UX for Textual Search in Archive
Should I program the archive-publishing subsite to present search results more or less the same way NodeBB does?
The user-experience-related aspect of the design for any website that supports a search through more than a trivial amount of data faces the problem that the search results can be any size but memory available for browsers is finite. It wouldn't be good to risk overfilling the primary memory on the user's computer with the search results. NodeBB solves this, as many sites do, with pagination of the results. You get a page of results and if you want more, you signal for the next page (or you can even skip forward).
A page of search results gives a title and/or some kind of brief summary for each result. In NodeBB, if you click on one of the results, you are taken to a page with a full rendering of the result, even including the tree of responses to it. At the top of the page is a breadcrumb trail that lets you navigate upward through the tree in which the result finds itself. Nothing about the breadcrumb trail mentions the search results or the search-pattern input or provides a way to get back there. Of course, you can do so with your browser back-arrow.
Would an approach more or less resembling that of NodeBB suffice for the archive? Or does anyone have any radically different ideas for it?
An even bigger question about the search is what forms of search to allow, and how they are to be expressed. There could be exact text match, conjunctions, alternatives, deeper Boolean combinations, and so on.