@josiah I mean, we had @jnolis writing about using R in prod at T-mobile 5 (!) years ago! And the context there implies that it's getting hit a bunch, if it's being used to classify customer responses.
@krzysztof_dyba I can check! I recall doing the comparison with geometries a few months back and jsonify edged out rjson in that case. There's also the license implication of GPL-2
@josiah Beyond your deft troll handling, what I really like about this post is that you invite actual resolution of the issues — like, demand performant solutions! Open source encourages this and has room for many ways to solve problems! — rather than sneer about What True Devs Tolerate.
@josiah Your blog post is 🔥 our slack channel (and yes we have "R in prod" albeit not something with a huge traffic load). Two of our SWE know R and can translate it into :python: or some JS.
@josiah At my university, no one uses R anymore in the scientific and numerical fields where it is supposed to shine. Of course, no one uses it in computer science too. Either Python or Julia.
I wouldn't put any of R, Python or Julia in production though. In my own opinion, they should only be used for interactive programming and scripting (about one file).
@mo8it interesting. I think not putting Python or R into production is a bad take. Good R and good Python is fast and typically uses interfaces to fast things written in C++ or C or Rust or Fortran. It’s a bit elitist ngl
And the Julia thing is super niche :P but VERY fast!
@mo8it maybe elitist is the wrong word? But it implies that only low level languages like C/++ / Rust should be put into prod which is a yuge ask for most
Never met a developer who liked R, and to put things into production, you have to work with devs.
The new people coming into data science all know python. Even among older data scientists, you rarely find one that prefers R (unless they're a statistician working in academia).
Who is going to maintain that code after the person who wrote it leaves for another company?
@orizuru Its a lot like how I was hired and I have to help edit / maintain C# tests. Even with your presumed stance of the world, its a pretty normal thing.
@orizuru@shibaprasad it’s absolutely terrible. Idk how you were using plumber on databricks though?? I gave them a whole slew of things to improve for R users. But they don’t care. Their customer facing team knows pandas and that’s it it feels like
@josiah@orizuru Haha kinda had the same experience. We had a 3 hours long interaction with them, they were clueless for most of the R stuffs. And our team mostly uses R. For ML-OR-Analytics, everything.
When we attended a hands-on workshop, the presenter said "R lovers, sorry to say, it is time to move on. It used to be cool but not anymore." (paraphrased)
@shibaprasad@orizuru nah. They’re wrong. They just don’t want to learn. It’s a defense mechanism for a bad platform. They’re stuck in 2011 with these notebooks.
@jrosell quarto also supports python. So it's a hard sell.
The thing R has got going for it is the plots, they are pretty.
I'm sorry, but the rest just lags behind, from the syntax to the ecosystem (except very specific stats packages).
Just having array indexes starting at 1 will make a Dev's skin crawl.
Dev are fragile creature and they cant handle Julia or Fortan (yup index starting at 1 or "pick your number"). Anyway I am always amazed how everyone can have enough confidence to claim something is dead from punk, to relational database or more recently data science.
Add comment