Deprecated: The each() function is deprecated. This message will be suppressed on further calls in /home/zhenxiangba/zhenxiangba.com/public_html/phproxy-improved-master/index.php on line 456
Steve Vickers's WWW Home Page
[go: Go Back, main page]

 
 

Dr Steve Vickers

School of Computer ScienceThe University of Birmingham,
Birmingham, B15 2TT, UK.

email address: s.j.vickers
post office: cs.bham.ac.uk

Use both, with "@" in between in the usual way.

Tel: 0121-414 3743

Teaching

I am Research Students Tutor. (Here is a link to information for current research students.)
I do not deal with admissions to research degree programmes (PhD or MPhil in Computer Science or Cognitive Science) - our admissions page is here.

I teach -

I also have lecture notes for research seminars I have given (primarily for PhD students) on Category Theory and "Topology via Logic".

Check my personal timetable if you want to check when I am likely to be free.

Research

I am a member of the Mathematical Foundations of Computer Science research group.
My particular research interests are in geometric logic, topology and topos theory.

Before I joined Birmingham I was at the Department of Pure Maths at the Open University, and before that I was at the Department of Computing at Imperial College, where I worked in the Theory and Formal Methods group.

You can download many of my papers and talks, and also theses of some of my former PhD students.

Geometric Logic

Most of my work is in investigating "geometric logic" (so-called from its origins in algebraic geometry). This has some unusual properties. First, it makes a hard distinction between formulae and axioms: a logical formula is restricted in the connectives it can use, to conjunction, disjunction, equality and existential quantification, and an axiom (for a geometric theory) expresses relationships between formulae in the form "for all x, y, z, ... (formula1 - formula2)". Hence the missing connectives can be introduced in just one layer - no nesting - in axioms. Second, it allows infinite disjunctions in formulae.

Though these features may look weird, they have some conceptual justification in the idea of observation: formulae correspond to observable facts (and infinite disjunction is not a problem observationally) while the axioms correspond to background assumptions or scientific hypotheses.

A particularly pleasing property of the logic is its close relationship with topology: the class of models for a propositional geometric theory is automatically a topological space, and by working within a "geometrically" constrained constructivist mathematics, one finds that constructions of functions from one such space to another are automatically continuous - indeed, with a bit of oversimplification, I have argued ("Toposes pour les vraiment nuls") that since "geometric mathematics" is intrinsically continuous, it is only to compensate for the over-optimism of classical reasoning that one needs explicit topologies and continuity proofs. This "topology-free space" approach leads naturally, by generalizing to the predicate logic, to Grothendieck's toposes as generalized topological spaces.

To work with "topology-free spaces" is to accept certain constructivist constraints on one's mathematics, and a number of my more theoretical papers (in particular "Topical Categories of Domains" and "Localic Completion of Quasimetric Spaces") serve as case studies to investigate whether the constraints still permit sensible mathematics.

A number of features suggest that the ideas may be applicable to specification of computer systems:

and some preliminary work has been done on how geometric logic might be used as a specification language.

A key part of this application would be to remove the infinities from the logic, so that support tools can be based on a finitary formalism. To some extent this can be done by introducing inductive sort constructors in the many-sorted predicate logic: then propositional theories with infinite disjunctions can often be replaced by equivalent finite predicate theories in which the sorts are specified as being constructed in a given way. A promising mathematical approach to this is by the category of Joyal's "Arithmetic Universes".

Related to this field is my lecture Non-classical Reasoning, which was given as a first year undergraduate "Topics" lecture in the Department of Computing at Imperial College.

Lectures on Category Theory and Topology via Logic

In 2003-4 I gave two series of talks aimed primarily at research students. Copies of my slides are available from me.

The first series was on Category Theory, and comprised lectures I gave at Imperial introducing categories as part of an MSc module on Mathematical Structures for Computer Science.

The second series developed the ideas of my book Topology via Logic (Cambridge University Press 1989, ISBN 0 521 36062 5 and 0 521 57651 2), which was an introduction to topology and its use in computer science.

In many ways the lectures followed the book, but they went beyond it in bringing out the deep relationship with constructive logic. Part of their aim was to put together the shape of a constructive 2nd edition of Topology via Logic.

Outline

"via Logic"
In my book, "via Logic" referred to the way it presents topologies as Lindenbaum algebras for theories in a particular non-classical logic, the so called geometric logic. This lies behind an alternative approach to topology known as locale theory, which has been used in the semantics of programming languages - notably in Abramsky's thesis.
Observability
In my book I motivate the use of geometric logic by an informal discussion of how certain logical connective have observational content (this is also present in the work of Smyth and Abramsky). Then the basic deal in semantics is that a domain, with its Scott topology, embodies a logical theory of how one can observe some type of program by watching it run (without access to the source code).
Going constructive
The book is entirely classical in its mathematics. However, since it was written I have come to understand much better the key role of constructive mathematics in topology and have exploited it in my papers such as "Topical categories of domains". This is what I want to explain in the lectures.
Better constructive topology
One aspect of this is that if you are already committed to constructivism (maybe in order to take better account of computational issues), then the localic approach is known to give better results in constructive topology than topology itself.
Continuity for free
But there are deeper reasons topological reasons for being constructive. Associated with the geometric logic, there is a "geometric mathematics". It is constructive, and also has an intrinsic continuity. If you adhere to the constructive constraints of geometric reasoning then you do not have to give explicit continuity proofs - continuity is automatic. In effect, the usual continuity proofs in classical reasoning are bureaucratic certifications that you haven't transgressed the geometric constraints. So there are quite practical reasons for accepting those constraints.
Toposes
A major spinoff is that the same approach adapts quite naturally to a generalization of topological space, namely toposes. This aspect of toposes is normally concealed, but it is fundamental to their importance. Roughly speaking, from this point of view a (Grothendieck) topos is a topological space for which there are "not enough opens" and instead the topological structure has to be defined through what are known as its sheaves.

Downloadable papers

Below are preprint versions of some of my papers, the most recent first. A more complete list (with abstracts and citation details) is also available.

(with Achim Jung and Drew Moshier) "Presenting dcpos and dcpo algebras" describes the presentations of dcpo algebras by generators and relations.

"Issues of logic, algebra and topology in ontology" uses geometric logic as a case study for how a logic might be affected by the ontological commitment one is prepared to make.

"Fuzzy sets and geometric logic" describes how geometric logic can provide a conceptual simplification of Hoehle's sheaf approach to fuzzy sets. You can also download slides of my Linz talk on this topic.

"Cosheaves and connectedness in formal topology" gives a predicative introduction to some topos ideas in Bunge and Funk's book "Singular Coverings of Toposes".

"A localic theory of lower and upper integrals" is just what it says.

"The connected Vietoris powerlocale" describes a powerlocale construction whose points are connected, and uses it to give a choice-free constructive account of the Intermediate Value Theorem and Rolle's Theorem.

"Locales and toposes as spaces" is a first step towards a constructive version of "Topology via Logic". It explains how theories in geometric logic can be viewed as topological spaces (for the propositional fragment) or generalized topological spaces (toposes) for the full predicate logic.

(with Erik Palmgren) "Partial Horn logic and cartesian categories" gives a simple presentation of cartesian theories using a logic in which function symbols may be partial. It leads to an easy proof of the free model theorem, and a simple construction of classifying categories.

"Sublocales in formal topology" shows how various results about sublocales in topos-valid locale theory transfer to predicative formal topology.

"Some constructive roads to Tychonoff" uses Tychonoff's theorem as a case study for approaches to topology in constructive mathematics. The final version is published in this book.

(with Gillian Hill) "A language for configuring multi-level specifications" develops "Presheaves as configured specifications" to cope with specifications in which components of systems may themselves be configured systems.

"Localic completion of generalized metric spaces I" shows how generalized metric spaces may be completed to give locales. "Generalized" means esentially that only the zero self-distance and the triangle inequality are expected. "Localic completion of generalized metric spaces II: Powerlocales" investigates their powerlocales, gives a simple description of them as "generalized metric powerspaces", and describes some application.

"Entailment systems for stably locally compact locales" describes a way of presenting stably locally compact locales using a kind of non-reflexive sequent calculus. It is similar to the Multilingual Sequent Calculus of Jung, Kegelmann and Moshier, but without the explicit presence of  conjunction and disjunction in the sequents.

(with Christopher Townsend) "A universal characterization of the double powerlocale" shows that the double powerlocale PP(X) can be considered a double exponential $^($^X) even when X is not locally compact (so $X doesn't exist). It does this using the Yoneda embedding of the category of locales.

"Compactness in locales and in formal topology" describes a general criterion for compactness of a formal topology (with preordered base) and proves it in both topos-valid mathematics and predicative mathematics.

"The double powerlocale and exponentiation: a case study in geometric logic" examines the "double powerlocale", the composite of the upper and lower powerlocales, and relates it to exponentiation. It uses geometric reasoning (stable under pullback along geometric morphisms) to make locales appear like spaces.

(with Pedro Resende) "Localic sup-lattices and tropological systems" resets the quantale methods in a constructive localic framework.

(with Gillian Hill) "Presheaves as Configured Specifications" outlines a specification language for describing how components are put together with sharing. It uses presheaves to give a mathematical treatment that is both simple and specificationally natural.

"Strongly Algebraic = SFP (Topically)" develops the work of "Topical Categories of Domains" to cover the equivalence between two of Plotkin's characterizations of SFP domains. It also serves as an introduction to how presheaf toposes are interpreted as generalized spaces.

"Localic Completion of Quasimetric Spaces" presents a localic account of completion of quasimetric spaces, using Lawvere's approach by enriched categories.

"Topical Categories of Domains" sets domain theory, and in particular Abramsky's account in "Domain Theory in Logical Form", in a framework of geometric logic. Ordinary categories of domains are replaced by "topical" categories, internal in the category of toposes and geometric morphisms: the classes of objects and morphisms are replaced by toposes classifying them.

"Topology via Constructive Logic" presents the idea that constructive mathematics has practical usefulness in studying topology.

"Constructive points of powerlocales" gives a constructively sound account of the points of the lower, upper and Vietoris powerlocales, and shows how in certain cases a powerlocale of D is homeomorphic to $^D.

"Toposes pour les vraiment nuls" is a further introduction to Grothendieck's idea of toposes as generalized topological spaces, using geometric logic to hide topology.

"Toposes pour les nuls" is a brief introduction to Grothendieck's idea of toposes as generalized topological spaces, written from a topological viewpoint.

"Locales are not pointless" presents some preliminary results for reasoning about arbitrary locales through their points, with particular applications to powerlocales.

(With Mark Dawson) "Towards a GeoZ toolkit" discusses how a mathematical toolkit for GeoZ must differ from that for Z, and illustrates the use of schema entailment in presenting it.

"Geometric logic as a specification language" is a short introduction to GeoZ, a Z-like specification language based on geometric logic. Includes a description of "schema entailment" and its use as a definitional mechanism.

"Geometric Logic in Computer Science" is an introduction to geometric logic, paying particular attention to the deep mathematical properties that make it differ from other logics.

(With Samson Abramsky) "Quantales, Observational Logic and Process Semantics" studies quantales (complete lattices equipped with monoid structure, the binary operation distributing over all joins) with applications to process semantics.

"Information Systems for Continuous Posets" defines an "infosys" as an idempotent relation, and shows how they may be used to present continuous dcpos. Includes a proof that the Vietoris powerlocale of a continuous dcpo is again a continuous dcpo.

"Geometric Theories and Databases" describes a topos-theoretic construction of categorical powerdomains (bagtoposes) with tentative application to database theory.

(With Peter Johnstone) "Preframe Presentations Present" investigates preframes (directed complete posets with finite meets that distribute over the directed joins) with applications to locale theory, including an easy proof of the localic Tychonoff theorem.

Talks

"Aspects of Topology" (Departmental Seminar, School of Computer Science, University of Brimingham) - Introduces topology and some of its application in theoretical computer science in the "point-free" form, and then leads on to toposes and a novel use (Imperial, Nijmegen) in the formulation of quantum mechanics.

"On the Trail of the Formal Topos" - Advances in Constructive Topology and Logical Foundations, Padua 8-11 October 2008. Discusses routes for extracting predicative content from topos theory.

"Locales via Bundles" - Sheaves in Geometry and Quantum Theory, Nijmegen 3-5 Spetember 2008; slightly briefer version given at Categories, Logic and Physics, Oxford, 23-24 August 2008. Introduces locales through the idea of bundles, with reference to the topos approaches to quantum theory.

"Fuzzy Sets and Geometric Logic" - 29th Linz Seminar on Fuzzy Set Theory: Foundations of Lattice-Valued Mathematics with Applications to Algebra and Topology, Linz 12-16 February 2008. Explains sheaf accounts of fuzzy sets in terms of geometric logic.

"Schemas as Toposes" - Open University on 2nd July 2002.

Past PhD students

(New PhD students welcome!)