The second OCaml-Paris meetup took place on May 21 at IRILL; I think it was a very interesting event, with good presentations on surprisingly diverse and interesting topics.
I made a presentation on a topic suggested by Fabrice Le Fessant: how to interact with the upstream OCaml development. You can get the slides here, as a gorgeous HTML presentation; but I’m afraid it may feel a bit bland without the audio.
Of course, the topic is a bit limited; I think that in the Grand Scheme of Things – at least on an OCaml scale – contributing to the OCaml compiler is possibly a minor point in the space of potential improvements that could benefit the community as a whole. I tried to open the perspective a bit in my last slide, with some suggestions of other things to do, and that may be seen as a contribution to this year’s edition of the “how to conquer the world” debate currently ongoing on the caml-list.
I very much enjoyed all the talks. Esther Baruk gave a high-level description of Lexifi’s activity (besides “paying Alain Frisch to do cool things”). It was a balance of the abstract idea of bringing the elegance of formalism (the very kind that abounds in functional programming) to financial contracts, and practical use of a programming language (a dialect of OCaml) as a product to sell to finance companies.
Benoît Vaugon talked about the OCaml bytecode, not only as a backend for the OCaml compiler, but as an input format for various OCaml-related tools: dead code elimination, his embedded PIC runtime, and finally his OCamlCC project, compiling bytecode programs to native C code. It was extremely interesting, but probably more technical than the audience expected. It is difficult not to be a bit overwhelmed when you are presented a subtle program analyzer, an implementation of the whole OCaml runtime in 8-bit assembly, and an optimizing compiler to C all at the same time.
Finally, Frédéric Bour talked about Merlin, a project that provides semantics-aware editor modes for OCaml for both Vim and Emacs, and promises to scale neatly to even more cool features and other editors. I suspect the presentation itself was a bit rough, but the software demo was definitely impressive enough. I discussed a bit Merlin’s happy design choices in the caml-list announce thread.
We then had some time for informal discussion (and pizzas), and I enjoyed seeing a lot of different people I sometimes hadn’t heard about in a while.
The sources of the slides are actually a Markdown file and the compilation process is extremely simple, using Pandoc with the compilation options detailed at the end of the slides sources – an idea I stole from Nicolas Pouillard. I wouldn’t use that for a technical presentation with maths, but it was useful to be in a browser to easily follow links.
I think it would be nice to have a webpage to aggregate the slides of all presenters. I opened a suggestion on the ocaml.org issue tracker, but didn’t get an answer so far.