies A = ΣObs φ. The next proposition shows that the behaviour functor defined in Theorem 3.13 is compatible with the observational and the standard satisfaction relations. Proposition 4.4 For any A Alg Obs (Σ Obs ) and any Σ-formula φ, A = Σ Obs φ if and only if F ΣObs (Α) = φ. Proof: Follows from Theorem 3.11 of [BHW 95] and from the fact that for observational algebras A, Σ Obs,A coincides with SObs,A (cf. Lemma 3.8). As a consequence of Corollary 3.14 and Proposition 4.4 we can generalize Scott s theorem (cf. e.g. [K 71]) to observational algebras and observational satisfaction (taking into account that Σ-formulas may be infinitary; cf. Section 2). 6 Corollary 4.5 (Observational version of Scott s theorem) Let A, B Alg Obs (Σ Obs ) be two observational Σ Obs -algebras such that F Σ Obs (Α) and F Σ Obs (Β) are countable. The following conditions are equivalent: (1) There exists an observational Σ Obs -isomorphism ϕ: A B. (2) For all (possibly infinitary) Σ-formulas φ, A = Σ Obs φ if and only if B = ΣObs φ. Proof: (1) (2): Is a direct consequence of Corollary 3.14 and Proposition In [DF 98] some hints are given that one may still extend the observational satisfaction relation to arbitrary sentences. However, then the congruence rule of the equational calculus is only sound if one can prove that all operations of a specification are "behaviourally coherent" which in our terminology would mean that all models of the specification are indeed observational algebras. 6 It may be interesting to note that this generalization relies only on the fact that the functor F Σ Obs is full and faithful and compatible with the satisfaction relations of the involved categories. Hence analogous generalizations of Scott s theorem can always be achieved if these assumptions are satisfied. 9

11 (2) (1): Assume (2). Then, by Proposition 4.4, for all Σ-formulas φ, F Σ Obs (Α) = φ iff F Σ Obs (Β) = φ. Since F ΣObs (Α) and F ΣObs (Β) are countable we can now apply Scott s theorem (cf. [K 71]) and obtain that F Σ Obs (Α) and F ΣObs (Β) are isomorphic Σ-algebras. Then we obtain (1) by applying Corollary The above characterization of observationally isomorphic algebras underlines the adequacy of our observational homomorphism notion. A related result, but formulated in terms of observational equivalence of algebras instead of observational isomorphism, was given in [BHW 95]. We are now able to define the syntax and semantics of flat observational specifications. Structured specifications will be considered in Section 7. Definition 4.6 (Flat observational specification) A flat observational specification SP = Σ Obs, Ax consists of an observational signature Σ Obs = (Σ, S Obs, OP Obs ) and a set Ax of Σ-sentences, called the axioms of SP. The semantics of SP is given by its signature Sig Obs (SP) and by its class of models Mod Obs (SP) which are defined by Sig Obs (SP) = def Σ Obs, Mod Obs (SP) = def {A Alg Obs (Σ Obs ) A = Σ Obs Ax}. For any observational specification SP, the class Mod Obs (SP) is closed under observational isomorphisms. Remark 4.7 If we consider the special case where all operation symbols are used as observers (cf. Remark 3.7(1)) then our semantical definition coincides with the one in [BHW 95] (for total observational equalities). Moreover, if we assume given a fixed universe of observable data and if we restrict to equational axioms then the model class of a flat observational specification coincides with the model class of a hidden specification as defined in [GM 97]. Example 4.8 The following specification of bank accounts has additionally to the account operations of Example 3.2 an operation "paycharge" which reduces the balance of an account by a constant monthly fee. spec ACCOUNT = sorts {account, int} observable sorts {int} observers { _.bal: account int, _.redo_ : account account} operations "operations for the integers" {new: account, _.change_ : account, int account, _.paycharge_ : account account} 10

12 axioms "axioms for the integers" 7 { x: int, s: account. new.bal = 0, new.redo = new, s.change(x).bal = s.bal+x, s.change(x).redo = s, s.paycharge.bal = s.bal-10, s.paycharge.redo = s} A possible model of the specification ACCOUNT which satisfies the axioms even literally can be defined in terms of lists of integers. Another model which satisfies the axioms observationally (but not literally) can be constructed by using the well-known array with pointer realization of lists. (Note that in this example all algebras which satisfy (observationally) the axioms are already observational algebras. In general this is obviously not true, for instance, if an operation is specified very loosely.) 5 The Institution of Observational Logic The category of observational algebras is the basis for defining an institution of observational logic which captures the model-theoretic view of the observable behaviour of systems. For the underlying notions of institutions we refer, for instance, to [ST 88]. An essential ingredient which, up to now, is still missing is an appropriate morphism notion for observational signatures. In the following definition we utilize again the flexibility provided by our notion of an observer. Definition 5.1 (Observational signature morphism) Let Σ Obs = (Σ, S Obs, OP Obs ) and Σ Obs = (Σ, S Obs, OP Obs ) be two observational signatures with Σ = (S, OP) and Σ = (S, OP ). An observational signature morphism σ Obs : Σ Obs Σ Obs is a signature morphism σ: Σ Σ such that the following conditions are satisfied: (1) σ(s Obs ) = σ(s) S Obs. (2) If (op, i) OP Obs then (σ(op), i) OP Obs, (3) If (op, i) OP Obs such that op : s 1,,s n s and s i σ(s) then there exists op OP such that (op, i) OP Obs and op = σ(op). Condition (1) is standard. It requires that observable and non observable sorts are preserved by σ. Condition (2) requires that also observers are preserved by σ. Condition (3) is essential for the satisfaction condition presented below. It says that whenever the image σ(s) of some "old" sort s of Σ Obs is observed by an observer op of Σ Obs then there must be a corresponding observer op of Σ Obs which observes s and which is mapped to op. Thus no "new" observations can be introduced for "old" sorts. However, it is important to note that this still allows us to introduce new operation symbols in Σ Obs which are not in the image of σ but nevertheless may have argument sorts which are in the image of σ. A standard example may be given by the signature of a 7 The axioms of the integer operations are as usual. For expressing the generation principle of the integers one may use either an infinitary disjunction (cf. also [K 71]) or "generated by" axioms which can easily be incorporated in our approach; cf. Conclusion. Another possibility is, of course, to assume a fixed universe of visible data like in hidden algebra; cf. e.g. [GM 97]. 11

13 specification of a bank which is based (i.e. imports) a specification of accounts. Then the bank specification may introduce an operation "_.add_: bank, account bank" for adding an account to a bank. This is not a problem as long as "add" is not used as an observer for accounts which indeed would be strange. Whether "add" is used as an observer for banks or not is irrelevant for establishing an observational signature inclusion from the account signature into the bank signature. Examples like this, where some argument sorts of an operation are imported from another signature frequently occur in practice, in particular, in object oriented programming. This situation cannot be dealt with by hidden signature morphisms; cf. e.g. [MG 94]. Convention We implicitly assume in the following that whenever we consider an observational signature morphism σ Obs : Σ Obs Σ Obs then the underlying signature morphism is σ: Σ Σ. Lemma 5.2 Observational signatures and observational signature morphisms form a category which has pushouts. (For the proof see Appendix.) Definition 5.3 (Observational reduct functor) For any observational signature morphism σ Obs : Σ Obs Σ Obs, Alg Obs (σ Obs ): Alg Obs (Σ Obs ) Alg Obs (Σ Obs ) is defined by: For each A Alg Obs (Σ Obs ), Alg Obs (σ Obs )(A ) = def A σ. For each observational Σ Obs -homomorphism ϕ : A B, Alg Obs (σ Obs )(ϕ ): A σ B σ is defined by Alg Obs(σ Obs )(ϕ ) = def ϕ σ. (See Section 2 for the definition of the reducts A σ and ϕ σ.) The following lemma is essential for proving that Alg Obs (σ Obs ) is indeed a well-defined functor (cf. Theorem 5.5) and also for checking the (observational) satisfaction condition (cf. Theorem 5.6). It says that the observational equality is preserved by observational reducts. (The proof is given in the Appendix.) Lemma 5.4 For any observational signature morphism σ Obs : Σ Obs Σ Obs and observational Σ Obs -algebra A Alg Obs (Σ Obs ), ( Σ Obs,A ) σ = ΣObs,(A σ). Theorem 5.5 For any observational signature morphism σ Obs : Σ Obs Σ Obs, Alg Obs (σ Obs ): Alg Obs (Σ Obs ) Alg Obs (Σ Obs ) is a well-defined functor. (For the proof see Appendix.) We are now able to state the satisfaction condition for observational logic. It guarantees encapsulation in the sense that observational properties are respected when composing systems. In Section 7 we will consider structured specifications and we will see how a 12

14 straightforward sound and complete proof system for structured specifications can be constructed which needs the validity of the satisfaction condition. Theorem 5.6 (Observational satisfaction condition) Let σ Obs : Σ Obs Σ Obs be an observational signature morphism. For any A Alg Obs (Σ Obs ) and Σ-sentence φ, A = Σ Obs σ(φ) if and only if Alg Obs(σ Obs )(A ) = Σ Obs φ where σ(φ) is the usual extension of a signature morphism to Σ-sentences. (For the proof see Appendix.) Corollary 5.7 (The institution of observational logic) The quadruple INS Obs = (Sig Obs, Sen IFOLEQ, Alg Obs, = Obs ) is an institution whereby: Sig Obs is the category of observational signatures and observational signature morphisms. The functor Sen IFOLEQ : Sig Obs Set maps - each observational signature Σ Obs = (Σ, S Obs, OP Obs ) to the set of (possibly infinitary) many-sorted first-order Σ-sentences (cf. Section 2) and - each observational signature morphism σ Obs : Σ Obs Σ Obs to the obvious translation function which transforms Σ-sentences into Σ -sentences. The functor Alg Obs : (Sig Obs ) op Cat maps - each observational signature Σ Obs to the category Alg Obs (Σ Obs ) of observational Σ Obs -algebras and observational Σ Obs -homomorphisms and - each observational signature morphism σ Obs : Σ Obs Σ Obs to the observational reduct functor Alg Obs (σ Obs ): Alg Obs (Σ Obs ) Alg Obs (Σ Obs ). = Obs = ( = ΣObs ) ΣObs Sig Obs where, for each observational signature Σ Obs, = ΣObs is the observational satisfaction relation of a Σ-sentence by an observational Σ Obs - algebra. Remark 5.8 In Section 3 a family (F Σ Obs ) ΣObs SigObs of (full and faithful) functors F Σ Obs : Alg Obs(Σ Obs ) Alg(Σ) is defined and in Proposition 4.4 we have shown that these functors are compatible with observational and standard satisfaction. Hence, the family of functors F Σ Obs can be extended to an institution morphism (cf. [T 96]) which maps the institution of observational logic to the institution of (standard) infinitary first order logic. 6 A Proof System for Observational Logic In this section we study the proof theory for observational logic. For defining an appropriate proof system we first associate to any observational signature Σ Obs the following set FA Σ Obs of Σ-sentences. 13

15 Definition 6.1 Let Σ Obs = (Σ, S Obs, OP Obs ) be an observational signature with Σ = (S, F). FA Σ Obs = def {FA Σ Obs (s) s S\S Obs} where for any s S\S Obs, FA Σ Obs (s) = def x L, x R :s. Var(c). c[x L ] = c[y R ] x L = x R. c C(ΣObs) s SObs Thereby Var(c) denotes the set of all variables occurring in c besides the context variable z s. The underying idea for considering FA Σ Obs stems from a result in [BHW 95] where it is shown that the behavioural theory of a class C of Σ-algebras coincides with the standard theory of the fully abstract algebras of C. Indeed it is easy to prove (using Lemma 3.8) that an observational Σ Obs -algebra satisies FA Σ Obs if and only if it is fully abstract. (Note, however, that this does not hold for arbitrary Σ-algebras.) The following theorem shows that the sentences FA Σ Obs allow us to characterize the observational consequence relation in terms of the standard consequence relation. (The proof is given in the Appendix.) Theorem 6.2 Let Σ Obs be an observational signature, Φ be a set of Σ-sentences and φ be a Σ-sentence. Φ = ΣObs φ if and only if Φ FA Σ Obs = φ. Remark 6.3 Theorem 6.2 shows that there exists a correspondence between observational theories and standard theories which can be expressed by a map that associates to an observational theoy (Σ Obs, Φ) the standard theory (Σ, Φ FA Σ Obs ) in the sense of a plain institution map in [C 93]. In the sequel we assume given, for each signature Σ, a sound and complete proof system Π(Σ) for (many-sorted) first-order logic. The proof system Π(Σ Obs ) for observational logic is then constructed by adding to the axioms and rules of Π(Σ) the sentences FA Σ Obs as further axioms. Definition 6.4 (Proof system for observational logic) For any observational signature Σ Obs, Π(Σ Obs ) = def Π(Σ) FA Σ Obs. We write Φ ΣObs φ (Φ Σ φ resp.) if φ is a Σ-sentence that can be deduced from a set Φ of Σ-sentences by the axioms and rules of Π(Σ Obs ) (Π(Σ) resp.). Using Theorem 6.2 we can prove the soundness and completeness of Π(Σ Obs ). Corollary 6.5 (Soundness and completeness) For any observational signature Σ Obs, for any set Φ of Σ-sentences and any Σ-sentence φ, Φ ΣObs φ if and only if Φ = ΣObs φ. Proof: Φ ΣObs φ iff, by definition of Π(Σ Obs ), Φ FA Σ Obs Σ φ iff, by soundness and completeness of Π(Σ), Φ FA Σ Obs = φ iff, by Theorem 6.2, Φ = Σ Obs φ. 14

16 Remark 6.6 (1) If Σ Obs contains indirect observers there may be infinitely many observable contexts and then FA Σ Obs contains infinitary conjunctions. In this case we choose for Π(Σ) a proof system for infinitary first-order logic (for instance, the manysorted variant of the proof system in [K 71]). 8 Otherwise FA Σ Obs is finitary (if contexts which are equal up to α-conversion are removed). Then one can choose a formal proof system for finitary first-order logic (for instance, the one of [M 87]) to achieve the above result (provided that Φ is finitary). In particular, any available theorem prover for firstorder logic can be used to prove that φ is an observational consequence of Φ. 9 (2) The axioms FA Σ Obs can be considered as a coinductive proof principle (cf. e.g. [JR 97]) which, together with Π(Σ), allows us to prove the observational validity not only of equations but of arbitrary first-order formulas. The essential difference to the proof technique proposed in [BH 94] is that there the observational semantics was not defined in terms of observational algebras and therefore it was necessary to check that the observational equality induced by the observers of Σ Obs is a Σ-congruence on the algebras under consideration (cf. also the hidden coinduction principle for behavioural equations in [GM 97]). Example 6.7 Consider the signature of the observational specification ACCOUNT of Example 4.8. It induces the infinitary sentence FA(account) = def s L, s R : account. i N s L.redo i.bal = s R.redo i.bal s L = s R. Now consider the implicitly universally quantified equation s.paycharge = s.change(-10). It is easy to prove by induction that for all i N, s.paycharge.redo i.bal = s.change(-10).redo i.bal can be derived from the axioms of ACCOUNT. Then, using FA(account), we deduce s.paycharge = s.change(-10) and therefore, by Corollary 6.5, this equation is an observational consequence of the ACCOUNT specification. 7 Structured Observational Specifications In [B 98] (and similarly in [ST 88]) a basic set of specification building operations is defined which allows one to build structured specifications over an arbitrary institution. We will now apply these operators to the particular institution of observational logic thus obtaining the following set of operations for constructing structured observational specifications. The semantics of such a specification SP is determined by its 8 In this case it is however possible to apply a result of [BH 96] to encode the infinitary formulas FA Σ Obs by finitary ones using auxiliary (hidden) symbols and reachability constraints. 9 For instance, using the Larch Prover one can directly implement the axioms FA Σ Obs by the "partitioned by" construct of LP; cf. [GH 93]. 15

17 (observational) signature, denoted by Sig Obs (SP), and by its class of models, denoted by Mod Obs (SP). In the following definition we assume that σ Obs : Σ Obs Σ Obs is an observational signature morphism such that its underlying signature morphism σ: Σ Σ is injective. 10 basic: union: translate: derive: Any presentation Σ Obs, Ax is an observational specification. Its semantics is defined in Definition 4.6. For any two observational specifications SP1 and SP2 with Sig Obs (SP1) = Sig Obs (SP2), the expression SP1 SP2 is an observational specification with semantics Sig Obs (SP1 SP2) = def Sig Obs (SP1), Mod Obs (SP1 SP2) = def Mod Obs (SP1) Mod Obs (SP2). For any observational specification SP with Sig Obs (SP) = Σ Obs, the expression translate SP by σ Obs is an observational specification with semantics Sig Obs (translate SP by σ Obs ) = def Σ Obs, Mod Obs (translate SP by σ Obs ) = def {A Alg Obs (Σ Obs ) Alg Obs (σ Obs )(A ) Mod Obs (SP)}. For any observational specification SP with Sig Obs (SP ) = Σ Obs, the expression derive from SP by σ Obs is an observational specification with semantics Sig Obs (derive from SP by σ Obs ) = def Σ Obs, Mod Obs (derive from SP by σ Obs ) = def {Alg Obs (σ Obs )(A ) A Mod Obs (SP )}. Definition 7.1 Let SP be an observational specification with signature Σ Obs. A Σ-sentence φ is called an observational theorem of SP, written SP = ΣObs φ, if Mod Obs (SP) = ΣObs φ. In the following we are interested in a proof system which allows us to prove observational theorems of structured (observational) specifications. For this purpose we instantiate the institution independent proof system of [B 98] and obtain the following rules which generate, for each observational signature Σ Obs, a relation ΣObs between observational specifications SP and Σ-sentences φ. φ Ax (basic) Σ, Ax ΣObs φ SP1 ΣObs φ SP2 ΣObs φ (union-1) (union-2) SP1+SP2 ΣObs φ SP1+SP2 ΣObs φ SP ΣObs φ (translate) translate SP by σ Obs Σ Obs σ(φ) 10 The injectivity requirement ensures that the interpolation property for institutions (cf. [B 98]) needed for the completeness proof is valid. 16

18 SP Σ Obs σ(φ) (derive) derive from SP by σ Obs ΣObs φ SP ΣObs φ i for i I, {φ i i I} ΣObs φ (pi-obs) SP ΣObs φ According to the rule (pi-obs) the proof system for structured specifications is based on the proof system Π(Σ Obs ) for observational logic presented in Section 6. The other rules correspond to the specification building operations and hence proofs of observational theorems can be performed according to the structure of a given specification. An institution independent proof of the soundness of the above rules is presented in [ST 88]. The completeness can be checked by applying the results of [B 98] to the institution INS Obs of observational logic. For this purpose one has to show that INS Obs satisfies the amalgamation and interpolation properties which is detailed in the Appendix. Theorem 7.2 (Soundness and completeness) Let SP be an observational specification with signature Σ Obs = (Σ, S Obs, OP Obs ) and let φ be a Σ-sentence. SP = ΣObs φ if and only if SP ΣObs φ. 8 Conclusion The main topic of our next research steps is the consideration of refinement relations between (structured) observational specifications with an emphasis on refinement proofs. We hope that we can reuse several results of [BH 95] and [H 97] but we are aware that these approaches do not deal with a built in (internalised) observational semantics of structured specifications as considered in this paper. Easy further tasks should be, first, to extend our approach to partial observational equalities (in the spirit of [BHW 95]) thus obtaining a means to forget irrelevant junk elements and, secondly, to incorporate reachability constraints (or "generated by" axioms) for expressing generation principles. Such generation principles may be useful not only for observable data but also for states, because in reality any state is reachable from the initial state by a finite sequence of transitions. Other important directions of future research deal with a specification methodology based on observational logic, with observational term rewriting and with an extension to concurrent systems. 17

19 References [B 98] T. Borzyszkowski: Completeness of a logical system for structured specifications. In: F. Parisi Presicce (ed.): Recent Trends in Data Type Specification, LNCS 1376, , [BB 91] G. Bernot, M. Bidoit: Proving the correctness of algebraically specified software: modularity and observability issues. Proc. AMAST 91, , Springer-Verlag Works. in Comp. Series, [BH 94] M. Bidoit, R. Hennicker: Proving behavioural theorems with standard first-order logic. In Proc. ALP 94, LNCS 850, 41-58, [BH 95] M. Bidoit, R. Hennicker: Proving the correctness of behavioural implementations. In Proc. AMAST 95, LNCS 936, , An extended version entitled "Modular correctness proofs of behavioural implementations" appears in Acta Informatica. [BH 96] M. Bidoit, R. Hennicker: Behavioural theories and the proof of behavioural properties. Theoretical Computer Science 175, 3-55, [BHW 95] M. Bidoit, R. Hennicker, M. Wirsing: Behavioural and abstractor specifications. Science of Computer Programming 25, , [BT 95] M. Bidoit, A. Tarlecki: Behavioural satisfaction and equivalence in concrete model categories. Technical Report, Warsaw University and Ecole Normale Supérieure, [C 93] M. Cerioli: Relationships between logical formalisms. Ph.D. thesis, Univ. di Pisa-Genova-Udine, [DF 98] R. Diaconescu, K. Futatsugi: CafeOBJ Report, Version , [GH 93] J. Guttag, J. Horning: Larch: Languages and Tools for Formal Specification. Texts and Monographs in Computer Science, Springer, [GM 97] J. Goguen, G. Malcolm: A hidden agenda. Report CS97-538, Univ. of Calif. at San Diego, [H 97] R. Hennicker: Structured specifications with behavioural operators: semantics, proof methods and applications. Habilitation thesis, Institut für Informatik, Universität München (LMU), [HN 93] R. Hennicker, F. Nickl: A behavioural algebraic framework for modular system design with reuse. In: H. Ehrig, F. Orejas (eds.): Recent Trends in Data Type Specification, LNCS 785, , [JR 97] B. Jacobs, J. Rutten: A Tutorial on (Co)Algebras and (Co)Induction. EATCS Bulletin 62, , [K 71] H. J. Keisler: Model theory for infinitary logic. North-Holland, [LEW 96] J. Loeckx, H.-D. Ehrich, M. Wolf: Specification of Abstract Data Types. Wiley and Teubner, [M 87] E. Mendelson: Introduction to Mathematical Logic. Math. Series. Wadswort & Brooks/Cole, [MG 94] G. Malcolm, J. A. Goguen: Proving correctness of refinement and implementation. Technical Monograph PRG-114, Oxford University Computing Laboratory, [P 96] P. Padawitz: Swinging data types: syntax, semantics, and theory. In: M. Haveraaen, O. Owe, O.-J. Dahl (eds.): Recent Trends in Data Type Specification, LNCS 1130, , [R 87] H. Reichel: Initial computability, algebraic specifications, and partial algebras. International Series of Monographs in Computer Science No. 2, Oxford: Clarendon Press, [R 95] H. Reichel: An approach to object semantics based on terminal co-algebras. Math. Struct. Comp. Sci., 5, , [Sch 90] O. Schoett: Behavioural correctness of data representation. Sci. of Comp. Prog. 14, 43-57, [ST 88] D. T. Sannella, A. Tarlecki: Specifications in an arbitrary institution. Information and Computation 76, , [T 96] A. Tarlecki: Moving between logical systems. In: M. Haveraaen, O. Owe, O.-J. Dahl (eds.): Recent Trends in Data Type Specification, LNCS 1130, ,

20 Appendix Proof of Theorem 3.11: It is straightforward to prove that the composition of observational homomorphisms yields an observational homomorphism. Obviously, the composition is associative. To show that Σ Obs,A is the identity on any observational Σ Obsalgebra A, one has first to check that Σ Obs,A is an observational homomorphism between A and A. Properties (1) - (3) of Definition 3.9 are satisfied because Σ Obs,A is an equivalence relation on A. Property (4) is satisfied since Σ Obs,A is a Σ-congruence for any observational Σ Obs -algebra A. Then it is straightforward to prove that for any observational Σ Obs -homomorphisms ϕ: A B and ψ: C A we have ϕ o Σ Obs,A = ϕ and Σ o Obs,A ψ = ψ. Thus Σ Obs,A is the identity for any A Alg Obs(Σ Obs ). Proof of Theorem 3.13: Using the properties of observational homomorphisms it is straightforward to prove that F Σ Obs (ϕ) is a well-defined Σ-homomorphism. It is also straightforward to prove that F Σ Obs is compatible with the composition of morphisms and preserves identities. Thus F Σ Obs is a functor. To show that F Σ Obs is full and faithful it is sufficient to prove that for any A, B Alg Obs (Σ Obs ) and for any Σ-homomorphism h: A/ Σ Obs,A B/ ΣObs,B there exists a unique observational Σ Obs -homomorphism ϕ: A B such that F Σ Obs (ϕ) = h. Assume that h: A/ Σ Obs,A B/ ΣObs,B is a Σ-homomorphism. Then h induces a relation ϕ A x B such that for all a A, b B, a ϕ b if and only if h([a]) = [b]. It is esay to prove that ϕ is indeed an observational Σ Obs -homomorphism between A and B such that F Σ Obs (ϕ) = h. Let us now prove the uniqueness of ϕ. Assume that ϕ : A B is an observational Σ Obs -homomorphism such that F Σ Obs (ϕ ) = h. We have to show ϕ = ϕ. But this is simple since for any a A and b B, a ϕ b iff F Σ Obs (ϕ)[a] = [b] iff h([a]) = [b] iff F Σ Obs (ϕ )[a] = [b] iff a ϕ b. Proof of Lemma 5.2: Let σ1 Obs : Σ Obs Σ1 Obs and σ2 Obs : Σ Obs Σ2 Obs be observational signature morphisms with underying signature morphisms σ1: Σ Σ1 and σ2: Σ Σ2 resp. It is well-known that in the category of algebraic signatures there exists a pushout σ1 Σ > Σ1 σ2 σ1 σ2 Σ2 > Σ Now let S Obs = σ1 (S1 Obs ) σ2 (S2 Obs ), OP Obs = {(σ1 (op1), i) (op1, i) OP1 Obs } {(σ2 (op2), i) (op2, i) OP2 Obs } and Σ Obs = (Σ, S Obs, OP Obs ). 19

Observational Logic. Rolf Hennicker*, Michel Bidoit**

Observational Logic. Rolf Hennicker*, Michel Bidoit** Observational Logic Rolf Hennicker*, Michel Bidoit** *Institut für Informatik, Ludwig-Maximilians-Universität München Oettingenstr. 67, D-80538 München, GERMANY **Laboratoire Spécification et Vérification,

More information

On the Duality between Observability and Reachability

On the Duality between Observability and Reachability On the Duality between Observability and Reachability Michel Bidoit 1, Rolf Hennicker 2, and Alexander Kurz 3 1 Laboratoire Spécification et Vérification (LSV), CNRS & ENS de Cachan, France 2 Institut

More information

Behavioural theories and the proof of. LIENS, C.N.R.S. U.R.A & Ecole Normale Superieure, 45 Rue d'ulm, F{75230 Paris Cedex 05, France

Behavioural theories and the proof of. LIENS, C.N.R.S. U.R.A & Ecole Normale Superieure, 45 Rue d'ulm, F{75230 Paris Cedex 05, France Behavioural theories and the proof of behavioural properties Michel Bidoit a and Rolf Hennicker b b a LIENS, C.N.R.S. U.R.A. 1327 & Ecole Normale Superieure, 45 Rue d'ulm, F{75230 Paris Cedex 05, France

More information

Interpolation in Logics with Constructors

Interpolation in Logics with Constructors Interpolation in Logics with Constructors Daniel Găină Japan Advanced Institute of Science and Technology School of Information Science Abstract We present a generic method for establishing the interpolation

More information

Theoretical Computer Science

Theoretical Computer Science Theoretical Computer Science 433 (202) 20 42 Contents lists available at SciVerse ScienceDirect Theoretical Computer Science journal homepage: www.elsevier.com/locate/tcs An axiomatic approach to structuring

More information

Relating Abstract Datatypes and Z-Schemata

Relating Abstract Datatypes and Z-Schemata Relating Abstract Datatypes and Z-Schemata Hubert Baumeister University of Munich, Institute of Computer Science, Oettingenstr. 67, D-80358 Munich, Germany baumeist@informatik.uni-muenchen.de Abstract.

More information

Quasi-Boolean Encodings and Conditionals in Algebraic Specification

Quasi-Boolean Encodings and Conditionals in Algebraic Specification Quasi-Boolean Encodings and Conditionals in Algebraic Specification Răzvan Diaconescu Institute of Mathematics Simion Stoilow of the Romanian Academy Abstract We develop a general study of the algebraic

More information

Equational Logic. Chapter Syntax Terms and Term Algebras

Equational Logic. Chapter Syntax Terms and Term Algebras Chapter 2 Equational Logic 2.1 Syntax 2.1.1 Terms and Term Algebras The natural logic of algebra is equational logic, whose propositions are universally quantified identities between terms built up from

More information

Löwenheim-Skolem Theorems, Countable Approximations, and L ω. David W. Kueker (Lecture Notes, Fall 2007)

Löwenheim-Skolem Theorems, Countable Approximations, and L ω. David W. Kueker (Lecture Notes, Fall 2007) Löwenheim-Skolem Theorems, Countable Approximations, and L ω 0. Introduction David W. Kueker (Lecture Notes, Fall 2007) In its simplest form the Löwenheim-Skolem Theorem for L ω1 ω states that if σ L ω1

More information

ACLT: Algebra, Categories, Logic in Topology - Grothendieck's generalized topological spaces (toposes)

ACLT: Algebra, Categories, Logic in Topology - Grothendieck's generalized topological spaces (toposes) ACLT: Algebra, Categories, Logic in Topology - Grothendieck's generalized topological spaces (toposes) Steve Vickers CS Theory Group Birmingham 2. Theories and models Categorical approach to many-sorted

More information

Syntactic Characterisations in Model Theory

Syntactic Characterisations in Model Theory Department of Mathematics Bachelor Thesis (7.5 ECTS) Syntactic Characterisations in Model Theory Author: Dionijs van Tuijl Supervisor: Dr. Jaap van Oosten June 15, 2016 Contents 1 Introduction 2 2 Preliminaries

More information

Syntax. Notation Throughout, and when not otherwise said, we assume a vocabulary V = C F P.

Syntax. Notation Throughout, and when not otherwise said, we assume a vocabulary V = C F P. First-Order Logic Syntax The alphabet of a first-order language is organised into the following categories. Logical connectives:,,,,, and. Auxiliary symbols:.,,, ( and ). Variables: we assume a countable

More information

What are Iteration Theories?

What are Iteration Theories? What are Iteration Theories? Jiří Adámek and Stefan Milius Institute of Theoretical Computer Science Technical University of Braunschweig Germany adamek,milius @iti.cs.tu-bs.de Jiří Velebil Department

More information

On the specification and verification of model transformations

On the specification and verification of model transformations On the specification and verification of model transformations Fernando Orejas 1 and Martin Wirsing 2 1 Universitat Politècnica de Catalunya, Barcelona (Spain), orejas@lsi.upc.edu 2 Ludwig-Maximilians

More information

The Essentially Equational Theory of Horn Classes

The Essentially Equational Theory of Horn Classes The Essentially Equational Theory of Horn Classes Hans E. Porst Dedicated to Professor Dr. Dieter Pumplün on the occasion of his retirement Abstract It is well known that the model categories of universal

More information

Universal Algebra for Logics

Universal Algebra for Logics Universal Algebra for Logics Joanna GRYGIEL University of Czestochowa Poland j.grygiel@ajd.czest.pl 2005 These notes form Lecture Notes of a short course which I will give at 1st School on Universal Logic

More information

Symbolic Graphs for Attributed Graph Constraints

Symbolic Graphs for Attributed Graph Constraints Symbolic Graphs for Attributed Graph Constraints Fernando Orejas Dpt. L.S.I., Universitat Politècnica de Catalunya, Campus Nord, Mòdul Omega, Jordi Girona 1-3, 08034 Barcelona, Spain. Abstract In this

More information

On the Complexity of the Reflected Logic of Proofs

On the Complexity of the Reflected Logic of Proofs On the Complexity of the Reflected Logic of Proofs Nikolai V. Krupski Department of Math. Logic and the Theory of Algorithms, Faculty of Mechanics and Mathematics, Moscow State University, Moscow 119899,

More information

A generalization of modal definability

A generalization of modal definability A generalization of modal definability Tin Perkov Polytechnic of Zagreb Abstract. Known results on global definability in basic modal logic are generalized in the following sense. A class of Kripke models

More information

Order Sorted Algebra. Japan Advanced Institute of Science and Technology. March 8, 2008

Order Sorted Algebra. Japan Advanced Institute of Science and Technology. March 8, 2008 Order Sorted Algebra Daniel Găină Japan Advanced Institute of Science and Technology March 8, 2008 Introduction There are many examples where all items of one sort are necessarily also items of some other

More information

1 First-order logic. 1 Syntax of first-order logic. 2 Semantics of first-order logic. 3 First-order logic queries. 2 First-order query evaluation

1 First-order logic. 1 Syntax of first-order logic. 2 Semantics of first-order logic. 3 First-order logic queries. 2 First-order query evaluation Knowledge Bases and Databases Part 1: First-Order Queries Diego Calvanese Faculty of Computer Science Master of Science in Computer Science A.Y. 2007/2008 Overview of Part 1: First-order queries 1 First-order

More information

Boolean Algebra and Propositional Logic

Boolean Algebra and Propositional Logic Boolean Algebra and Propositional Logic Takahiro Kato September 10, 2015 ABSTRACT. This article provides yet another characterization of Boolean algebras and, using this characterization, establishes a

More information

PART I. Abstract algebraic categories

PART I. Abstract algebraic categories PART I Abstract algebraic categories It should be observed first that the whole concept of category is essentially an auxiliary one; our basic concepts are those of a functor and a natural transformation.

More information

A MODEL-THEORETIC PROOF OF HILBERT S NULLSTELLENSATZ

A MODEL-THEORETIC PROOF OF HILBERT S NULLSTELLENSATZ A MODEL-THEORETIC PROOF OF HILBERT S NULLSTELLENSATZ NICOLAS FORD Abstract. The goal of this paper is to present a proof of the Nullstellensatz using tools from a branch of logic called model theory. In

More information

Lecture 2: Syntax. January 24, 2018

Lecture 2: Syntax. January 24, 2018 Lecture 2: Syntax January 24, 2018 We now review the basic definitions of first-order logic in more detail. Recall that a language consists of a collection of symbols {P i }, each of which has some specified

More information

Informatics and Mathematical Modelling, Techn. University of Denmark, Lyngby

Informatics and Mathematical Modelling, Techn. University of Denmark, Lyngby Computing and Informatics, Vol. 20, 2001,????, V 2003-Sep-4 Casl THE COMMON ALGEBRAIC SPECIFICATION LANGUAGE: SEMANTICS AND PROOF THEORY Till Mossakowski Department of Computer Science, University of Bremen,

More information

Introduction to Metalogic

Introduction to Metalogic Philosophy 135 Spring 2008 Tony Martin Introduction to Metalogic 1 The semantics of sentential logic. The language L of sentential logic. Symbols of L: Remarks: (i) sentence letters p 0, p 1, p 2,... (ii)

More information

Partially commutative linear logic: sequent calculus and phase semantics

Partially commutative linear logic: sequent calculus and phase semantics Partially commutative linear logic: sequent calculus and phase semantics Philippe de Groote Projet Calligramme INRIA-Lorraine & CRIN CNRS 615 rue du Jardin Botanique - B.P. 101 F 54602 Villers-lès-Nancy

More information

6 Coalgebraic modalities via predicate liftings

6 Coalgebraic modalities via predicate liftings 6 Coalgebraic modalities via predicate liftings In this chapter we take an approach to coalgebraic modal logic where the modalities are in 1-1 correspondence with so-called predicate liftings for the functor

More information

07 Equational Logic and Algebraic Reasoning

07 Equational Logic and Algebraic Reasoning CAS 701 Fall 2004 07 Equational Logic and Algebraic Reasoning Instructor: W. M. Farmer Revised: 17 November 2004 1 What is Equational Logic? Equational logic is first-order logic restricted to languages

More information

Automated Synthesis of Tableau Calculi

Automated Synthesis of Tableau Calculi Automated Synthesis of Tableau Calculi Renate A. Schmidt 1 and Dmitry Tishkovsky 1 School of Computer Science, The University of Manchester Abstract This paper presents a method for synthesising sound

More information

Boolean Algebra and Propositional Logic

Boolean Algebra and Propositional Logic Boolean Algebra and Propositional Logic Takahiro Kato June 23, 2015 This article provides yet another characterization of Boolean algebras and, using this characterization, establishes a more direct connection

More information

Topos Theory. Lectures 17-20: The interpretation of logic in categories. Olivia Caramello. Topos Theory. Olivia Caramello.

Topos Theory. Lectures 17-20: The interpretation of logic in categories. Olivia Caramello. Topos Theory. Olivia Caramello. logic s Lectures 17-20: logic in 2 / 40 logic s Interpreting first-order logic in In Logic, first-order s are a wide class of formal s used for talking about structures of any kind (where the restriction

More information

A Tableau Calculus for Minimal Modal Model Generation

A Tableau Calculus for Minimal Modal Model Generation M4M 2011 A Tableau Calculus for Minimal Modal Model Generation Fabio Papacchini 1 and Renate A. Schmidt 2 School of Computer Science, University of Manchester Abstract Model generation and minimal model

More information

Bisimulation for Neighbourhood Structures

Bisimulation for Neighbourhood Structures Bisimulation for Neighbourhood Structures Helle Hvid Hansen 1,2 Clemens Kupke 2 Eric Pacuit 3 1 Vrije Universiteit Amsterdam (VUA) 2 Centrum voor Wiskunde en Informatica (CWI) 3 Universiteit van Amsterdam

More information

A note on coinduction and weak bisimilarity for while programs

A note on coinduction and weak bisimilarity for while programs Centrum voor Wiskunde en Informatica A note on coinduction and weak bisimilarity for while programs J.J.M.M. Rutten Software Engineering (SEN) SEN-R9826 October 31, 1998 Report SEN-R9826 ISSN 1386-369X

More information

Part II. Logic and Set Theory. Year

Part II. Logic and Set Theory. Year Part II Year 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2018 60 Paper 4, Section II 16G State and prove the ǫ-recursion Theorem. [You may assume the Principle of ǫ- Induction.]

More information

A STUDY OF TRUTH PREDICATES IN MATRIX SEMANTICS

A STUDY OF TRUTH PREDICATES IN MATRIX SEMANTICS A STUDY OF TRUTH PREDICATES IN MATRIX SEMANTICS TOMMASO MORASCHINI Abstract. Abstract algebraic logic is a theory that provides general tools for the algebraic study of arbitrary propositional logics.

More information

Behavioral Reasoning for Conditional Equations

Behavioral Reasoning for Conditional Equations Behavioral Reasoning for Conditional Equations M A N U E L A N T Ó N I O M A R T I N S1 and D O N P I G O Z Z I 2 1 Departamento de Matemática Universidade de Aveiro 3810-193 Aveiro Portugal martins@mat.ua.pt

More information

Streams and Coalgebra Lecture 2

Streams and Coalgebra Lecture 2 Streams and Coalgebra Lecture 2 Helle Hvid Hansen and Jan Rutten Radboud University Nijmegen & CWI Amsterdam Representing Streams II, Lorentz Center, Leiden, January 2014 Tutorial Overview Lecture 1 (Hansen):

More information

Foundations of Mathematics

Foundations of Mathematics Foundations of Mathematics Andrew Monnot 1 Construction of the Language Loop We must yield to a cyclic approach in the foundations of mathematics. In this respect we begin with some assumptions of language

More information

Topos Theory. Lectures 21 and 22: Classifying toposes. Olivia Caramello. Topos Theory. Olivia Caramello. The notion of classifying topos

Topos Theory. Lectures 21 and 22: Classifying toposes. Olivia Caramello. Topos Theory. Olivia Caramello. The notion of classifying topos Lectures 21 and 22: toposes of 2 / 30 Toposes as mathematical universes of Recall that every Grothendieck topos E is an elementary topos. Thus, given the fact that arbitrary colimits exist in E, we can

More information

via Topos Theory Olivia Caramello University of Cambridge The unification of Mathematics via Topos Theory Olivia Caramello

via Topos Theory Olivia Caramello University of Cambridge The unification of Mathematics via Topos Theory Olivia Caramello in University of Cambridge 2 / 23 in in In this lecture, whenever I use the word topos, I really mean Grothendieck topos. Recall that a Grothendieck topos can be seen as: a generalized space a mathematical

More information

EQUATIONAL REASONING WITH SUBTYPES

EQUATIONAL REASONING WITH SUBTYPES EQUATIONAL REASONING WITH SUBTYPES GARY T. LEAVENS AND DON PIGOZZI Abstract. Using equational logic as a specification language, we investigate the proof theory of behavioral subtyping for object-oriented

More information

Introduction to Logic in Computer Science: Autumn 2006

Introduction to Logic in Computer Science: Autumn 2006 Introduction to Logic in Computer Science: Autumn 2006 Ulle Endriss Institute for Logic, Language and Computation University of Amsterdam Ulle Endriss 1 Plan for Today Today s class will be an introduction

More information

Step-indexed models of call-by-name: a tutorial example

Step-indexed models of call-by-name: a tutorial example Step-indexed models of call-by-name: a tutorial example Aleš Bizjak 1 and Lars Birkedal 1 1 Aarhus University {abizjak,birkedal}@cs.au.dk June 19, 2014 Abstract In this tutorial paper we show how to construct

More information

hal , version 1-21 Oct 2009

hal , version 1-21 Oct 2009 ON SKOLEMISING ZERMELO S SET THEORY ALEXANDRE MIQUEL Abstract. We give a Skolemised presentation of Zermelo s set theory (with notations for comprehension, powerset, etc.) and show that this presentation

More information

Propositional and Predicate Logic - V

Propositional and Predicate Logic - V Propositional and Predicate Logic - V Petr Gregor KTIML MFF UK WS 2016/2017 Petr Gregor (KTIML MFF UK) Propositional and Predicate Logic - V WS 2016/2017 1 / 21 Formal proof systems Hilbert s calculus

More information

Reconsidering MacLane. Peter M. Hines

Reconsidering MacLane. Peter M. Hines Reconsidering MacLane Coherence for associativity in infinitary and untyped settings Peter M. Hines Oxford March 2013 Topic of the talk: Pure category theory... for its own sake. This talk is about the

More information

1 Introduction. 2 Categories. Mitchell Faulk June 22, 2014 Equivalence of Categories for Affine Varieties

1 Introduction. 2 Categories. Mitchell Faulk June 22, 2014 Equivalence of Categories for Affine Varieties Mitchell Faulk June 22, 2014 Equivalence of Categories for Affine Varieties 1 Introduction Recall from last time that every affine algebraic variety V A n determines a unique finitely generated, reduced

More information

Categorical logics for contravariant simulations, partial bisimulations, modal refinements and mixed transition systems

Categorical logics for contravariant simulations, partial bisimulations, modal refinements and mixed transition systems Categorical logics for contravariant simulations, partial bisimulations, modal refinements and mixed transition systems Ignacio Fábregas, Miguel Palomino, and David de Frutos-Escrig Departamento de Sistemas

More information

Adjunctions! Everywhere!

Adjunctions! Everywhere! Adjunctions! Everywhere! Carnegie Mellon University Thursday 19 th September 2013 Clive Newstead Abstract What do free groups, existential quantifiers and Stone-Čech compactifications all have in common?

More information

Closure operators on sets and algebraic lattices

Closure operators on sets and algebraic lattices Closure operators on sets and algebraic lattices Sergiu Rudeanu University of Bucharest Romania Closure operators are abundant in mathematics; here are a few examples. Given an algebraic structure, such

More information

Modular Construction of Modal Logics

Modular Construction of Modal Logics Modular Construction of Modal Logics Corina Cîrstea 1 and Dirk Pattinson 2 1 School of Electronics and Computer Science, University of Southampton, UK cc2@ecs.soton.ac.uk 2 Institut für Informatik, LMU

More information

Universität Augsburg

Universität Augsburg Universität Augsburg Properties of Overwriting for Updates in Typed Kleene Algebras Thorsten Ehm Report 2000-7 Dezember 2000 Institut für Informatik D-86135 Augsburg Copyright c Thorsten Ehm Institut für

More information

A Note on Graded Modal Logic

A Note on Graded Modal Logic A Note on Graded Modal Logic Maarten de Rijke Studia Logica, vol. 64 (2000), pp. 271 283 Abstract We introduce a notion of bisimulation for graded modal logic. Using these bisimulations the model theory

More information

The non-logical symbols determine a specific F OL language and consists of the following sets. Σ = {Σ n } n<ω

The non-logical symbols determine a specific F OL language and consists of the following sets. Σ = {Σ n } n<ω 1 Preliminaries In this chapter we first give a summary of the basic notations, terminology and results which will be used in this thesis. The treatment here is reduced to a list of definitions. For the

More information

Another algorithm for nonnegative matrices

Another algorithm for nonnegative matrices Linear Algebra and its Applications 365 (2003) 3 12 www.elsevier.com/locate/laa Another algorithm for nonnegative matrices Manfred J. Bauch University of Bayreuth, Institute of Mathematics, D-95440 Bayreuth,

More information

TABLEAU SYSTEM FOR LOGIC OF CATEGORIAL PROPOSITIONS AND DECIDABILITY

TABLEAU SYSTEM FOR LOGIC OF CATEGORIAL PROPOSITIONS AND DECIDABILITY Bulletin of the Section of Logic Volume 37:3/4 (2008), pp. 223 231 Tomasz Jarmużek TABLEAU SYSTEM FOR LOGIC OF CATEGORIAL PROPOSITIONS AND DECIDABILITY Abstract In the article we present an application

More information

Bringing class diagrams to life

Bringing class diagrams to life Bringing class diagrams to life Luis S. Barbosa & Sun Meng DI-CCTC, Minho University, Braga & CWI, Amsterdam UML & FM Workshop 2009 Rio de Janeiro 8 December, 2009 Formal Methods proofs problems structures

More information

On Modal µ-calculus And Non-Well-Founded Set Theory

On Modal µ-calculus And Non-Well-Founded Set Theory On Modal µ-calculus And Non-Well-Founded Set Theory Luca Alberucci (albe@iam.unibe.ch) and Vincenzo Salipante (salipant@iam.unibe.ch) Institut für Informatik und angewandte Mathematik, Universität Bern,

More information

185.A09 Advanced Mathematical Logic

185.A09 Advanced Mathematical Logic 185.A09 Advanced Mathematical Logic www.volny.cz/behounek/logic/teaching/mathlog13 Libor Běhounek, behounek@cs.cas.cz Lecture #1, October 15, 2013 Organizational matters Study materials will be posted

More information

Bisimulation for Neighbourhood Structures

Bisimulation for Neighbourhood Structures Bisimulation for Neighbourhood Structures Helle Hvid Hansen 1,2 Clemens Kupke 2 Eric Pacuit 3 1 Vrije Universiteit Amsterdam (VUA) 2 Centrum voor Wiskunde en Informatica (CWI) 3 Universiteit van Amsterdam

More information

Algebraizing Hybrid Logic. Evangelos Tzanis University of Amsterdam Institute of Logic, Language and Computation

Algebraizing Hybrid Logic. Evangelos Tzanis University of Amsterdam Institute of Logic, Language and Computation Algebraizing Hybrid Logic Evangelos Tzanis University of Amsterdam Institute of Logic, Language and Computation etzanis@science.uva.nl May 1, 2005 2 Contents 1 Introduction 5 1.1 A guide to this thesis..........................

More information

Partial model checking via abstract interpretation

Partial model checking via abstract interpretation Partial model checking via abstract interpretation N. De Francesco, G. Lettieri, L. Martini, G. Vaglini Università di Pisa, Dipartimento di Ingegneria dell Informazione, sez. Informatica, Via Diotisalvi

More information

FROM COHERENT TO FINITENESS SPACES

FROM COHERENT TO FINITENESS SPACES FROM COHERENT TO FINITENESS SPACES PIERRE HYVERNAT Laboratoire de Mathématiques, Université de Savoie, 73376 Le Bourget-du-Lac Cedex, France. e-mail address: Pierre.Hyvernat@univ-savoie.fr Abstract. This

More information

An Introduction to Modal Logic III

An Introduction to Modal Logic III An Introduction to Modal Logic III Soundness of Normal Modal Logics Marco Cerami Palacký University in Olomouc Department of Computer Science Olomouc, Czech Republic Olomouc, October 24 th 2013 Marco Cerami

More information

Higher Order Containers

Higher Order Containers Higher Order Containers Thorsten Altenkirch 1, Paul Levy 2, and Sam Staton 3 1 University of Nottingham 2 University of Birmingham 3 University of Cambridge Abstract. Containers are a semantic way to talk

More information

Logics above S4 and the Lebesgue measure algebra

Logics above S4 and the Lebesgue measure algebra Logics above S4 and the Lebesgue measure algebra Tamar Lando Abstract We study the measure semantics for propositional modal logics, in which formulas are interpreted in the Lebesgue measure algebra M,

More information

Fibres. Temesghen Kahsai. Fibres in Concrete category. Generalized Fibres. Fibres. Temesghen Kahsai 14/02/ 2007

Fibres. Temesghen Kahsai. Fibres in Concrete category. Generalized Fibres. Fibres. Temesghen Kahsai 14/02/ 2007 14/02/ 2007 Table of Contents ... and back to theory Example Let Σ = (S, TF) be a signature and Φ be a set of FOL formulae: 1. SPres is the of strict presentation with: objects: < Σ, Φ >, morphism σ :

More information

EXAMPLES AND EXERCISES IN BASIC CATEGORY THEORY

EXAMPLES AND EXERCISES IN BASIC CATEGORY THEORY EXAMPLES AND EXERCISES IN BASIC CATEGORY THEORY 1. Categories 1.1. Generalities. I ve tried to be as consistent as possible. In particular, throughout the text below, categories will be denoted by capital

More information

arxiv: v1 [cs.pl] 19 May 2016

arxiv: v1 [cs.pl] 19 May 2016 arxiv:1605.05858v1 [cs.pl] 19 May 2016 Domain Theory: An Introduction Robert Cartwright Rice University Rebecca Parsons ThoughtWorks, Inc. Moez AbdelGawad SRTA-City Hunan University This monograph is an

More information

AN EXTENSION OF THE PROBABILITY LOGIC LP P 2. Tatjana Stojanović 1, Ana Kaplarević-Mališić 1 and Zoran Ognjanović 2

AN EXTENSION OF THE PROBABILITY LOGIC LP P 2. Tatjana Stojanović 1, Ana Kaplarević-Mališić 1 and Zoran Ognjanović 2 45 Kragujevac J. Math. 33 (2010) 45 62. AN EXTENSION OF THE PROBABILITY LOGIC LP P 2 Tatjana Stojanović 1, Ana Kaplarević-Mališić 1 and Zoran Ognjanović 2 1 University of Kragujevac, Faculty of Science,

More information

Basics of Relation Algebra. Jules Desharnais Département d informatique et de génie logiciel Université Laval, Québec, Canada

Basics of Relation Algebra. Jules Desharnais Département d informatique et de génie logiciel Université Laval, Québec, Canada Basics of Relation Algebra Jules Desharnais Département d informatique et de génie logiciel Université Laval, Québec, Canada Plan 1. Relations on a set 2. Relation algebra 3. Heterogeneous relation algebra

More information

TR : Binding Modalities

TR : Binding Modalities City University of New York (CUNY) CUNY Academic Works Computer Science Technical Reports Graduate Center 2012 TR-2012011: Binding Modalities Sergei N. Artemov Tatiana Yavorskaya (Sidon) Follow this and

More information

From Algebras and Coalgebras to Dialgebras

From Algebras and Coalgebras to Dialgebras From Algebras and Coalgebras to Dialgebras Erik Poll and Jan Zwanenburg University of Nijmegen Toernooiveld 1, 6525 ED Nijmegen, The Netherlands http://www.cs.kun.nl/~{erikpoll,janz} {erikpoll,janz}@cs.kun.nl

More information

Category theory for computer science. Overall idea

Category theory for computer science. Overall idea Category theory for computer science generality abstraction convenience constructiveness Overall idea look at all objects exclusively through relationships between them capture relationships between objects

More information

Design of Distributed Systems Melinda Tóth, Zoltán Horváth

Design of Distributed Systems Melinda Tóth, Zoltán Horváth Design of Distributed Systems Melinda Tóth, Zoltán Horváth Design of Distributed Systems Melinda Tóth, Zoltán Horváth Publication date 2014 Copyright 2014 Melinda Tóth, Zoltán Horváth Supported by TÁMOP-412A/1-11/1-2011-0052

More information

3. Categories and Functors We recall the definition of a category: Definition 3.1. A category C is the data of two collections. The first collection

3. Categories and Functors We recall the definition of a category: Definition 3.1. A category C is the data of two collections. The first collection 3. Categories and Functors We recall the definition of a category: Definition 3.1. A category C is the data of two collections. The first collection is called the objects of C and is denoted Obj(C). Given

More information

Lax Extensions of Coalgebra Functors and Their Logic

Lax Extensions of Coalgebra Functors and Their Logic Lax Extensions of Coalgebra Functors and Their Logic Johannes Marti, Yde Venema ILLC, University of Amsterdam Abstract We discuss the use of relation lifting in the theory of set-based coalgebra and coalgebraic

More information

A Discrete Duality Between Nonmonotonic Consequence Relations and Convex Geometries

A Discrete Duality Between Nonmonotonic Consequence Relations and Convex Geometries A Discrete Duality Between Nonmonotonic Consequence Relations and Convex Geometries Johannes Marti and Riccardo Pinosio Draft from April 5, 2018 Abstract In this paper we present a duality between nonmonotonic

More information

Dual Adjunctions Between Algebras and Coalgebras

Dual Adjunctions Between Algebras and Coalgebras Dual Adjunctions Between Algebras and Coalgebras Hans E. Porst Department of Mathematics University of Bremen, 28359 Bremen, Germany porst@math.uni-bremen.de Abstract It is shown that the dual algebra

More information

Solving equations over small unary algebras

Solving equations over small unary algebras Computational Logic and Applications, CLA 05 DMTCS proc AF, 2006, 49 60 Solving equations over small unary algebras Przemyslaw Broniek broniek@iiujedupl, Algorithmics Research Group, Jagiellonian University,

More information

Symbol Index Group GermAnal Ring AbMonoid

Symbol Index Group GermAnal Ring AbMonoid Symbol Index 409 Symbol Index Symbols of standard and uncontroversial usage are generally not included here. As in the word index, boldface page-numbers indicate pages where definitions are given. If a

More information

2. Prime and Maximal Ideals

2. Prime and Maximal Ideals 18 Andreas Gathmann 2. Prime and Maximal Ideals There are two special kinds of ideals that are of particular importance, both algebraically and geometrically: the so-called prime and maximal ideals. Let

More information

Fundamentals of Software Engineering

Fundamentals of Software Engineering Fundamentals of Software Engineering First-Order Logic Ina Schaefer Institute for Software Systems Engineering TU Braunschweig, Germany Slides by Wolfgang Ahrendt, Richard Bubel, Reiner Hähnle (Chalmers

More information

Krivine s Intuitionistic Proof of Classical Completeness (for countable languages)

Krivine s Intuitionistic Proof of Classical Completeness (for countable languages) Krivine s Intuitionistic Proof of Classical Completeness (for countable languages) Berardi Stefano Valentini Silvio Dip. Informatica Dip. Mat. Pura ed Applicata Univ. Torino Univ. Padova c.so Svizzera

More information

Fundamentals of Software Engineering

Fundamentals of Software Engineering Fundamentals of Software Engineering First-Order Logic Ina Schaefer Institute for Software Systems Engineering TU Braunschweig, Germany Slides by Wolfgang Ahrendt, Richard Bubel, Reiner Hähnle (Chalmers

More information

Vietoris bisimulations

Vietoris bisimulations Vietoris bisimulations N. Bezhanishvili, G. Fontaine and Y. Venema July 17, 2008 Abstract Building on the fact that descriptive frames are coalgebras for the Vietoris functor on the category of Stone spaces,

More information

Abstract model theory for extensions of modal logic

Abstract model theory for extensions of modal logic Abstract model theory for extensions of modal logic Balder ten Cate Stanford, May 13, 2008 Largely based on joint work with Johan van Benthem and Jouko Väänänen Balder ten Cate Abstract model theory for

More information

Derived Algebraic Geometry IX: Closed Immersions

Derived Algebraic Geometry IX: Closed Immersions Derived Algebraic Geometry I: Closed Immersions November 5, 2011 Contents 1 Unramified Pregeometries and Closed Immersions 4 2 Resolutions of T-Structures 7 3 The Proof of Proposition 1.0.10 14 4 Closed

More information

Morita-equivalences for MV-algebras

Morita-equivalences for MV-algebras Morita-equivalences for MV-algebras Olivia Caramello* University of Insubria Geometry and non-classical logics 5-8 September 2017 *Joint work with Anna Carla Russo O. Caramello Morita-equivalences for

More information

Jónsson posets and unary Jónsson algebras

Jónsson posets and unary Jónsson algebras Jónsson posets and unary Jónsson algebras Keith A. Kearnes and Greg Oman Abstract. We show that if P is an infinite poset whose proper order ideals have cardinality strictly less than P, and κ is a cardinal

More information

A Semantic Approach to Interpolation

A Semantic Approach to Interpolation A Semantic Approach to Interpolation Andrei Popescu and Traian Florin Şerbănuţă and Grigore Roşu Department of Computer Science, University of Illinois at Urbana-Champaign. {popescu2,tserban2,grosu}@cs.uiuc.edu

More information

On the Boolean Algebra of Shape Analysis Constraints

On the Boolean Algebra of Shape Analysis Constraints On the Boolean Algebra of Shape Analysis Constraints Viktor Kuncak and Martin Rinard Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology Cambridge, MA 02139, USA

More information

Lectures - XXIII and XXIV Coproducts and Pushouts

Lectures - XXIII and XXIV Coproducts and Pushouts Lectures - XXIII and XXIV Coproducts and Pushouts We now discuss further categorical constructions that are essential for the formulation of the Seifert Van Kampen theorem. We first discuss the notion

More information

The equivalence axiom and univalent models of type theory.

The equivalence axiom and univalent models of type theory. The equivalence axiom and univalent models of type theory. (Talk at CMU on February 4, 2010) By Vladimir Voevodsky Abstract I will show how to define, in any type system with dependent sums, products and

More information

The Vaught Conjecture Do uncountable models count?

The Vaught Conjecture Do uncountable models count? The Vaught Conjecture Do uncountable models count? John T. Baldwin Department of Mathematics, Statistics and Computer Science University of Illinois at Chicago May 22, 2005 1 Is the Vaught Conjecture model

More information

Matching Logic: Syntax and Semantics

Matching Logic: Syntax and Semantics Matching Logic: Syntax and Semantics Grigore Roșu 1 and Traian Florin Șerbănuță 2 1 University of Illinois at Urbana-Champaign, USA grosu@illinois.edu 2 University of Bucharest, Romania traian.serbanuta@unibuc.ro

More information

1. Propositional Calculus

1. Propositional Calculus 1. Propositional Calculus Some notes for Math 601, Fall 2010 based on Elliott Mendelson, Introduction to Mathematical Logic, Fifth edition, 2010, Chapman & Hall. 2. Syntax ( grammar ). 1.1, p. 1. Given:

More information