Computer science, maths, and occasionally something completely different

Possession of Knowledge Mandates an Ethical Responsibility

 
Rod Burstall, Professor Emeritus, School of Informatics, the University of Edinburgh Rod Burstall is one of the founding members of Laboratory for Foundations of Computer Science. The photograph is from Rodney Topor's flickr.

Rod Burstall dropped by the LFCS lab lunch yesterday. Lab lunch is an informal gathering of LFCS members which takes place every Tuesday. On rota a LFCS member gives a (supposedly lightweight) talk which is usually followed by a round of questions. Yesterday, after the questions, Rod stood up to remind us that we, as researchers, carry an ethical responsibility for how our research impact the surrounding society. Read more.

GDP70

 
Leslie Lamport congratulates Gordon Plotkin on his 0th birthday. By the end of his talk, Leslie Lamport congratulated Gordon Plotkin.

Today is Gordon Plotkin's 70th birthday! The Laboratory for Foundations of Computer Science at the University of Edinburgh held an event in his honour today to celebrate his birthday. It was a full day of talks by friends, colleagues, and former students of Gordon's.

There is absolutely no doubt that Gordon is my favourite computer scientist. Gordon has made numerous contributions to computer science, and I have been exposed to many of them throughout my undergraduate and postgraduate studies. In fact, my current research topic comes from his line of work. Therefore I am also deeply grateful to be able to take part in the celebration of him on his 70th birthday. Read more.

A Compiler for Multi-shot Effect Handlers

 
Computer Laboratory entrance, Cambridge The pond at Pembroke College, Cambridge. I have seen deeper rain puddles in Edinburgh.

I recently spent two weeks at OCaml Labs, the University of Cambridge implementing a compiler for the experimental, research language Links with effect handlers. During those two weeks I managed to compile a substantial subset of the Links language to native code, in particular, I managed to compile some programs that use handlers. My Links compiler reuses most of the infrastructure from the OCaml effects/multicore compiler.

At OCaml Labs I was working with KC Sivaramakrishnan, whom is working on the Multicore OCaml project.

In this post I shall discuss numerous three things: I will briefly describe the compiler infrastructure, discuss an example program that compiles, and how I encode multi-shot handlers. Read more.

Visiting OCaml Labs -- Day #1

 
Computer Laboratory entrance, Cambridge Entrance to the Computer Laboratory, Cambridge.

Today I arrived in Cambridge, where I am staying for next two weeks, while I work on a native backend for Links with effect handlers. My plan is to target either the surface syntax of OCaml or the OCaml IR directly. I have been invited by KC Sivaramakrishnan, whom is working on adding multicore support to OCaml using handlers. I plan to do something similar with Links, albeit my focus at the moment is on concurrency and not on parallelism. However, I might obtain multicore support for free.

KC gave me an overview of the OCaml compiler and the multicore project today. I will give a brief overview/summary of our discussion. Read more.

Programming with Effect Handlers in Links (1)

 

Plotkin and Power's algebraic effects [1] combined with Plotkin and Pretnar's handlers [2] provide an approach to model and control side-effecting computations. In my master's project at the University of Edinburgh I implemented effect handlers in the functional programming language Links [3]. Through a series of posts I will demonstrate how to program with effect handlers in Links. In this first part, we will mainly focus on building intuition for programming with handlers by using a coin toss as a running example. Our coin toss model will encompass effects such as nondeterminism and exceptions. Through handlers we will give different interpretations of a coin toss. Read more.

See the archive for older posts.