I added this edition in this edition group. Both are eBooks published by Standard Ebooks (the one I’ve added is a newer revision). The book is essentially a collection of short stories taking place in a fictional world created by the author. Now, the first edition was already in the database and rather than linking a work for the whole book, each story there is linked to it separately. This is something I can understand, since these stories don’t constitute a single narrative - they just reveal different aspects of Lord Dansany’s world. When I added the revision, I linked each separate story, however I did create a work for the whole book and linked each story using the “has part” relationship. My question is, now that there’s a work for the whole book, should link it to both editions and remove the relationships of each story?
The first question is, of course, whether the decision to create a ‘work’ for the collection is the right one.
This is a difficult decision, as there is no ‘narrative frame’ that connects the stories. The stories are linked ‘by Dunsany’s invented pantheon of deities who dwell in Pegāna’ (wiki). Is this a sufficient condition for a separate ‘work’ in BB? It’s OK for me, but I think some users will see it differently.
Assuming the decision was correct, I would leave it as it is, as we don’t yet have a way to display the parts of a work in the edition form.
I basically agree with @indy133 . But just to clarify further, the idea of “narrative frame” is for things like the One Thousand and One Nights, where in addition to the stories themselves you have an interlocking narrative, e.g. “Scheherazade, the vizier’s daughter, offers herself as the next bride. On the night of their marriage she tells a tale [enter tale here] but does not finish it, promising to continue on the next day. On the next day…” and so on and so forth. The tales are complete stories and can be published separately, so require their own works (when they are), but if we only add the internal stories for the edition, we would be missing part of the narrative.
In this book, The Gods of Pegāna, there is nothing but the stories, there is no frame narrative (in this sense) to lose, so it doesn’t make sense to add an additional work.
I’m not familiar with the book in question, but I think if this particular collection is commonly published as one, it might make sense to create a work for the particular grouping in question, whether or not there’s an added narrative like the 1,001 Nights
however, if it’s a one-off collection, I agree it wouldn’t make sense to create such a collection work
Unfortunately, I haven’t read One Thousand and One Nights to compare but, as far as I know, the stories in this book are always published together (although I could see someone only including a few of them in an anthology book). And after reading it, they don’t feel “standalone” enough to separate them, because the whole book feels like a mystical religious text in a way. While there isn’t one unified narrative presented from beginning to end, there is some connection between the stories - especially towards the end where the book talks about some prophets back to back and actually the last story talks about how the end of the world would happen. But even in the beginning, there’s the creation of the world and the gods, which is necessary to know in order to understand the rest of the stories in the book.
An interesting topic that has popped up before (and not just on BookBrainz: Short story cycle - Wikipedia).
Where is the line between a collection of separate short stories and a self-contained work that is presented as a cycle of short stories?
In some cases it is fairly clear, like fix-up novels (e.g. Ray Bradbury’s Martian Chronicles and The Illustrated Man), but in other cases like the one you brought up, the decision is harder to make.
My personal opinion is that if a user deems that the short stories are linked by a common theme, and are not commonly published separately, it makes sense to have a containing work entity (with “contains” relationships to the other short stories works).
That would represent the general understanding of this collection as a work.
After all, we treat some collections of poems in the same way (Like Baudelaire’s Les Fleurs du Mal: Les Fleurs du mal (Work) – BookBrainz )
However that leaves the decision open to interpretation, which some users do not like.
If we were to try to formalize the decision process, would the criteria suggested above (united by a common theme, not usually published separately) be sufficient?
Does anyone have other/better criteria to suggest?
If we were to try to formalize the decision process, would the criteria suggested above (united by a common theme, not usually published separately) be sufficient?
In my opinion: no. As you say, themes or probability of separate publishing are a matter of opinion. There is no need to fill BB with duplicates because different editors have different feelings. If the works have a connection, a sequence, you can just add them to a series.
It would be like someone on MB saying “this album is never published separately, and it’s too much work to add the data for each separate recording, so I’ll just add it as one recording”, and other people replying “Sure, if you feel it has a common theme, why not?” Because the whole point of the project is not doing that.
The way I see it, the main benefit of BB is precisely that we can have works instead of books. Of course that is much harder to add it work by work, rather than just copying the book details from some other catalogue, but that’s the whole point. There are millions of book catalogues, there is only one BB. If we lose that, there’s no point wasting our time here.
Actually what you say about recordings in mb isn’t quite accurate. Instead it’s more “these separate recordings (work) are part of a bigger whole so let’s add an overarching work”
which we actually do in mb! (eg for classical, but also things like musicals)
Myself I think that a Work should express an actual existing concept, that is, something that has a narrative purpose (eg 1000 & 1 Night),
an overarching concept (eg short story circles)
or even just a re-framing concept later added (some anthologies and fix ups do this)
The difficulty is that in real life the concept of “what is a work, what is a series, what is a later added framing device and what is just a collection” is really really blurry.
For example, is a newspaper a work? Is a dictionary or a reference book containing large amounts of referable/citable credited sub-works (articles) a work in its own right?
When/when not? When does a framing device become a work?
I am currently working on the “work” section of the docs and in-fact am a bit stuck on this exact part, eg, the “framing works”.
For transparency:
Myself I had in the past (for simplicities sake) added works that could with grace be considered “framing works”. But I’m still not sure about that. The data in the database should be real data and not added for the database’s structure’s ease of use either. But sometimes a framing work can be both?
Sorry, it seems I missed your reply. I didn’t mean that works in BB are the same as recordings in MB — it obviously depends on the kind of release/genre, we can never compare written works and recordings 1:1. My point is that the common argument of “it’s a lot of work doing it like this, so let’s do it in an easier way” to me is not valid. And it would be similar to saying we shouldn’t add recordings in a release of popular music (where songs are generally clearly defined) because it’s too much work. It can be a lot of work, but the whole point is to do the work and have good, accurate data.
The difficulty is that in real life the concept of “what is a work, what is a series, what is a later added framing device and what is just a collection” is really really blurry.
I don’t think it is really blurry. I think there will be specific edge cases that are hard to define and that we can discuss here. But for the vast majority of cases, it’s very clear.
For example, is a newspaper a work? Is a dictionary or a reference book containing large amounts of referable/citable credited sub-works (articles) a work in its own right?
A newspaper, clearly no. It’s an edition (by the terms we use here) including many different texts (works) of different types (news articles, essays, opinion pieces, reviews — traditionally but not commonly any more also short stories and poems, etc).
A dictionary and encyclopedia-like work are different things. An encyclopaedia includes a series of essays (works) on different subjects, by different authors; if signed, they should be attributed to the author. A dictionary or glossary, etc, is mostly a list of words, it does not consist of separate texts with specific authors that could be entered separately — generally not even full sentences. But it may include other works in addition to the dictionary itself that are separate works, that can have specific authors and should be entered separately (such as prefaces and essays, texts about the dictionary, languages or grammar, etc.)
When/when not? When does a framing device become a work?
I am currently working on the “work” section of the docs and in-fact am a bit stuck on this exact part, eg, the “framing works”.
My feeling is you’re interpreting “framing work” differently from me and from what we discussed when we worked on this. Correct me if I’m wrong, but you seem to see “framing work” as an imaginary frame. Such as existing within the same narrative, or within the same fantasy world. That is not what I mean. We previously defined work for BB as a written work, a piece of text. When I say “framing work” I mean there is additional text that connects these stories, and that isn’t a work itself because it only introduces the stories, it can’t be read by itself (unlike the stories it frames). I’m not talking about imaginary works, but actual text.
Actually I think we basically agree here, I pretty much agree with everything you’ve said
The only issue is formalising these things into (hopefully) unambiguous guidelines that people will not misinterpret or argue (too) much about! 
I’m actually very glad to hear that. I thought we were seeing this very differently, but if we’re not, that’s much better.