> "unquestionably belong in an encyclopedia"

This is the problem. Wikipedia takes its lead from encyclopedias, but it's not an encyclopedia, because it's fundamentally different. It doesn't have the same space constraints, it isn't written in the same way, it isn't referenced in the same way, etc.

"Does this article feel like it belongs in an 18th century concept of an encyclopedia" kind of test for inclusion or deletion in Wikipedia just seems wrong-headed to me.

Wikipedia set out to be an encyclopaedia, not a collection of random facts. Complaining that wikipedia isn't a collection of random facts is like complaining that vi isn't Nethack. There's no reason why it couldn't be, but it isn't, and the people maintaining it don't want it to be, so that's that.

I've never understood why those who do want wikipedia to be wiki-random-non-notable-factia don't just start wiki-random-non-notable-factia. I predict it will be fairly unmaintainable, but I'm totally keen to be proven wrong.

The textual content and most of the image content is all licensed under licenses that allow you to download a copy and put it on your own server. I have a copy of English, Spanish, and Portuguese Wikipedias (text only!) in a pendrive on my desk. There's a conveniently downloadable textual tarball. Furthermore, you can even follow Recent Changes to keep things updated. And there are legit companies like Answers.com that actually do this.

I suggest that you start an "expanded Wikipedia" project with more inclusionist criteria. I'll be happy to contribute, although I'm pretty happy with the English Wikipedia's notability policy. I'm sure thousands of other people will be too, especially if the result is any good.

Jimmy Wales already did something like this, by the way.

It's not fundamentally different. Wikipedia is explicitly like traditional encyclopedias in that it's not a primary source, and that it only wants articles about things which can be reliable cited in multiple primary sources.

Articles about random Internet personalities or some garage band are thus inherently not in the scope of the project, until they become more notable.

There's no shortage of source material on lots of topics deemed "non-notable", especially when the people deeming those topics "non-notable" are people who know next to nothing about the subject itself, or are too bound by process and policy to think about doing the right thing for the project.