Journalism is the first rough draft of history, as the shopworn phrase goes, and it’s a clever one, but it’s never seemed quite right to me. Daily journalism is the reportage of events which may or may not be deemed worthy of reflection and remembrance; it’s in the subequent commentaries and essays—and even supposedly neutral online encyclopedias—where “History” begins to come together.

So I’m with John Overholt, a curator at Harvard’s Houghton Library, who launched a “concept Tumblr”^[1]I’m coining that, by the way as a personal project, earlier this fall, devoted to the first version of Wikipedia entries: First Drafts of History. The idea is dead simple and all but infinitely replicable: for every subject Wikipedia covers, there was once a first version of this entry—and it’s just three clicks away from any Wikipedia article, so long as you know which three^[2]“View history” > “Oldest” > First time-stamped entry.

That’s where Overholt began, as he told me last month: “I was suddenly struck by how interesting and unusual it is that Wikipedia’s entire (or mostly so) history is easily available and that you can peel back the layers of each article to its genesis. As someone with a keen interest in history, that’s very appealing to me, and I was curious to know what the articles were like in those early stages.”

Funny enough, this is close to an idea that I once started to explore, in a post on this very site. Way back in May of 2009 I copied the text over from the first version of the entry about the rock band Radiohead and used it to muse about how Wikipedia’s standards have changed. I announced it as the first in a series, but I never did it again. Ideas are cheap, execution is what matters, and Overholt is executing it like crazy. Every day he posts screen shots with links to the the article and first version every single day, often matching entries to the calendar (Black Friday (shopping) on November 28) or focusing on pop culture goofery (Metal umlaut).

And looking back at the origins of entries reveals something about where Wikipedia came from. The second paragraph of the first Merlot article describes the varietal in three succinct sentences before concluding: “Merlot is also the name of an XML Editor….:-).”

Very, very early early articles, such as the first draft about Venezuela, are just one sentence. Others are written in in shorthand, omitting direct references to subject in a long-abandoned style, i.e. Putin: “Born October 7, 1952… KGB officer from 1975 to 1992…” and so forth.

It also offers glimpses into recent-but-forever-ago history, when Facebook was Thefacebook.com, and the iPhone was just a nickname for an Apple partnership with Motorola (later redirected to Motorola ROKR, at least for a time), then rendered “IPhone” due to limitations of the software. This first concludes: “Note of author : please rewritting my article in a correct english. thank you”

I asked Overholt what his take on all of this was, and I’ll do no better than by quoting him at length:

Obviously it’s funny when articles have a really eccentric start, or a tone that’s very different from the standard style of Wikipedia today, but the thing I’m really struck by is how ambitious and difficult a task it is to think about, in essence, organizing all knowledge. It’s a problem that historians and philosophers have grappled with for centuries. I was tickled by the Pastrami article I posted the other day, which had the edit summary “What can one say about pastrami?” What indeed! But the important thing is thinking to say anything about pastrami at all. The genius of Wikipedia is that it didn’t really stop to solve the overarching problem of how to organize all knowledge first (because it’s all but unsolvable) but rather decided, “Well, we’ll just start with something, and hopefully make that something better little by little.” So even if the first draft of an article is terrible, it’s already done the very hardest thing just by existing.

What else I think is important about it is that it might help to demystify the Wikipedia process, even if just a bit. Many readers have no idea how articles or written, and few probably ever think about what they once looked like, or what the best version may be.

An example I’ve considered with friends: do you prefer the version of the Wikipedia article Dog from 2014 or the article Dog from 2004? I’ll still take today’s entry for a number of reasons, but a decade ago it was arguably more accessible, and about one-quarter the size.

It makes you wonder: what should a Wikipedia article be? What’s the ideal Wikipedia article? The answer to that has changed over time, and probably will keep changing so long as it’s an active project. Reminding readers that Wikipedia once was very different is a good way to remind them that it can still be better.

All images ultimately via Wikipedia.org; first and third courtesy of Overholt.