Hi,
This thread is interesting.
It's clear that authors utilize ordered list to achieve the desired formatting when step-like things are needed within choice and choicetable. Do we contend that having step-like things inside of choice and choicetable are not a valid use-cases? I believe that they are valid use-cases, especially when you consider that Information Mapping advocates this construct in their if-then tables.
If we agree that step-like things within choice and choicetable are valid use-cases, then it would be better to offer the semantically correct element for the use cases in choice and chdesc. I deliberately omitted choption from having substeps because I contend the choption content is to help the reader select among choices.
Ordered list would have to remain in the content model for backward compatibility. But, a newly adopting organization could remove it with a constraint.
I suspect that the current models were arrived at only after lengthy discussion. Should we decide to pursue this, it would be useful to have the rationale for the decisions.