taw comments on The Importance of Goodhart's Law - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (112)
I'll agree that "most people [...] don't even bother looking at the data [...]" - I, in particular, am not sufficiently invested in this argument to go to the inconvenience of reading a PDF. The effect of modifiers "this site" and "Soviet Russia" I have no interesting opinion on.
(By the way: horrible format for Internet content. If you can read this, please don't upload your information to the Internet in PDF format. Make an HTML file.)
We have to learn to live with PDFs as virtually all research is formatted as PDFs. Sane (single column portrait-only) PDFs like the linked paper are not particularly worse than constant-width websites. You are exaggerating the inconvenience.
The problem are PDFs which do things that make sense only on paper - like double column / alternating portrait-landscape - these are really really bad for reading on screen. But - what stops PDF readers from having some hacks to make them bearable? I cannot think of any reason. And it would definitely be easier to hack PDF readers than to make all researchers and all research journals in the world switch to HTML.
Related problem of tables being in appendix as opposed to floating seems harder to solve, but it's nowhere near as bad as double columns PDFs.
The biggest three problems with PDFs as a format for Internet content are:
The text display does not adapt to your window.
Viewing the content requires running additional processes, adding CPU and memory usage.
PDF viruses.
You pointed out (1), but (2) is no less annoying to me personally. That said: yeah, I got no control over this.