It's an extremely cursed idea... but it kind of makes sense. I can't stand R but things like dplyr and ggplot are incredible. I wouldn't mind having access to them in a sane shell.
There is a reason that statisticians love R. It's because the alternative to R for statistical work is SAS, not python. Compared to SAS, R is a godsend.
I'll just point out that R started as a Scheme dialect. If you dig into the source, you'll see that the main C data structure is the SEXP, for s-expression. And as the other comment says, you can call everything as a function. You can write an s-expression form of R in about 30 minutes. Literally all you have to do is identify opening and closing parens and rewrite in `function`(arguments) form.
Ooooooh wow, I remember digging into R internals a while ago for something and seeing `sexp`, and only NOW am I connecting it to Lisp's SEXP!! That makes so much sense!
Does RData derive from LISP images? Because if so, that'd be quite the revelation to me.
RData only stores the environment of the session. It doesn't store the current search path (you need to reload libraries). I believe it's not a full binary image like LISP
I submit R is easily the most misunderstood language among programmers.
And it's not just under the hood where it's very lisp like. Heck you can write it with an almost lisp notation if you really wanted to, `+`(1, 2), and the syntax is incredibly pliable. I'm a big fan of lisp and I'd argue it's one dialect of lisp I actually get to use.
Here I am preferring the base plotting. To each their own. I read words like "aes" "geom" "grammer of graphics" and my eyes glaze over. plot(data, main="title" etc.) is easy enough for me.
There's a few arbitrary codes/conventions to learn at the start. My take is not to look too deeply into them - every DSL/framework has them and in ggplot2's case they only take a very short time to get used to before you have the flexibility and incredible power/productivity of ggplot2 at your fingertips!
If the plot is simple, ggplot may seem like overkill if you aren’t familiar with the library. But once you start wanting to customize things substantially or doing non-simplistic plots ggplot becomes much easier to use than base graphics.
I could imagine that proper Lisp-like metaprogramming might come in handy. R can do metaprogramming, but it's quite tricky, even with all the tidyeval tooling.
Curious to know what you/others consider to be R's warts? I have my own gripes with it, but mostly love it. I ask because it's interesting to learn what's missing form a language, what other languages do better, etc.