general style question about definitions

[solpahi]

Should the definitions include sumti types? Should they be phrased "x1 (li)..." or "x1 (number)", "x2 (ka)" or "x2 (property)", etc?
Either way, it should be consistent.
And what all those gismu that used to be used with {nu}, but are now more commonly used with {ka}?

[lagleki]

It was agreed in IRC that "event", "property" etc. but not "nu", "ka".

tersumti interactions is a more important aspect. If not in definitions then either in a separate new field either as full definitions or as their signatures.

[dustinlacewell]

If we are to do this, we should decide now, since this will affect the entire process.

[teleological]

Since abstractions are required in some places by the semantics of particular gismu, and since the BPFK proscribes implicit sumti raising, it seems to me like a good idea to include these requirements in the formal definition of the relevant gismu.

However, I think it's important to distinguish the requirement for abstractions in some positions -- which is grammatically defined -- from non-abstraction type suggestions which are part of some gismu definitions, e.g. "(agent)".