Skip to content
/ nuscr Public

A toolkit to manipulate Scribble-style multiparty protocols, based on classical multiparty session type theory.

License

Notifications You must be signed in to change notification settings

nuscr/nuscr

Repository files navigation

𝝼Scr (nuScr) CI status

A toolkit to manipulate Scribble-style multiparty protocols, based on classical multiparty session type theory. The toolkit provides means to define global protocols, project to local protocols, convert local protocols to a CFSM representation, and generate OCaml code for protocol implementations.

You can quickly try the live version here.

How to install

The easiest way to install is to use opam via

  opam install nuscr

then you can check your installation via

  nuscr --help

which prints help and usage information.

Development

We use the dune build system. To contribute, clone the repo and run

  opam pin . --with-version=dev

to pin the nuscr version to dev, and then run

  dune build

to build.

You can run the executable by

  dune exec nuscr -- [args]

where [args] are the arguments to the main CLI program.

You can run the tests via

  dune runtest

When you submit a pull request, the continuous integration system will build, run tests and check the code formatting. You may find utils/fmt.sh helpful for formatting the code.

Please also remember to update the CHANGES.md file in order to keep track of the changes in each version.

Goals and Non-goals

Our vision of nuScr is to provide a simple and lightweight implementation based on multiparty session types (MPST). We aim to make it easy for researchers to extend the original MPST, and to prototype implementations. A future goal is to provide a certified implementation (extracted from Coq).

With that said, nuScr is not ready for production (better use Scribble), and it will not be compatible with Scribble. Scribble has been extended beyond the original MPST to model checking with automata, whereas we wish to stick to the top-down approach of original MPST. NuScr is likely to be not so user-friendly in the current form, and may remain so in the future.