The article outlines five qualities that make models useful in the classroom, and concedes that not all models are effective at engaging students in the scientific modeling process. Especially important was the idea that models should represent processes rather than "things," since "things" are easily found in textbooks and do not provide much room for further exploration and model manipulation. This ties into the fact that models should be revisable, the fifth quality that the article talks about. Another key takeaway from the article was the emphasis on the both the seen and the unseen, especially when looking at "invisible" things such as waves or atoms, and how guiding the students to look at unobservable factors can open new ideas.
Something I found to be important during this course that the article emphasized was the use of time and showing time passing in models. If you eliminate the time factor from a model, you often end up just looking at a "thing" rather than a process. Putting extra consideration into time and its effects on the individual processes of a model will allow students to really understand the causes of each event and its effects on other events as time passes.