Tom MacWright

Microsoft Excel and the Folk Tradition of Technology

Right now I’m on a Megabus back from New York City to Washington DC. I just gave a talk at a conference called Visualized about my running maps and work on seeing the progress of OpenStreetMap’s attempt to redo several hundreds of years of building roads in seven or eight years of typing and drawing on computers.

The talk went okay; it was nothing to write home about. I’m relieved to be finished with it, since the lineup and fit-and-finish of the conference was intimidating in a way that replicated the pre-cross-country-race fidgety nervousness that doomed my high school athletic pursuits.

Anyway, that’s done now. The only interesting thing I really said was at the after-party, where I professed a love of Microsoft Excel and a hatred of professional criticism and then did a lackluster job of supporting that point.

Here’s that point, anyway. This is unlikely to be one of those pieces with the perfect mix of poignant and inane and the perfect end-loop - read Paul Ford or Aaron Straup Cope instead.

Microsoft Excel is the folk tradition of data visualization and our current strategy of making charts and graphs unique and technical does not have a proper focus on what folk is.

This is a loose idea of what folk is, in the context of these words; it’s guaranteed to miss the academic definition and using folk as the term for these attributes does not imply that it is the first or only object to encapsulate them.

Folk is guitars, now. At earlier points in history, it was the violin, the banjo, dulcimer, lute, and so on. Guitars are popular enough to be boring, limited in their ability to play interlocking melodies like the piano or manipulate the entire life of a played note like the flute. But they’re inexpensive, durable, loud, and a shockingly effective tool of communication.

Please don’t cringe at the use of ‘communication’ like I just did when I typed it. This isn’t to say communication as-in communicating emotions, wooing beloveds, and so on; it’s communication as in copying. Making music is a series of abstractions upon the basic idea of sound: waveform, notation, key, song structure, chord fingerings, vocal technique, and instrument itself are some of the levels.

Reductionism is always possible; sounds are the lowest level, copyable and subject to analysis in themselves; but this isn’t a practical way to look at music if you want to create it. Creating music starts in the opposite direction: playing guitar means that you can copy from more people, take lessons from the guy down the street and imitate YouTube videos in your bedroom.

Guitar is a folk instrument because it’s a popular abstraction that lets people communicate.

When I say copying, I mean this: authorship in a folk tradition is about creative force rather than creative origin. A girl I went to school with is now a popular folk ballad singer known for her incredible commitment to learning tradition on a both academic and instinctive level. When she performs a song, she notes not only the original author, if known, but the author of that particular version, and, if any, the performer that she learned her performance from.

By inverting authorship, tradition sustains it.

Linux is a folk movement. It’s the obvious example, and the paragraph writes itself; communities post help for each other, adapt each other’s work, and eventually create a sense of tradition that goes beyond any material or informational work and is more a way of continuing.

But so is Photoshop. It doesn’t matter that Photoshop technically costs hundreds of dollars or that the thing itself is not as mutable as Linux. There are millions of Photoshop tutorials on the Internet that are all a means of sharing, and useful because of this shared abstraction.

The inverse of this idea isn’t as clear; is there anything like an instrument or a program that is not a folk instrument? It’s not easy to say, but what is clear is a sort of tension.

The first source of tension is medium.

The talks at Visualized ranged from people who have created software that was downloadable during the talk, to an artist who created a certain work in vinyl, laser-cut paper, and frosted plastic. Of those who push pixels, the gap between the Illustrator & Photoshop crowd and the code crowd was significant.

It’s easy to talk about certain specifics of medium; the digital artists lusting after high-resolution printing that makes serifs & tiny charts possible, and the print artists after interactivity that attracts larger audiences (or shunning this desire by calling it pandering, or unproductive).

But there are some other differences.

Physical media makes everything seem far away. After one talk which mentioned the transition from print to digital, I tweeted that ‘The Internet makes everything seem similar’, and Brandon Martin-Anderson summarized that idea much better: ‘The internet makes every idea in the world smell like the room that my computer is in.’. Physical presentation is not only filled with more cues, like your having to drive to a gallery or the price of a wall-art, but it also presents context in a vastly different way; a poster on the wall can cite influences by name, but the viewer can’t ctrl-` though tabs to see what influence really means.

The Internet flips this; like the geographic principle that things closer together are more related to each other, the Internet’s collapse of everything into a flat index means that simultaneous invention is the norm. Everything new will be notified of its predecessor a comment on the post.

Copying is different. The sweeping term covers everything from being taught and copying knowledge from a professor to redesigning an interface and copying what you don’t create.

As more design is dictated by computation - and I see no sign of this trend reversing - the profession will have to grapple with the concept of copying in a way quite different from copying-as-false-authorship.

The clearest example I can think of is tutorials versus libraries. The idea of tutorials is well-known, but for the non-software initiated, a library is a set of functionality, like tools to create graphics or do math, that can be shared by multiple ‘applications’. For instance, browsers and desktop games can and do share libraries for networking and graphics.

Using a library and following a tutorial are ways of building on the work of others; smaller problems, like how to do a fast fourier transform or how to pan-sharpen images are more common and more shared than the larger ideas that tend to be thought of as ‘the final work’.

But the feel different. In programming, you install a library easily and can more or less ignore its internal works, authorship, and particulars while taking advantage of the time saved. This is because, to get a statistics library, I type npm install -g simple-statistics and to use it I type require('simple-statistics') and I’m done; it’s not performative, it’s declarative: I want this, and then it is.

Following a tutorial is different: there’s no script that you can copy & paste; you mime the actions of the person who created the tutorial, read their narrative, understand the problem area just a little bit more. You feel helped, but instead of feeling like you’re using someone’s work, you feel like you’ve learned someone’s technique, and I think that’s a real difference.

I’ll circle back someday. Okay, here it is:

A great deal of the visualization community’s energy is expended in criticism and avoidance. We now know that rainbow scales, nonzero axes, pie charts, 3D charts, all-at-once transitions, RGB interpolation, big numbers, choropleths for absolutes, scaled points for relatives, bubble charts, and, of course the spherical mercator projection, are bad, evil terrible things. Our recent heroes like Tufte and Bertin had cautionary lessons about the limitations of understanding and the dangers of complexity. They have been for the most part Reformation-era purifiers instead of Renaissance-era creators.

Anyway, I like Excel. I like it because there are tutorials all over the internet for how to use it. I like it because it enables people who are neither designers nor developers to create things. I like it because it isn’t intimidating: one of Aaron Straup Cope’s recent bits where he talks about how Flickr failed at being a place where everyone felt welcome to create and Instagram succeeded.

The reason that I have such a problem with all the talk about how awful and retrograde people filtering their photos is that if we did anything wrong, at Flickr, it was letting people believe that their photos weren’t good enough to be worth uploading to the site.

It’s important that people create, even if they aren’t good at it. Having created a thing makes you look at the thing differently. If you’ve ever copied a master painting in art class you not only know that painting better than you’ve ever known a painting, but you can then look paintings as makeable things, as an interaction between skill, ability, materials, environment, and so on.

The chorus of visualization experts is that this should save us from boring charts; that we have suffered long under mediocre aesthetics and poor design and now it is time to move beyond. I don’t believe this. The relevance design in charts and even the idea that we need custom-designed charts is because bad tools like Microsoft Excel made visualization normal; they turned a specific, cloistered practice into something that we can picture in our heads and feel like we understand.

So popularity, normality, and mediocrity aren’t things to be eliminated; they fill out any real tradition.

And so I guess I should conjure some take-aways.

The critical view of visualization is negative, and should only be the counterpoint to creation, not the center of attention by itself. Like the typographer’s collective witchhunt of Comic Sans, criticism has given us something to hate but nothing to love - and much less to say about what we do love.

Knee-jerk criticism over specific parts of visualization rather than the end-result communication ability is insider ball at best and puritanical at worst. If something perceptually imperfect draws attention in a way that the visual-truth version could never, it seems best to communicate and be heard rather than not.

This month I’m recording an album of some folk and some electronic music. It’s bound to be pretty bad, but the thing that excites me most is that two of my good friends are also going to try writing and recording an album.

I would rather that happen than to just leave it to talented and beautiful people.