remodeling spGrp to self-nest for #2695

[ebeshero]

See if you agree with how I modified the content model and description of <spGrp> and the somewhat more generalized new values I've added for @org on att.divLike.

Right now, this branch isn't passing tests because I haven't added the new source for the example to BIBL. But when I build the Guidelines schema (with make exemplars) I'm validating a test file with @lb42 's example on the ticket. I notice that my new values for @org are breaking test1 and test2 in Stylesheets so we'll need to do some repairs—but we should review them anyway.

[ebeshero]

This is passing tests now. @lb42 , how does this look, and do you want to do something different with the BIBL entry for your example? I'm not sure it needs more than this update to the spec if we think the chapter prose in https://tei-c.org/release/doc/tei-p5-doc/en/html/DR.html#DRSPG is strong enough.

[joeytakeda]

Reviewed in Jan 20, 2026 community call (with thanks!) -- group notes that, by and large, this look great and agrees that prose may need to be updated to be a bit more explanatory (which @lb42 has offered to draft).

However, one small quibble is the addition of both unison and parallel as values. Community Call group suggests just using parallel to allow for broader use (as unison is semantically quite similar to parallel, but may not be as widely applicable).

[lb42]

The bibliographic record for the nice new example should probably not use the date 1830 since this is the date of first performance, not the publication date, which is unknown. And it should definitely spell the publisher's name correctly ("Hailes Lacy") . I suggest adding the serial title as well "Lacy's Acting Edition, volume 18, No. 0262" but that's not essential.

[lb42]

As Joey says, it's not clear why both "unison" and "parallel" are provided, or what the distinction between them is. I would stick with just "parallel" as a third value, complementing the existing values of "composite" (no claim is made as to how child elements should be processed), and "uniform" (all children are to be precessed in sequence). Have you a use case for "UNISON"?

[lb42]

There's a lot I would rewrite in the existing DR chapter, but none of it is particularly wrong. Perhaps the section https://tei-c.org/release/doc/tei-p5-doc/en/html/DR.html#DRSIM might be moved to follow https://tei-c.org/release/doc/tei-p5-doc/en/html/DR.html#DRSPG to show how spGrp is useful for things other than musical numbers. Certainly the Shavian example needs to be corrected to use @org rather than @type

<spGrp org='parallel'  rend="braced">
 <sp>
  <speaker>Lady Utterword</speaker>
  <p>Mr Mangan!</p>
 </sp>
 <sp>
  <speaker>Captain Shotover</speaker>
  <p>Whats that?</p>
 </sp>
 <sp>
  <speaker>Hector</speaker>
  <p>Ha! ha! Do. Do.</p>
 </sp>
 <sp>
  <speaker>Ellie</speaker>
  <p>Please dont.</p>
 </sp>
 <stage type="delivery">in consternation</stage>
</spGrp>

[martindholmes]

I agree with @lb42 and the community group that the value "unison" is essentially no different from "parallel", unless the latter is supposed to suggest speaking different text while the former means the same text; but if speakers are speaking the same text in unison, that would be a single speech, surely?

[ebeshero]

@lb42 @joeytakeda @martindholmes I've implemented your recommended revisions for this branch, including:

• removed the "unison" value for @org (I agree, not necessary, just an idea I had while initially working on the ticket.)
• rearranged DR to move #DRSIM after #DRSEG. (That works nicely, yes.)
• fixed the BIBL entry for the Lacy's play edition.
• While in the neighborhood of DR, saw some ancient discussion over references to G. B. Shaw plays. Shaw is an old favorite of mine, so I'm familiar with his punctuation quirks (really he was being as minimalist as possible), and just updated both BIBL entries to Shaw to the proper published Constable editions. I just looked this up, and apparently Shaw called apostrophes "uncouth bacilli" and wanted them to be eliminated from English writing! (He didnt win that–see what I did there?) but his plays are properly published without such "uncouth bacilli".

[sydb]

I do not feel qualified to review the changes to BIB.

[sydb]

Any reason not to have the @xml:id follow the usual pattern? (All of the 1,576 other <egXML> IDs that start with "gi-" end with a disambiguation component comprising a hyphen and two lowercase letters, a–z.)

[ebeshero]

Okay, fine! I'm changing it. :-) Never thought to look at all of them, but was just pairing the egXML id to the id I made for the new BIB entry.

[sydb]

Wouldn’t “… for encodings which focus on the actual performance …” say the same thing in a simpler way?

[ebeshero]

Yep. Revised.

[sydb]

I am concerned about the implications of the inconsistent use of space in this example.

• 18 of 42 end tags are preceded by an extraneous space. I do not think this alone is a problem, but since I am writing a comment, anyway …
• One start tag has an extraneous space before the TAGC delimiter (‘>’).
• Four metrical line elements (<l>) start with 1–3 occurrences of U+00A0. Is this standard encoding practice for indented lines? (I have always used @rend on <l> to indicate the indentation, but <space> exists for this purpose, and is quite reasonable, too. Use of U+00A0 does not strike me as wrong, but it does strike me as fragile, and thus perhaps not something we should be recommending.)

[ebeshero]

Thanks for catching this, @sydb ! I went through with this tidy regex (\S) (</) and located 27 extraneous spaces before the end tags, all of which I've now removed. And I carefully removed the ones at the start tags. (I agree, it's not a good idea to suggest we handle indentation that way, and I'm pretty sure it doesn't matter for this example, but if someone objects we can put a nice @rend in there instead.)
Fixed the weird space in the start tag as well.

[ebeshero]

@martindholmes Did the build for this come out okay in PDF / HTML? I can rename the branch if that will help, and also make sure I can build from Jenkins a little later today.

[martindholmes]

Test build on Jenkins works well; PDF looks good.