letter_to_tufte.txt

Dear Dr. Tufte,

I was inspired by your post and challenge regarding slope graphs, and while I know your challenge has been completed, I have a few thoughts and I am interested in your perspective.

Specifically, I have recently been very concerned with the media used for presenting scientific information, and the resources available for scientists in standardizing these media. It seems as though a shift has occurred in the presentation of science, away from print media towards digital media; though LaTeX and typeset scientific results remain an important component of scientific production.

There seems to be some heterogeneity in the technologies used in generating slope graphs, and I'm curious about this. A practical consideration is the intended medium of a chart's presentation, and I fear many of the submissions for your challenge are lacking in this department.

It is important for chart-generating tools such as you solicited be available in the final format it will be presented in, and for scientists, I envision this as digital (websites) and print (LaTeX). The first possible standardized solution is to merely use a graphical format such as PNG or GIF, which is portable across horizons; I think it is telling that in your initial post, you posted the graphic as a GIF!

On top of this, though, I think it is important that there be implementations in a native language---and that we limit the number of languages in use, so that as practicing visualizers, there is no "paradox of choice" with too many options to choose from, for a basic product.

All of this is to say that a collaborator and I have implemented slopegraphs in three formats, and we believe that from the perspective of a publishing scientist, this is an important innovation: we implement slope graphs in PNG format, using Python and its package matplotlib; we implement a LaTeX version, for journal publication; and we include a web-based D3.js implementation.

We believe the medium is as important as the message, and that these three formats---PNG, typeset PDF, and HTML-based---are the three most relevant to practicing scientists. They allow for portability on the one hand, and native implementation on the other, such that they can be used in a variety of settings.

What are your thoughts on the recent glut of visualization software? Do you think it is time to start standardizing?

Best,
M. Henry Linder