A Pamphlet against R | Hacker News Reader

A Pamphlet against R

panicz.github.io

So, are there out of the box libraries to estimate a zero inflated negative binomial regression model in guile? How about to create multi-faceted plots with < 10 lines of code?

I would LOVE to do my statsing with any lisp variant, but the reality is that R has a killer ecosystem of statistical and visualization packages. Python had an enormous uphill battle but now has a really great story to tell in part because it hitched itself to the rising fortunes of Bayesian stats.

Any discussion of languages has to include amount and maturity of libraries for the purposes at hand.

This PDF has some good content. But I feel a somewhat harsh response is warranted if your starting premise is positioning yourself against an existing solution that many individuals use.

There's RCL which is a lisp interface to R.

https://common-lisp.net/project/rcl/

A bit long in tooth but might get you what you want.

Cheers! Though my personal, someday dream, is to port R to Racket such that I can freely interop. It shall be called "Arket" and verily be googleable. If someone beats me to my nerdy, nerdy dream, so much the better.

If we talk about the Lisp family of languages (Lisp, Scheme, Racket, Logo, Dylan, ISLisp, Picolisp, MDL, etc. etc.), then R is actually a Lisp dialect. From the past. R's basic technology is rooted in the Lisp's of the 70s. Even Fexprs are back.

And was actually implemented on top of scheme originally. Ross Ihaka talks about it in his "back to the future" paper.

It is mentioned on page 2: "However, the case of the R programming language is particularly interesting, because initially it was just a harmless implementation of Scheme; but then as a result of the irresponsible experiments of mad scientists it mutated into a monster [3].", where [3] is the link to https://www.stat.auckland.ac.nz/~ihaka/downloads/JSM-2010.pd...

R turned into a language which used older Lisp features. It no longer resembles Scheme, but both older and then newer Lisp systems. It even provides OOP derived from CLOS. It's really a vey strange language which provides/uses a lot of features, which can make it difficult to maintain (my guess having read about the language, but have not used it). For example a mix of lazy evaluation and fexprs can make understanding and predicting the behavior of code very difficult.

Before R there was a somewhat healthy LispStat ecosystem, more recently there is: http://www.cliki.net/common-lisp-stat

An 85 page pamphlet.

Someone is rediscovering literate programming where the code is so meaningful it serves as documentation of itself.

I vote for a troll disguising a boring scheme tutorial in a satire.

But it is very metaphysical à la Artistote or Giodarno Bruno ... I would say like almost any stuff written by the functional programming adepts. I am not sure if FP is mathematics or philosophy sometimes. I like them like some weird religious monks producing free beers. I drink the beer, I avoid the church.

But I would say that even if I do not like scheme, I value meaningful code too.

This is performance art.

Awesome! I am going to read this pdf.

I'd love a summary when you're done. It looks interesting but intimidatingly long.

TLDR; Has nothing to do with R (barley mentioned, pamphlet title is clickbait). Advocates (and teaches) Scheme for computational intelligence. Might be satire, or might be a tutorial on Scheme ... I honestly don't know. Ends with the line "$ chmod -R 666 /" making me lean towards satire/intelligent trolling. Is nicely typeset.

Reply on news.ycombinator.com