Wikipedia used to have a warning about articles of a certain size. If I recall correctly, it was 64KB. As far as I understand, the reason for this was more engineering-oriented than user-experience-oriented: Loading a larger page was slower, because networks were slower, or at least so some people thought.
Wikipedia no longer has this warning. It’s not unusual to have a page of 250KB or more. I don’t participate in discussions about performance, but the discussions that I do see are about the time that it takes to parse the templates server-side, to load JavaScript modules, and to render the CSS; they are not so much about the kilobyte size of the pages themselves.
I suspect, however, that there is a problem with page length. Not one of performance engineering, but of user experience. Do people actually read whole encyclopedic articles in Wikipedia? In case you haven’t guessed it already, my hypothesis is that most people don’t.
This is my hypothesis because of the famous debunking of a designer myth: people usually don’t read texts.
It should be clarified right away that the notion that people don’t read whole Wikipedia article is not, by itself, a problem. It may be a bit sad for people who invest hours (or years!) in writing the brilliant prose of each excellent article, but the point of Wikipedia is not supposed to be getting millions of people to read very long articles. Rather, it’s making information that they need accessible, and making it as easy as possible for everybody to edit this information.
Do long articles make finding information easy? Probably not. Experienced Wikipedia editors are familiar with article structure, with tricks like Find in Page, and so on, but a lot of readers are not.
So here’s my call: Let’s bring back article length warning in some form. The importance of a topic doesn’t necessarily justify having a very long article about it. The purpose is not to have a long page, but to make information easy to find. If splitting an article to several pages makes the information easier to find, then the readers will of course be happy, and the editors who invest their effort in writing a lot about a topic should be happy, too, because their writing is more likely to be actually read.
Aharoni in Unicode 