Solid

The larger issue here is the fact that "MDF Unicode" and "MDF Alternate Unicode" aren't 100% correct with regard to the official MDF standard, as far as I can tell. I think we need to work together with one of the "owners" of the MDF standard in order to rectify this.

See also:
http://projects.palaso.org/issues/show/170
http://projects.palaso.org/issues/show/171

Maybe we should document all suggested template changes here (in issue 427)?

On second thought, I've created a more-generically-named issue for tracking problems with the templates:
http://projects.palaso.org/issues/show/432
("Templates need to match the official standards more precisely")

Personally, I don't think this should be changed.

It doesn't match real language to have more than one POS per sense. My interpretation of the MDF alternate hierarchy is simply that \ps is required to be specified for each sense, and if it's not there, nothing should be inferred, whereas in the standard hierarchy, a missing \ps implies this sense shares the same POS as the previous sense.

I think finding data that actually uses more than one \ps per sense is pretty unlikely, and it probably would be data that needed reviewing anyway. I think it's better to leave the default as only one per sense, and let the user adjust things if they really have it the other way.

When I work on a dictionary, I try to have as many things under either \lx(and \se) or \sn as possible. If \ps is below \sn, in my structures I just make it a sibling of \de and \ge etc.