case doesn’t evaluate the left sides of it’s clauses.

https://stackoverflow.com/questions/3782970/how-can-i-display-the-definition-of-a-function-in-clojure-at-the-repl

In the way you defined it, no. (+ x y) evaluates to z and unless you re-read the source code there is no way

You can use macros to have that info though (give me a moment to work up an example

So one example, if you want the original form to have the definition as metadata

You can also use the :file metadata along with the line and col info to read your own source and extract it out

You might also be interested in https://github.com/xsc/rewrite-clj

It does not let you take a compiled function and extract the syntax tree from it, but it does have some tools that are somewhat related to the topic, so maybe some interesting stuff to explore there.

I believe it is implemented in Rust

Source: http://tank.hyperfiddle.com/:dustingetz!niko-gobel-differential-dataflow/ C-f "Rust"

Thanks for the link and doing that interview!

Maybe I was expecting these incremental processors to be something more fundamental and composing like a transducer and they’re actually just lots of hand tuned mutation code?

Differential Dataflow's raison d'être is to generalize incremental computation to work in a distributed setting. The two fundamental algorithms are progress tracking (implemented in Timely dataflow) and the logic of partially ordered times. Any incremental computation requires a previous state to make use of. In a distributed setting, determining the previous state to use is a hard problem, which Differential solves for you. This is almost a separate issue from how you express incremental computations themselves. Differential comes with a bunch of incrementalized operators built-in (hand tuned mutation code, if you will), which happen to be great for many common use cases (Datalog evaluation being one of them). But there are other ways of expressing incremental computations, which could all be made to work with Differential, e.g: http://www.umut-acar.org/self-adjusting-computation https://www.youtube.com/watch?v=R3xX37RGJKE

FWIW I too am interested in simplified implementations of the core concepts for educational purposes, but I probably won't have the time to work on them anytime soon.

Thanks Nikolas - this information is super helpful. I’ll be digging into all this for a while I think 🙂

Feel free to send me a mail at any point!