You do not follow my argument. It's not that "experts should want to write encyclopedias", it's that experts do write encyclopedias. This is not a disagreement in our viewpoints, it's an observation that your statement does not fit observed facts.

You believe "experts should do what experts do best, and Wikipedia should do what encyclopedias do best: to wit, experts should conduct research to generate new primary sources or write books to generate new secondary sources, and Wikipedia should continue finding secondary sources to summarize." This also shows that you believe the goals of Wikipedia are aligned with that of encyclopedias in general.

I've listed two dozen titles of encyclopedias written by experts and links to several hundred more. Quite obviously one of the things that experts do is write encyclopedias!

This means that your philosophical view regarding encyclopedias is very different than that used a large number of other encyclopedia projects. Granted, individually each of those projects is smaller than Wikipedia, but collectively they far surpass the amount of content in Wikipedia.

I cannot have the dialog with you about the article we're commenting on until you acknowledge and understand the big difference between the Wikipedia model and that of most other encyclopedia projects.

Once you do that, you'll understand autarch's original observation "that wikipedia attempts to substitute policy for expertise." With that understood, it's easy to consider how an encyclopedia maintained by domain experts would not have the same deference to secondary sources (and avoidance of primary sources and new research) as described by this article.