Whenever I mention that I've been thinking about writing some kind of free and/or open source dataviz manifesto I'm met with skepticism. In one recent conversation in a data visualization focused Slack, I mentioned that the New York Times' much-lauded data graphics probably cannot be be archived effectively. This immediately turned into a conversation about whether Observable, which I gather is some kind of interface for D3.js, is genuinely open source or not.
Maybe my mistake was that I had used the phrase "open source". I should have asked whether someone decades from now will be able to verify, reproduce, edit, or share a NYT data graphic published today. I think it's very plausible they won't even be able to read it in the first place. The NYT's data graphics rely on a pretty extensive stack of technologies, some at least partly proprietary, that might be tough to get working again once they're obsolete. If I came across a bar chart in an archived newspaper issue from decades ago in the disused microfiche room of some small town public library today, I could certainly read it, and I could probably also share it without too much trouble. And that will probably still be true decades from now, assuming there are still people who know how to repair microfiche readers. Or, more likely, if digital scans of old newspapers are still available.
I hope it will be easy to use something made with Observable decades from now after all. And I honestly don't even care very much about software per se, let alone all of the fine distinctions involved in free and/or open source software.
I just think we should make data graphics today that are verifiable, reproducible, editable, and shareable. Right now, the data graphics which are arguably the very best being published for mass consumption check zero of those boxes. Or at least, I think it's zero. I can't be certain because they're all paywalled.