create a model.sensePart class and change content model of <sense> accordingly

[laurentromary]

The content models of <entry> and <sense> are doomed to being aligned because they both use model.entryPart.top. This is a real issue when ne wants to customize one or the other elements and more generally, it does not make real 'sense' to have <sense> in the same group as <entry>, <hom> and <re>. A simple correction for this would be to create a model.sensePart class, and ensure a smooth transition by making all current members of model.entryPart.top members of this class, adding <sense> itself, in order to make the element recursive automatically (and thus simplify the current content model). This proposal is not changing anything from a behavioral point of view, but would allow, for instance, the TEI lex 0 group to fine tune the content model of <entry> without impacting on <sense>.

[ttasovac]

Just a vote of support for this idea from my part.

Leaving the world of TEI for a moment, it doesn't make sense — lexicographically speaking — to think of <entry> and <sense> as equals. An entry is a basic unit of dictionary macrostructure, whereas sesnses belong to the dictionary microstructure, i.e. they appear at two very different levels in the dictionary hierarchy. Which is why lumping them together in the same TEI model class is kind of sloppy.

As @laurentromary already pointed out, we ran against the wall in TEI Lex-0 because we couldn't modify the content models of entry and sense separately because they were members of the same class. Keeping model.entryPart.top as is and creating an additional model.sensePart class would be totally non-disruptive from the point of view of TEI, but it would be both lexicographically more sound and practically very helpful for more fine-grained customizations in various projects.

[scstanley7]

Many apologies for the delay in addressing this! Just to clarify: the proposed model.sensePart would contain <entry>, <hom>, <re> and <sense>?

How do you feel about the following language in "remarks" for this new model class?

groups elements belonging to the microstructures of dictionary entries

Thanks for your patience.

[laurentromary]

No, they would not. We suggest it contains: sense cit def dictScrap etym form gramGrp lbl re usg xr (the previous members of model.entryPart.top to which we add sense). The result is pure continuity from the resulting content model.

[martinascholger]

Council agrees at F2F