This is a limitation for physical encyclopedias because paper, and even more so shelf space, are expensive so some facts have to be prioritized over others. And you can't easily make corrections to a book after it is printed and sold, and since research never stops a paper encyclopedia will frequently be out of date, thus it can never be that accurate - so why expend a ton effort to make it more accurate in the first place?
Electronic encyclopedias, like Wikipedia, don't have the same inherent limitations. Wikipedia just happens to have re-created them for the same reason electronic calendars by default show you one month at a time, even on the 30th and 31st, rather than a chunk of next and last months.
The reason so many people are so motivated to complain about Wikipedia and all it's limitation is because everyone intuitively can sense there ought to be a better way.
1. Notability of facts is a red herring. Text is easy to compress and storage space is cheap.
2. The need to summarize topics is a red herring. Having a summary "front page", or "top", or "above the fold" etc, combined with more in depth, detailed sections, is a standard way of organizing information on the web. An electronic encyclopedia really ought to use that kind of presentation strategy.
3. Primary sources, new research and minority opinions would naturally be part of the lengthy, detailed version of a topic.
And if you want to go completely crazy you could do things like allow voting. Allow people to sort what the default view is by general popularity.
Even crazier, sort by popularity based on experts opinion, work out a way for experts to electronically sign or approve articles, and allow people to choose a set of "experts".
Go totally bonkers and verify the experts so that someone could choose to see the top evolution articles as rated by Richard Dawkins.
Or alternatively re-crate all the limitations of a physical set of encyclopedias.
