I really like the concept of the Textual Analysis. You can extract text from documents and turn them into model elements, but the logic behind it is unclear. Perhaps, I am missing the point of the feature. The documentation has been unhelpful clarifying it.
After extracting candidate items in the document, it lets you set the element type. You can set the modeling element type, but available list of types seem to be an incomplete, random selection of elements from various UML diagramming types. Why can’t I, for instance, extract a term and turn it into a class data member or method, or into a swimlane?
There’s also a column labeled “Candidate Class” which adds further confusion. It lets you assign a different name than the one extracted from the document, which is useful. The column naming, however, leads you to think it applies to class names or it applies to the categorization of the element.