This is an interesting request that requires serious consideration.
- Because not all resources are searched many searches, including those built by the system, are of a double nature such as <topic creation> OR creation ... this results an additional results to search through as the "normal" condition.
- With a number of resource providing massive numbers of titles - Text Creation Partnership, Perseus, et. al. - plain textual results return massive but unpredictable results as spelling variants, computer parsing, mixed languages ... expand the number of items matching a character string.
- The net results is large numbers of resources containing matches to skim through to find the resources actually needed for your research.
Being able to sort results in the output is certainly part of the solution. I would suggest that being able to assign languages to transliterated texts, add add labels so that some searches don't need to be compound (think of outlines), be able to provide an "interlinear" with standardized spelling ...
So yes, this needs support but everyone also needs to think what else is needed to truly solve the entire problem.
