For example, consider an essay prompt that asks students to discuss why Twain chose to have Huck and Jim perceive particular events in different ways. There's a standard that addresses character's point of view that you'd definitely want in such a rubric, but the flexibility of the tool allows you to name that rubric element "Huck's Perspective" instead of the more generic "Character's Perspective". That means you could then re-use the **same** standard and call the next one "Jim's Perspective". You can then evaluate the students' ability to articulate both perspectives separately.
Because of this super flexibility there's no way the system could auto-generate adequate rubric performance "descriptors" to insert into the cells of the rubric grid. That's left up to the teacher.
I've realized that rubric descriptors are really for setting expectations before students complete the assignment. However I have to confess that I question the practical value of this. I know this flies in the face of proper rubric methodology, but meticulously defining rubric descriptor text seems more rooted in teachers' inherent nerdosity than actual student benefit.
I worked hard to set expectations in class before, during, and after the assignment by having students discuss and evaluate tons of samples. So I never felt that rubric descriptor text was necessary; they largely knew what a "Proficient" thesis looked like. Besides, how many students are really going to pore over all X-by-Y rubric grid squares of descriptors?
Without descriptors a rubric simply becomes an assessment recording/reporting tool, a way to capture at a high level how a reader assessed a writer's work and communicate that back to the writer. This approach is the core of the EssayTagger system.