This style of teaching (I feel) gives the illusion of confidence that the student is on the right track although its barely more than blind leading the blind.
Possibly an unpopular opinion but I have been tricked by folks like this in the working world and academia...I feel it is the source of much of my frustration with CS as a profession.
> I could not live code in this class. I didn’t know the material well enough. It took me too long to code, and I couldn’t recover from mistakes in front of the class.
Does it bother anyone else that the author was teaching something they weren't proficient in? Is this normal for colleges?
I only remember having one prof who clearly didn’t know the material. He spent most of every class just copying code onto the board by hand, then trying to talk about it a bit. I really didn’t learn much in that class. Unfortunately it was the foundational C++ class at my university and really kind of handicapped me in successive classes that were also taught in C++.
My alma mater? The University of Michigan.
You could, but they all have jobs doing actual programming. Universities pay sub-poverty wages to adjuncts to teach most of their classes and so this is what they get. It sounds like this person at least cares and is working hard to teach the material.
Yes, it concerns me - but it doesn't really surprise me (my own UG CS experiences have really disillusioned me with regard to how much of a priority teaching quality really is, particularly in research-intensive institutions). It doesn't sound like the author had much say in the situation "I also had to teach it", so unsure if the department is the root problem here.
But it's doing a disservice to students - particularly given in the US they'll be paying much, much more than the tuition fees here in the UK (which I think are excessive as it is). It's not good enough. If you're going to offer a class that's more practical, covering real web technologies, do it properly or don't offer it at all. We risk teaching poor practices otherwise and I've seen these lead to real security issues when students enter industry and want to put these skills to use.
And my belief, as far as I gathered from hearsay, was that ‘Western’ approach was different, what with having professors who actually work in the field between lectures. So I was, uh, rather surprised to read this.
What does bother me is the lack of education regarding the HTML elements. How often have I seen anyone with a MDN page reference for an element? That never happens! Yet it is the dictionary and there aren't a lot of elements to remember unless you are trying to draw SVG shapes with masks, arrow heads on them and fancy fills.
Is there a good place to get a good idea of what is being taught to students these days? I am sure that the basic HTML elements are being skipped over and I would not mind confirming that.
I found this approach is very effective for students because they basically learn the skills from their peers or even roommates, and if they are friends on campus, they can ask questions any time outside of TA hours. My alma mater has an excellent undergrad TA program where students who performed well in the course are recruited to be TAs when the course is offered later again.
Universities have a short list of options. Having instructors who are smart and understand the material but may not be masters of it? That's a really good situation to be in, compared to the worst case scenarios. Sure, maybe Mark doesn't feel comfortable live coding in front of a classroom for web dev (a skill, by the way, that changes dramatically every year). But he's someone who cares about teaching, plays around with his teaching strategy, and is smart (I mean, he's freaking Mark Guzdial). With those three facts, his students are in a significantly better situation than the average university class.
I had to teach "Intro to Algorithms" the past two semesters. I got an A in that class when I took it, and I'm able to solve most problems from the textbook. Does that mean I can stand up in front of the class and solve any question from the textbook? Absolutely not. Does that make me less qualified than Skienna or Djikstra? Absolutely. Does that mean I can't teach the class? I don't think it does. I can build and grade assignments, I can lecture in class, I can answer most reasonable questions that students have, and I can build good exams that fairly measure my students' knowledge. I'm sure my students would prefer to have an instructor who can do all of those things and also be able to perfectly answer any possible Algorithms question they may have. But unfortunately, that person doesn't exist at University of Delaware - or if they do, then they're teaching one of the other required courses, and we couldn't spare them to teach Intro to Algorithms. It's just the way things are with hiring. I don't think my students' careers are going to suffer because I can't whip out a proof for the reduceability of Vertex Cover to 3-Sat live in front of them.
Ironically, I would feel extremely confident about live coding for a Web Dev course. Hopefully I get a chance to teach one at some point, because I'm pretty sick of teaching Intro to Algorithms :)
Especially not since there typically isn't a good university career path for teachers. Research is generally valued a lot more when it comes to getting a permanent position etc.
In an intro course, you seldom get into the super advanced parts of that topic. He's still a very distinguished CS instructor, and even if a student were to ask some question about, say, a CSS cascade question, he'd be more than capable of looking up the answer and explaining things, even if he doesn't know something on the spot.
I'll add that when live coding, you need to generally be more than just comfortable explaining a topic -- even if you understand say, using jQuery callbacks with AJAX requests, you'd probably avoid teaching it live until you know how to recover from errors in reasonable ways or exactly where in the docs you need to look. It's not that you couldn't figure out the solution, but you want to be comfortable debugging something on the order of seconds of to a minute or two so you don't loose lecture time.
This removes some of the subjectivity from a completely open-ended coding assignment, as well as reducing the amount of effort required, because you can do a few small problems rather than a single huge one that typically involves working over a weekend.
The only downside is it's a lot more effort to set up than either an open assignment or a standard Q&A-based interview. You don't want to throw anyone into real production problems for a variety of reasons. The initial setup has to be understandable within a few minutes of code review.
It's true that I was just a step ahead of the students. However, I was using course materials designed by people who knew the material, I had veteran TA's (who could answer the questions), and I do have over 30 years of teaching experience. This was an advantage when teaching this class. I could figure out when the lecture slides weren't enough for me to understand, so I added more to them to make them understandable for the students. My expertise is in teaching: Figuring out how to make things learnable by students.
In the end, the Web programming part was only a portion of the course -- the rest of it was on design and development process, architecture, and issues like accessibility, all of which I am more familiar with. (For example, because I'm an old Smalltalk programmer, I can explain where MVC came from and strengths/weaknesses better than many.)
The course evaluations were very positive. I think the class worked.
I have created a tool that allows me to move quickly through my code where students can still see how the solution evolves. I can comment at interesting points in the development of the code and attach hand drawn pics and screenshots. Here are some examples:
https://ourcodestories.com/markm208/Playlist/All
(Click on the >> button to move from comment to comment. These work best on a big screen)
I show some of these in class before asking the students to work on a related problem. I require the students to view the rest outside of class.
My students really like these. They prefer them to textbooks (no surprise) and videos covering similar material.
The tool to create these 'playbacks' is free and open source (currently a VS Code plugin). I have also created the site to share these with others. I am looking for feedback.
At https://www.missionbit.com/ I taught a web development class for a while and we'd start with a pre-built but very simple web application. It was a "social wall" that displayed a list of messages and had a textbox that let people create new messages.
It used a real database, was ready to deploy on day 0, but did only that one thing (two endpoints, one database table, etc.).
Over the class students would add new features (liking, sorting, user authentication, etc.).
It works well if the project(s) connect!
The classic model just seems so turned around. Using class time for live teachers to deliver static content seems insane. And then making sure during the time when students have the most questions(at home doing the work) is also the time when they it's the hardest to ask teacher's questions.
In real world: They give me the goal, let me find the theory.
It's the finding the right theory for the implementation a goal is the toughest part here for a graduated.
