Any unwanted snippet/excerpt/context-of-match lines processed seems like a waste of resources of various kinds. For indexed searches, the only significant //server processing// is the snippet, so offering "no snippets" is a win-win.
Times when snippets are unwanted (for other types of resource savings) include:
- "researchers" just want a count of certain pages, e.g. like a template developer using hastemplate:. This is like how although a personal medical doctor, a social worker, or a police officer may be interested in the "content" of your life, a psychologist or sociologist or politician can't afford to care about any one person, since they are only focused on "the way things are trending".
- "offline processors" just want a list of titles, (AWB + database, text processing of search results page text, etc.)
- when more results per page are wanted
- when //only relevant info// is wanted, i.e. when clean results are wanted
- to hide errors when matching is incorrect
- to make consistent look and feel to the report when some regexp terms show snippets and some don't
- for slow or low bandwidth terminals
- for printouts of search results to use as checklists
Later maybe evolve "excerpt lines" from zero to some number. How about //titlesonly:// or #0 as the query term to signal "no snippets"?
Removing or not removing document size and document date info from the search results page doesn't seem nearly as relevant.