Programming Models, Part 4: One Week in Louvain-la-Neuve

21 Dec 2014

As we discussed in our first post, Derflow is the now former name of our distributed deterministic programming model that is the basis of our research into providing a more expressive way of working with CRDTs and eventual consistency.

I’ve just returned from a one week vacation in Louvain-la-Neuve, Belgium, where I spent the week working with Peter Van Roy at Université catholique de Louvain on what was formally referred to as Derflow. It was an amazingly productive week where we had a chance to iterate on the model, prototype some new features, and fix a selection of bugs.

Highlights

  • We renamed the programming model. Derflow wasn’t the greatest name in the world and was a product of focusing on the research over marketing. Inspired by Lisp, short for list processing, we’ve chosen Lasp as the new name for Derflow, following the same pattern for deriving the name: Lasp for lattice processing.
  • Pushed 59 commits, which resulted in over 6,000 additions and 4,000 removals, closing twelve GitHub issues. This work can be viewed on GitHub. This involved both new feature work and bug fixes.
  • Loïc Schils, a graduate student at UCL will be doing his master’s thesis, supervised by Peter and I, on our programming model and will begin contributing to the project in January 2015.
  • Viet Bui, a graduate student at UCL, supervised by Peter and Manuel Bravo, will be looking at doing explicit causality in our programming model as well for his master’s thesis.
  • Adrian Colyer highlighted our work on his blog.
  • Peter convinced me that lightning has struck four times, in the same place.

It was an amazing week, and I can’t wait to return in January.

comments powered by Disqus