Rule no 1: No matter what you do as a data scientist, there's always gonna be someone doing something slightly more advanced or different telling you you're not doing "real data science"

@ekaitz_zarraga I'm wrapping Apache POI with Clojure, while I am at it I'm trying to create an intermediate edn representation of an Excel file that can be generated, read, stored and put under some kind of version control

@ekaitz_zarraga <- is used to distinguish name assignment from keyword arguments assignment in function calls. f(k = 1, z = y <- 3) will work as expected with the z argument having y as value that will be resolved to 3.

There are historical reasons for this and one of them is that R is lispy enough to have similar quirks. The issue is that most R users are not trained to code in a functional way, but you can do pretty powerful stuff with it.

With RStudio alt + - gives you <-

@yogthos exactly, but many developers are used to the Java, Python and JavaScript ecosystem where there are many "broken" things.

The other day a coworker was using etaoin (https://github.com/igrishaev/etaoin) and he asked me for a replacement because he couldn't find anything about using a proxy, so I told him: "Look at the code". He is new to Clojure, but in 20 minutes he found a perfectly fine solution and prepared a pull request.

I challenge anyone after just a couple of Python or Java projects to do the same

@aral the EU DPA should make available an open endpoint that can be called to dump whatever breach automatically if a DNT request is not honored or has the issues you explained