Unlearning the Tidyverse R My motivation to unlearn the Tidyverse after eight years using it. Author Mauricio “Pachá” Vargas S. Published June 29, 2026 Why I have been working in a project with the University of Texas that showed some ABI issues (e.g., incompatibilities between software) that ultimately led me to propose a few changes to the ‘httpuv’ package, a package central to multiple parts of the Tidyverse and also Shiny. I think Hadley Wickham is an amazing person and I met him at Latin R and the RStudio Conf over the lapse of a few years before COVID. This is nothing against people and I know that Hadley does not maintain ‘httpuv’. Initially, I started working on adapting ‘httpuv’ to minimize its dependencies and do it in a very old fashioned way using the R’s C API. I am quite confident using it even when the notation feels like a time travel to the DOS days. I have used computers since I was 5 years old and I got my first computer to start Kindergarden back in 1995 and I learned commands such as “CD GAMES/” and “KEEN.EXE” to play Commander Keen back then. What made me feel very dissapointed was what I consider being gaslighted in a public forum. For PR 437 I read things such as “this is AI” and “you do not understand software development”. We all come from different places but starting from there is probably not the best start. Furthermore, after answering a few questions about a PR I worked on for two weeks because I want to keep UT staff happy, I decided to open a new PR to do the same using Posit’s cpp11 (PR 438 ), which did not receive any comments despite my follow up on GitHub and via email. I started using R back in 2015, initially I struggled a lot and I started with data.table. A few years after starting with R, I was told about the Tidyverse which, besides the struggles to install it on Linux, I find it is an amazing end-user tool. Meeting comments like what I found with the ‘httpuv’ package does not feel like “this is the R way”. I have co-maintained the CRAN WebTechnologies task view for a few years now, I am soon attending an R sprint to address bugs in R base, and I maintained a wide range of packages including wbstats and tradestatistics . Once in a while I receive PRs to the multiple repositories I have, including the Shiny-dependent d3po package, which is one of the very few fully FOSS packages created as an alternative to the excellent highcharter package made by my friend Joshua Kunst (highcharter depends on highcharts which requires payment for some uses). As side contributions, I have dedicated a few years of my life to improve pointblank , an excellent package for data validation. In other words, R has been a significant percentage of my life and I have used it from 9 to 5 Mon-Fri for over a decade now. This is why I do not appreciate certain forms of communication, especially after having been involved with a global community where I’ve met fantastic people (and unfortunately people that have insulted me over the email, which is not the ‘httpuv’ case). At some point I considered myself to be some random guy from Chile living in the far end of the world. Now I consider myself to be the same infinitesimal person but living in London and writing a PhD thesis where two-thirds of it are about improving R for Econometrics and I am glad that during all these years I played a major role in making R Para Ciencia de Datos happen. Stay tuned. I am really rusty with apply() , mapply() and most of base R that is not %*% and t() . I have just removed the Tidyverse from my laptop and I will be posting how do to common data cleaning tasks with minimal typing in base R.