The best writing tip I’ve received

Probably the best writing tip that anyone has given me is that the last sentence of any given paragraph I write should often be the topic sentence of the next paragraph. Once this was pointed out to me, I noticed it in almost all of my writing and often in student writing as well. This error probably happens because it’s natural to build up a paragraph until it connects to the next point to be made. But how we write is not how we read. Often I can go through a first draft and move the last sentence of many paragraphs to the beginning of the next.

Bonus tip: When a colleague read a draft of my book manuscript, he said that I should try to shorten break up the paragraphs, because the paragraphs would appear longer on the printed page. Sure enough, when the proofs arrived, he was right.

My year in tech 2016

Following Alan Jacobs and Boone Gorges, here is a brief report on my year in technology.

Compared to the list of technologies that I used in 2014, not much has changed. My preferred tools for writing remain Markdown, Pandoc, LaTeX, and Vim, with Zotero for citations. I’ve never had a system for taking notes that entirely pleased me. After a brief flirtation with an open wiki notebook, following the lead of Caleb McDaniel and Jason Heppler, I gave it up this year. The idea still seems sounds, but in practice I’m unwilling to have all my notes public, the separation of public and private notes was onerous to maintain, and a fair bit of my note-taking is in the form of “lab notebooks” for which a different system is necessary. At the moment all of my notes are in plain text in a Dropbox directory, which I can access on iOS with 1Writer or on computers in Vim. Perhaps someday I will return to a private wiki.

For computational history, R maintained pride of place. It is astonishing how much better and easier to use the language has become since I started to teach it to grad students a couple of years ago. For one project I was able to do everything from machine learning on Mason’s computing cluster to interactive visualizations in Shiny. I foresee a lot more D3.js in my future, but mostly to make interactive R packages.

Continue reading “My year in tech 2016”