Another revision of gimste done by me is complete.
I based it on excellent simple definitions of gismu made by la .aionys. (.io) but changed them here and there.
The most important difference is that I specified possible sumti types for each place.
E.g.
(event) - an event place. {nu} or any other place like nicte1 that has {nu} inside is possible for this place.
(assertion) - du'u
(property) - ka with ce'u referring to x1 of the current gismu
(property of x2) - ka with ce'u referring to x2 of the current gismu
One of the rare sumti types is e.g. (ordered set) that requires using {ce'o}.
Places with just x1, x2... without a sumti type specified imply "object" sumti type (e.g. mlatu1)
This revision primarily deals not with rearranging the place structure but instead with ignoring final places if they are not necessary.
Thus it is 99% compatible with the official gimste.
It is a part of a project of making tutorials for Simple Lojban.
The gismu that changed their place structure are:
zabna, mabla, traji, satci, konju
A few gismu had weird places in non-final places. Unlike other proposals I am not suggesting to change the place structure.
I just removed those gismu from the list and replaced them with new fu'ivla created by me.
There are three of them: {anli} intead of {djine}, {lianzi} instead of {linsi}, {kombitu} instead of {komcu}.
I also included some important cimjvo and fu'ivla into the list so that it can reflect the basic ontology more precisely (although, ofc. no such ontology would be perfect).
The current version can be seen here:
https://docs.google.com/spreadsheet/ccc?key=0Ahngu1CNj7wddFltdkVpZFNLWlQ5dVN1VVpXSnphVVE&usp=drive_web#gid=7
Feedback is welcome.