[tei-council] listZZZ things
Lou Burnard
lou.burnard at oucs.ox.ac.uk
Thu Aug 9 09:44:07 EDT 2007
When I rename <witList> to <listWit> (which I probably should), P5 will
have six listXXX things, which together with <list> itself constitute
the model.listLike class
Would it be reasonable to expect them all to have similar content
models? Currently, they don't. Here's what I just noticed (alerted by
a useful sourceforge ticket from the indefatigable T Schassen)
listBibl: (divTop|global)*, biblLike+, global*, (divBottom, global*)*
listNym: (nym|listNym)+
listOrg: (pLike+ | org+)
listPerson: (plike+ | (personLike+, particLinks?))
listPlace: (placeLike+, relation*)
listWit: (headLike?, (witness|listWit)+)
I note the following anomalies:
a) listWit and listNym can recurse, which seems like a generally useful
idea; listPlace can too, but only because listPlace and place are both
members of model.placeLike. None of the others can.
b) listOrg and listPerson can have just paragraph content but none of
the others can.
c) listBibl permits all sorts of tosh at beginning and end, but none of
the others does.
d) listPerson and listPlace permit encoding of relationships between
their children by means of a relation elements, albeit in different
ways, but none of the others does.
What's to do?
Here's a straw person proposal for bringing these content models to heel:
listBibl: (headLike*, (biblLike|listBibl)+)
listNym: (headLike*, (nym|listNym)+)
listOrg: (headLike* (org|listOrg)+, relation*)
listPerson: (headLike* (personLike|listPerson)+, relation*)
listPlace: (headLike*, (placeLike|listPlace)+, relation*)
listWit: (headLike*, (witness|listWit)+)
This would also mean changing places where listPerson and listOrg are
referenced to permit pLike as an alternative; possibly moving some
things currently permitted within listBibl out of it; moving listPlace
out of the model.placeLike class.
Or we could just let sleeping dogs lie.
More information about the tei-council
mailing list