Allow nested <facsimile>

[MatijaOgrin]

After some debate on the TEI-L (https://lists.psu.edu/cgi-bin/wa?A2=TEI-L;f4bdaac4.2406&S=), I would like to raise the topic of nested <facsimile>.

The starting premise is that a <facsimile> element contains a set of images that represent a material artefact. In the description of a composite manuscript, we introduce <msPart> elements which represent a single material part of that primary source as an artefact. If this is the case, then it would be reasonable that <facsimile> will have nested facsimile parts corresponding to those material parts of the manuscript:

<facsimle>
<facsimle> ... </facsimle>
<facsimle> ... </facsimle>
</facsimle>

Of course, this is not valid TEI; <facsimile> elements are not supposed to nest. However, nesting indeed represents the relation "I am a part of" as Magdalena pointed out. Nesting is a primary structural device in XML. Hence, <msPart> elements are nested as children of <msDesc>, they do not follow as its siblings.

In the last mail to the discussion on the TEI-L, I tried to explain why <surfaceGrp> is not a good solution to the problem.

I propose to reconsider allowing the <facsimile> as a possibly nested element.
Related project: Register of Older Slovenian Manuscripts.

[tuurma]

Thanks @MatijaOgrin for opening the issue, just to give a clear example, MS 013 is a collection of 6 sermons, each originally a separate manuscript, written on different occassions between 1787 and 1813. Since each of these could easily be an object of separate study and standalone encoding, therefore "worthy" of its own facsimile element, I do believe it would be most appropriate to allow facsimiles to nest, to represent clearly relations between parts of a manuscript.

Furthermore, I don't see any danger in doing so, since facsimile can only occur in a very specific context, namely as a child of TEI or teiCorpus, therefore there's zero potential for weird side effects.

[martinascholger]

Council decided at F2F Buenos Aires to allow nested <facsimile> elements

[martinascholger]

I repeat the discussion with @martindholmes, @sydb, and @tuurma from PR #2654 here.

Instead of allowing facsimiles to be interleaved with surfaces etc., the content model should be an alternate between the existing content model and one or more facsimiles.

This is the proposed content model:
facsimile = ( front?, (model.graphicLike | surfaceGrp | surface )+, back?) | facsimile+