That’s a most distressing position. I can think of many cases where this solution will function just as well, not better, than creating a Work for the short forms collection. And i can think of many cases where this solution will function significantly worse than going with a Work which includes a fixed set of other Works. For many well-known authors, the same, titled*, set of shorter works travels from book to book unchanged, treated as a whole, solo or grouped with other works by any authors in all kinds of combinations, across lots of different publishers, for tens or hundreds of years. And if it’s a translated author then most likely the same canonic set of translations, possibly by various translators, will also be reused again and again. But for unknown authors, quite possibly, the whole set of shorter works will be published once and that’s it. In all cases some knowledge and new research may be required from the BB editor to decide whether the collection’s Work should be supplemented by the full list of separate Works of all shorter titles included, but there’s no harm if the effort of creating separate sub-Works is delayed until someone is certain it needs to be done. But when it’s excessive, then the BB editor shouldn’t be forced to waste effort and lose in visual clarity of the result. Yes, the simple text list of contents may be superior in clarity against the list of separate database items, even on advanced stages of the site develompent. And the key problem? I feel and believe that the hierarchical set of Works, together with its set of relations to Contributors and Editions, is far more intuitive for a website user, because Work and its attributes are natural, with clear parallels in physical reality, and the Edition Group is an artificial entity that is harder to embrace. Its one purpose i can easily understand is when the SAME publisher around the same point in time releases the exact content on a variety of media. Not, for example, if a different publisher reprints a classic book with the same textual content but (and they will do that) having replaced the illustrations, it’s not an ‘edition that belongs to the same group’ intuitively then.
And whenever a Work is structured into smaller forms, it should be provided a simple clear text list of its contents as part of its descriprion. If contents are altered in physical Editions while keeping the same title, that creates a version of the Work (just like with versions of a one-piece novel), with differences that may or may not be ‘described’ by linking the different database items corresponding to included smaller Works, but the textual list of contents is necessary either way, for convenience of a human reader. That’s what going to need to eventually be done to all kinds of ‘Bible’ works for example, to explain how they differ. Just looking at list of sub-Works in a database will tell a website user nothing.
*‘Titled’ is probably important: even if the same set of short stories (like for an author who only wrote a few short stories) are almost always published as a set, still the BB editor shouldn’t be inventing an artificial name for it, and they’re forced to create separate db entries. But if an entity already has a name, that’s what allows our mind to operate it as a single item.
And a logical note just in case: if B and C are parts of A and there are no other parts of A in the database, that’s still okay for the A to be bigger than just B+C, as long as this stance suffices to represent all relations that are not directly inherited from A by B or C, but are totally inherited by everything that is ‘A–(B+C)’.