When you start working in the programming language theory business, you'll soon find out that
lambda abstraction and functions break many nice ideas or, at least, make your life very hard.
For example, type inference is easy if you only have
var x = ..., but it gets hard once you
want to infer type of
x in something like
fun x -> ... because we do not know what is
x. Distributed programming is another example - sending around data is easy, but
once you start sending around function values, things become hard.
Every programming language researcher soon learns this trick. When someone tells you about a nice idea, you reply "Interesting... but how does this interact with lambda abstraction?" and the other person replies "Whoa, hmm, let me think more about this." Then they go back and either give up, because it does not work, or produce something that works, in theory, well with lambda abstraction, but is otherwise quite unusable.
When working on The Gamma project and the little scripting language it runs, I recently went through a similar thinking process. Instead of letting lambda abstraction spoil the party again, I think we need to think about different ways of code reuse.
Olympic Games are perfect opportunity to do a fun data visualization project - just like New Year, you can easily predict when they will happen and you can get something interesting ready in advance. I used this year's Games in Rio as a motivation to resume working on The Gamma project. If you did not read my previous article, the idea is to build tooling for open, reproducible and interactive data-driven storytelling. When you see a visualization, not only you should be able to see how it has been created (what data it uses & how), but you should also be able to modify it, without much programming experience, and look at other interesting aspects of the data.
The first version of The Gamma project tries to do some of this, using historical and current data on Olympic medals as a sample dataset. You can play with the project at The Gamma web site:
Without further ado, here are the most important links if you want to explore how it all works on your own. However, continue reading and I'll describe the most important parts!
The project is still in early stages and the code needs more documentation (and ReadMe files, I know!) However, if you would be interested in using it for something or you have some interesting data to visualize, do not hesitate to ping me at @tomaspetricek. Also, thanks to the DNI Innovation Fund for funding the project and to the Alan Turing Institute for providing a place to work on this over the coming months!