Cardinality of core objects

[DavidFatDavidF]

Do we really want to allow lexicographic resources with zero entries? Similarly, entries with zero senses?

[michmech]

Re: lexicographic resources with zero entries

I think we do want to allow it if we want to allow the existence of lexicographic resources that contain only relations (between entries/senses in other lexicographic resources).

We should probably say that a valid lexicographic resource must contains at least one of: entry, relation.

But, relations are defined in the (optional) Linking Module. So, we should probably swing it like this:

• In the Core, say that a lexicographic resource must contain at least one entry.

• In the Linking Module, relax that and say that, if you’re implementing the Linking Module, a lexicographic resource must contain at least one of: entry, relation.

Re: entries with zero senses

I think we want to allow it in order to allow the existence of entries which serve only as members of relations, such as spelling variants whose only purpose is to redirect human users to other entries.

An example of that is shown in "A.1.16 Modelling variants" in the text. There are two entries there, with headwords "colour" and "color" respectively. The second one is empty (no senses) and labelled as "American spelling". The first one is non-empty (has senses) and labelled as "European spelling". The two entries are linked through a relation object of type "variants".

[DavidFatDavidF]

Decision Dec 12 meeting:

1. top level
   We stick with allowing empty lexicographic resources, we add a note explaining how a resource w/o entries can be useful via linking module or otherwise
2. entry
   entry w/o senses makes sense due to linking or as WIP
   we will also have an explanatory note