The Bluebook for legal citation has a similar example for citations to the Bible.

The format of the cite is independent, not only of the physical edition, but also of the citation style used? Examples of what doesn't work correctly would help us figure out what the problem is an how to solve it.

Perhaps I misunderstood. Is this issue about citation formatting, or about the behavior of openURL resolvers?

Hmm. There may be two formatting issues here. The item needs a particular citation format in the text or footnote, but should it also be suppressed in the bibliography? If that's so, an item type for material of this kind may be needed.

Concerning the format of the citation in the text, my point was that some citation style will govern the formatting of other, non-classical references in the document. Is the form "Aeschylus. Supplices 40-57" meant to be used regardless of whether the chosen citation style for other references is, say, Chicago Note or MLA?

The simplest thing would be to decouple the citation from the bibliography entry. So you would register the English translation independently, as an uncited reference for inclusion in the bibliography, and we would need a special type for "classical" or "standard reference" items, that inserts a simplified reference in-text, and does not appear in the bibliography.

Does that sound like a workable solution?

No, not yet. It would be a feature of the CSL processor; the plugin doesn't need to know anything about the special behavior. Some provision needs to be made to suppress inclusion in the bibliography for particular items. At the moment, I'm thinking that it would be enough to add a "classical" item type to CSL, and to suppress all bibliography output for such items. I'll float the proposal with the CSL group, and we'll see what emerges.

waboo - what I'd suggest you do for now is to just write the in-text citation per hand (i.e. without any Zotero involvement) and then add them to the bibliography using the "edit bibliography" function.

ok, then try this: insert the reference, but use the "show editor" field to modify it to the form you want - I believe that should work - it won't update anymore, but that doesn't seem like a problem.

This issue won't make it into CSL 1.0 and the first release of the new processor (early next year). But when the new processor comes out, a less painful workaround may be available, if selective bibliography output is possible. We'll have to wait and see how things develop.

PS: If you're interested in the nitty-gritty behind the scenes, the documentation on the processor functionality for selective production of bibliographies is here. Zotero would need to provide a mechanism for passing through category tags to the processor's bibliography function. With that in place, it would just be a matter of working out how to produce the desired in-text form in the in-text or in-footnote citations in the style.

Zotero 2.1 will account for non-Zotero footnotes when evaluating the position of a cite, both when the citations are originally created, and when any citation in the document is subsequently edited. There shouldn't be a need for a manual override.

With "TB Sotah 8a" entered manually, no, I don't see a way to avoid the ibid. and continue to have ibid in general (there is obviously the no ibid version of the style).

With regular Zotero what you're doing is exactly right (in the sense of being the best available solution), yes.
With MLZ and the abbreviation filter plugin you can configure an abbreviation (like TB or TB Sotah) for a classical source and it will be excluded from the bibliography automatically so this may well be worthwhile for you to explore. Personally my recommendation would be to test MLZ in a separate Firefox profile rather than update your regular Zotero copy, since going back from MLZ is a bit cumbersome (though not impossible).
https://support.mozilla.org/en-US/kb/profile-manager-create-and-remove-firefox-profiles

The fact that there is no Standalone version is probably the biggest one.
Also, while Frank has been very good about supporting it and by all accounts it's been very stable, you do have less support infrastructure and it has fewer users, so you do get less testing.

In general, the MLZ fork is meant to be temporary, so most of the features are intended for eventual inclusion into Zotero, yes. Zotero devs may decide against some or change how they are implemented and it's likely going to take a good amount of time, but the general answer is yes.
The biggest reason they're not currently implemented is that doing this right would require database model changes and those in turn affect syncing so implementing these changes is rather complex. There are changes under way that should make this easier in the future. A first update of fields and item types is planned for 4.2

The abbreviation filter does, I believe, already work with regular Zotero, but the specific functionality you're after wouldn't work because Zotero doesn't have classical sources.

That's basically correct, yes. One can force workarounds, but none of them are particularly satisfying.