Pascal Cuoq and Damien Doligez. Weak pointers and Hashconsing in Ocaml 3.10.2
Abstract: This article describes the implementations of weak pointers, weak hashtables
and hashconsing in version 3.10.2 of the Objective Caml system, with focus
on several performance pitfalls and their solutions.
Benjamin Canou, Vincent Balat and Emmmanuel CHAILLOUX. O'Browser : Objective Caml on browsers
Abstract: We present a way to run Objective Caml programs on a standard, unmodified web browser, with a compatible data representation and execution model, including concurrency.  To achieve this, we designed a bytecode interpreter in JavaScript, as well as an implementation of the runtime library.  Since the web browser does not provide the same interaction mechanisms as a typical Objective Caml  environment, we provide an add-on to the standard library, enabling interaction with the web page.  As a result, one can now build the client side of a web application with the standard Objective Caml compiler and run it on any modern web browser.
luc maranget. Compiling Pattern Matching to good Decisions Trees
Abstract: We address the issue of compiling ML pattern matching
to compact and efficient decisions trees.
Traditionally, compilation to decision trees is optimized
by (1) implementing decision trees as dags with maximal sharing;
(2) guiding a simple compiler with heuristics.

We first design new heuristics
that are inspired by necessity,
a concept from lazy pattern matching that we rephrase
in terms of decision tree semantics.
Thereby, we simplify previous semantical frameworks
and demonstrate a straightforward connection between necessity and
decision tree runtime efficiency.

We complete our study by experiments, showing that optimizing compilation
to decision trees is competitive with the optimizing match compiler
of Le Fessant and Maranget (2001).
Jérôme Vouillon. Lwt: a cooperative thread library
Abstract: We present a cooperative thread library for Objective Caml.  The
library is entirely written in Objective Caml and does not rely on
any external C function.  Programs involving threads are written in
a monadic style.  This makes it possible to write threaded code
almost as regular ML code, even though it has a different semantics.
Cooperative threads are especially well suited for concurrent
network applications, where threads perform little computation and
spend most of their time waiting for input or output, at which time
other threads can run.  This library has been successfully used in the
Unison file synchronizer and the Ocsigen Web server.
Jean-Christophe Filliatre. A Functional Implementation of the Garsia-Wachs Algorithm
Abstract: This functional pearl proposes an ML implementation of
  the Garsia-Wachs algorithm.
  This somewhat obscure algorithm builds a binary tree with minimum
  weighted path length from weighted leaf nodes given in symmetric
  order. Our solution exhibits the usual benefits of functional
  programming (use of immutable data structures, pattern-matching,
  polymorphism) and nicely compares to the purely imperative
  implementation from The Art of Computer Programming.
Sam Lindley. Many holes in Hindley-Milner
Abstract: We implement statically-typed multi-holed contexts in OCaml using an
underlying algebraic datatype augmented with phantom types. Existing
approaches require dynamic checks or more complex type systems. In
order to support concatenation we use two type parameters to represent
the number of holes in a context as the difference between two Peano
numbers. In order to support plugging a context with contexts of
different arity we introduce a datatype of lists of contexts of length
n with a total of m holes. Further, we extend our representation
to allow holes to be marked with additional type information. As an
example, we use these marks to implement statically-typed multi-holed
XHTML contexts. We take advantage of Garrigue's relaxed value
restriction.
Alec Heller and Jesse Tov. Caml-Shcaml: An OCaml Library for UNIX Shell Programming
Abstract: Objective Caml is a flexible, expressive programming language, but for manipulating Unix processes, nothing beats the terseness and clarity of Bourne Shell:

    ls *.docx | wc -l

To achieve the same effect in C requires several more lines of code and very likely a glance at the manual page for readdir(3).  Despite OCaml's excellent Unix module, which provides access to a wide variety of system calls, as simple as counting the ".docx" files in the current directory is hardly easier in OCaml than in C.  The code looks largely the same.

Caml-Shcaml addresses this problem by bringing high-level abstractions for Unix systems and shell programming to OCaml.  In particular, we take advantage of OCaml's type system to offer statically-checked data pipelines.  External Unix processes and internal OCaml stream transducers communicate seamlessly within these pipelines.  Caml-Shcaml brings other essential systems concepts into the world of typed functional programming as well, including high-level interfaces to Unix facilities for I/O redirection, signal handling, program execution, and subprocess reaping.
Michael Rainey, John Reppy and Matthias Blume. Calling variadic functions from a strongly typed language
Abstract: The importance of providing a mechanism to call C functions from high-level languages has been understood for many years and, these days, almost all statically-typed high-level-language implementations provide such a mechanism. One glaring omission, however, has been support for calling variadic C functions, such as printf. Variadic functions have been ignored because it is not obvious how to give static types to them  and because it is not clear how to generate calling sequence when the arguments to the function may not be known until runtime. In this paper, we address this longstanding omission with an extension to the NLFFI foreign-interface framework used by Standard ML of New Jersey (SML/NJ) and the MLton SML compiler. We describe two different ways of typing variadic functions in NLFFI and an implementation technique based on the idea of using state machines to describe calling conventions. Our implementation is easily retargeted to new architectures and ABIs, and can also be easily added to any HOT language implementation that supports calling C functions.
Andy Gill, Johan Nordlander and Magnus Carlsson. Unrestricted call-by-value recursion
Abstract: Call-by-value languages commonly restrict recursive definitions by
only allowing functions and syntactically explicit values in the
right-hand sides.  As a consequence, some very appealing programming
patterns that work well in lazy functional languages are hard to apply
in a call-by-value setting, even though they might not be using
laziness for any other purpose than to enable the desired form of
recursion.

In this paper we present an operational semantics as well as a
straightforward implementation technique for unrestricted recursion
under call-by-value.  On that basis we are able to demonstrate that
highly recursive programming idioms such as combinator-based parsing
are indeed compatible with call-by-value evaluation.