Last November, I learned the very sad news from Michael Friendly that John Fox had passed away. That brought my memory back to 2006 when I emailed John for the first time asking for his help on a problem about Structural Equation Models (SEM), which had puzzled me for months. I reached out to him because I had seen an appendix on SEM (freely available on his website) in his book, An R and S-PLUS Companion to Applied Regression. He patiently guided me all the way to a specific section in Greene’s Econometric Analysis, which finally solved my problem.

At that time, no one in the R community had heard of me, and I was just a beginner. You can imagine how grateful I was to receive such helpful replies from a professor at McMaster University. One funny anecdote is that when he first told me to find the derivation in Greene’s book, I quickly found a PDF copy online (I don’t remember where now), thumbed through it, and had no idea where to find the needle in the haystack, so I asked him again and attached the whole PDF. He was shocked because he thought I would go to a library to check out the book. I was embarrassed, but he looked into the PDF and found the section for me anyway. I imagine he burned the pirated copy in horror immediately after that… When I re-read those emails today, I also feel a little embarrassed about my poor English at the time.

My next major correspondence with John was in May 2008 when I submitted my first English paper to R News (now The R Journal). John was the Editor-in-Chief at the time. The paper was about the animation package. Again, John was very helpful and patient during the review and editing process. I was lucky to have him as the editor for my first English paper!

Then in August that year, I attended the useR! conference for the first time in Dortmund, Germany. I had a chance to meet him in person in the conference hall. I don’t remember what we talked about (other than my thanking him for his help), but my impression was that he was a very kind and humble person. Nowadays, I rarely take pictures, but I’m glad that I brought a camera (I think it was a Nikon) to the conference and took a picture with him:

Between 2006 and 2009, I used his famous Rcmdr package extensively and introduced it to many people. One reason I liked it was that although it was a GUI package (which is usually friendly to beginners), it also showed me the R code that was executed when I clicked on the GUI, so I could learn R this way as well. Another package of his that I used frequently was car (in particular, the recode() function, mainly because SPSS was the most popular statistical software at my school at the time, and this feature was frequently used in SPSS).

Rcmdr sparked my interest in building GUI applications in R, although I later found that Tcl/Tk was not my cup of tea. After I came to the US in 2009, I gradually fell in love with John Verzani’s gWidgets. Then a couple of years later, Shiny came out, and I was naturally attracted to it. The rest is history.

Thank you, and rest in peace, John. You will be missed by many people in the R community and beyond.