Hal,
I'm all in favor of it. For each convention that is invented for diagramming, the question to be asked is how does it instruct the viewer and clarify the structure? There are hundreds of selma'o, which is too many to come up with a unique diagramming method for each one. I would recommend that we come up with diagramming methods for these more general categories of cmavo instead:
- Abstracts - Attitudinals - Connectives - Descriptors - Discursives - Keyboard (alphanumeric, {ro}, verbal punctuation, shifts, etc.) - Math expressions - Modals - Prosumti/bridi - Qualifiers (NAI, LAhE, NA, NAhE, LUhU) - Relativizers (clause/phrase) - Sentence Structure (TUhE, FA, TUhU, NUhI, FAhO, CEhE, FUhE, ZEI, ad infinitum) - Tenses - Vocatives
Perhaps even some of these can share a diagramming method, such as vocatives and attitudinals.