Referee report of "mathml: Translate R Expressions to MathML and LaTeX".

• Id: 2023-47

General

This article introduces an R package, mathml, to translate R expressions to MathML and LaTeX. I find this package both interesting and useful. I'm not familiar with the underlying technology (Prolog), but the article has show many convincing examples. Being able to see the (LaTeX/MathML) math representation of R code is a great advantage, since it's much easier to read and understand the math when reading the formulas in PDF than reading them as R source code.

I think the major missing part of the article is a mention or comparison or comment of other existing work, e.g.,

• https://github.com/datalorax/equatiomatic
• https://github.com/cran/simsalapar/blob/master/R/expr2latex.R Examples: https://rstudio-pubs-static.s3.amazonaws.com/268872_483cb8ee959642b68e9cf2b84eda1e06.html
• https://github.com/tomicapretto/latex2r Response: References to equatiomatic and latex2r have been added to the manuscript (Introduction and Conclusion). For simsalapar, we refer to R's own plotmath syntax (Introduction).

Detailed comments

p2: I have no idea what "the predicate math/2" means, although I understand what it does. Sounds like a technical jargon. May be worth a brief explanation. Same thing for "ml/3" and "jax/3", etc. later.

Response: We added a few sentences explaining the Prolog-specific terminology (e.g., predicate), as well as the name/arity-notation.

p4: "parentheses are automatically placed around g + h in the nodot-example". Do you mean "d + e" instead of "g + h"?

Response: Thanks for noticing, this has been corrected.

p12: "it is currently not possible to use the functions of mathml for writing inline formulas, here, the user has to adopt the usual LateX notation". I'm curious why it's impossible to use `r mathout()`. It can be used inline if it can return a character string.

Response: The reviewer is correct, the limitation has been removed and an example with inline math is given in the introduction.

Minor correction of some names: RMarkdown -> R Markdown; Latex -> LaTeX; ShinyApp webpages -> Shiny Apps.

Response: The names have been corrected accordingly.

Code

The article defined a mathout() function in it. I wonder if it would be more convenient to just define it in the mathml package.

Response: Two functions mathout() and inline() have been added to the package.

To work better with R Markdown documents, it may be worth defining a `knitr::knit_print` method for the results returned from mathjax()/mathml().

Response: A function math() has been added to the package that does exactly this, thank you, nice improvement.