Rubric

Both the topic presentation and the project presentation are graded on the same six criteria. We use specifications grading: each criterion is judged met or not met, with no partial credit on any single criterion.

There is nothing hidden here. Read this page before you start preparing, and you will know exactly what the professor is writing down.

1. Accuracy

Met when your talk contains no incorrect technical claim that would leave the audience believing something false about how Python behaves.

Two things explicitly do not count as errors. A slip you notice and correct on the spot is not an error; catching your own mistake in front of an audience is a skill, not a defect. A simplification appropriate to CS 149 is not an error either — if you describe a list as a row of numbered boxes, you have not said anything wrong, even though the real object is more complicated.

Not met when a claim is wrong in a way the instructor has to correct for the class. The grader writes down the claim, so the record is a specific sentence you said rather than an impression of how well you knew the material.

2. Demonstration

Met when you show at least one concrete instance of your topic and walk the audience through it. Any of these count: code on a slide or the whiteboard that you explain line by line, a program you run live, a hand trace of what a fragment does step by step, or a drawing of what the data looks like in memory.

Not met when the talk describes the topic without ever showing it working. Code that appears on a slide but is never explained does not count. The criterion is the walk-through, not the display.

3. Motivation

Met when you name at least one specific situation in which the topic is useful — a task, a problem, or a bug it solves.

"You would use a dictionary to count how many times each word appears in a file" is specific. "Dictionaries are useful in programming" and "you will need this later in the course" are not.

You can deliver the motivation as part of your demonstration rather than as a separate section. The criterion is that the situation is stated somewhere in the talk, not that it gets its own slide.

4. Delivery

Met when both of these hold: you are audible from the back of the room throughout, and you address the audience rather than reading your slides or a script word for word for the whole talk. Glancing at notes is fine, and so is reading a definition or a line of code aloud.

This criterion judges those two behaviors and nothing else. It is not a judgment of confidence, enthusiasm, polish, accent, filler words, or how nervous you looked. A nervous, halting, audible talk delivered to the room meets this criterion.

5. Time

Met when your talk runs between 8 and 12 minutes, for a 10-minute target.

Timing starts at your first substantive sentence and stops when you finish your presentation. Question time is included and can be throughout or at the end. An instructor who has to stop a talk at 12 minutes marks this criterion not met and moves to questions — you still get your questions, so a long talk fails one criterion rather than two.

Run through your talk out loud with a timer at least once before class. Ten minutes is shorter than it sounds.

6. Questions

Met when you take at least one question and respond to it.

Answer quality is not graded. "I don't know, but I would look that up in the documentation for the random module" is a response, and it meets the criterion. Grading the quality of your answer would mean your score depends on how hard a question you happened to be asked, which is outside your control.

If nobody in the class asks a question, the instructor asks one, so this criterion is always available to be met.

Your grade

• All six met — full credit.
• Five met — most of the credit.
• Four or fewer — partial credit.

For each criterion you did not meet, the grader records one sentence of why. That sentence is your feedback.