Previous week Up Next week


Here is the latest OCaml Weekly News, for the week of June 28 to July 05, 2016.

  1. how to encourage adoption of OCaml?
  2. js_of_ocaml 2.8.1
  3. Other OCaml News

how to encourage adoption of OCaml?


Deep in this thread, Jeremy Yallop said:
> It is hard for me to judge because I came through RWO, but it appears to me
> that the lack of consensus on standard library comes up pretty quickly.

I think the standard library situation is much less of a concern than
it once was, now that it's easy to distribute small OCaml packages and
manage dependencies.

In the past distribution difficulties discouraged dependencies: for
example, even though many people prefer the design of ocaml-re and
ocaml-pcre to the regular expression facilities in the standard
library, the administrative overhead of an additional dependency made
the standard library the easier choice overall.  In that situation
it's desirable for the standard library to be large and featureful.
Nowadays there's much less benefit to having regular expression
support in the standard library, since depending on ocaml-re or
ocaml-pcre is just a matter of adding a line to an opam file and a few
lines to the build configuration.

The standard library still has a useful role to play, since it's
easier to make libraries interoperate if they can communicate via
common types, and several recent and proposed changes have that kind
of role in mind.  For example, the latest release of OCaml added a
'result' type to the standard library, which was previously defined in
incompatible but essentially equivalent ways in several libraries:

and there's a proposal for adding iterators to various container types
for the next release currently under discussion:
Dean Thompson then said and Yaron Minsky replied:
> From my understanding so far, it seems to me that mixing and matching Core
> and not-Core would be tough? Everything from result types to Lwt vs Async?
> Given the inspirational and educational power of Real World OCaml, many
> newcomers will face this issue.

A few thoughts: 

As Anil said, we're working on an updated RWO, which should resolve the camlp4

As for mixing and matching between libraries that do and don't depend on Core,
there's actually little difficulty here. Core sticks to the standard interchange
types (array, string, option, list, char, and now result) that are provided by
the stdlib, so whether you use Core (or Core_kernel) becomes more a matter of
personal preference, and shouldn't hinder interoperability.

One remaining problem with Core is the minimal executable size, which is
currently much bigger if you use Core. We're considering some work in three next
few months to make this much better.

Async and Lwt are a real problem. They provide very similar functionality, and
mixing and matching between two schedulers is not so easy. I'd love to see some
resolution here, but it's not clear what the solution would be. Perhaps once we
resolve the executable size issues of Core, there will be more appetite for some
kind of merger of the two libraries. In the meantime, we're highly committed to
continuing development and support for Async.
Ivan Gotovchits then said:
The solution would be to use the same approach as with standard types. We need a
common base inductive type for `Lwt.t` (aka `Ivar.t`), which will represent a
value which is defined in some point in the future (hence a `future` is a good
name). Another type is for capturing a concept of a variable that can have
multiple values in the future, that is represented as `Lwt_stream.t` or `Pipe`.
Currently in both Lwt and Async the main thread type is tightly coupled with the
underlying implementation, especially in Async (Lwt.t can be easily decoupled). 
Yaron Minsky then said:
I'm not at all sure that the decoupling is possible or wise for Async.
My intuition is that this is too complex of a problem with too much
need for careful optimization to be able to have a simple, shared
generic data structure for this.

The solution that seems most plausible to me is to settle on one
implementation, and port the API of one library to run on top of the
other. There was indeed an experiment in this direction that was done
by Jeremie Dimino:

That said, until we resolve the binary size issues with Core and
therefore Async, I doubt that this solution would be appealing to the
full community of lwt users.
Hendrik Boom asked and Yaron Minsky replied:
> What are the conceptual differences between Async and lwt?  Does either
> of them manage to take advantage of a shared-heap multicore system?

The conceptual differences are pretty thin; there are some error
handling differences, but they're not huge.

As for the shared heap question: Async greatly improves the
programmers ability to reason about concurrent code by providing clear
bounds on when interleaving can happen.  In particular, a given job
scheduled to run as part of Async can never be interrupted by another
async job. That means interleavings can effectively only happen where
you use a bind operator to string two closures together, and so one is
rarely exposed to race conditions. That means that Async program
typically don't need to think much about locking, and race conditions
stop being your primary source of bugs.

Using Async in the context of shared heap parallelism gives up these
guarantees, and puts you back in mutex/sempahore/condition-variable
hell.  My guess is that we'll end up taking advantage of the multicore
GC by running multiple Async schedulers, one per domain (e.g., OS
thread), and having carefully written primitives for sharing
data-structures and efficiently sending immutable messages between
these domains.  But I think just freely scheduling Async jobs across
multiple physical threads seems like a disaster from the point of view
of producing comprehensible, reliable code.

js_of_ocaml 2.8.1


hugo announced:
I'm happy to announce the release of js_of_ocaml 2.8 & 2.8.1.
They are already available via OPAM.

Js_of_ocaml is a compiler from OCaml bytecode to JavaScript. It makes it
possible to run pure OCaml programs in JavaScript environment like browsers and

Here are some notable changes:

Support for OCaml 4.03

Async_kernel (and Core_kernel) support
In the findlib package js_of_ocaml.async.
Note that it is only available with a 32bit OCaml compiler.

Dynlink and separate compilation
One can now compile OCaml module (.cmo) and OCaml library (.cma) to JavaScirpt.
Loading theses JavaScript files is enough to dynlink them.

Improved JavaScript code generation
* preserve as much as possible OCaml variable names (in pretty mode)
* Better source map support
* Drastically reduce memory usage when assigning short variable names

The full change log is available on github

Happy hacking

Other OCaml News

From the ocamlcore planet blog:
Here are links from many OCaml blogs aggregated at OCaml Planet,

Minimising the virtual machine monitor

Behavioural types

Software Engineer (ClojureScript, F#) at Xledger (Full-time)

FP Meetup: OCaml, Facebook and Docker at Jane Street,_Facebook_and_Docker_at_Jane_Street

Old cwn

If you happen to miss a CWN, you can send me a message and I'll mail it to you, or go take a look at the archive or the RSS feed of the archives.

If you also wish to receive it every week by mail, you may subscribe online.

Alan Schmitt