Permalink
Switch branches/tags
Nothing to show
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
36 lines (26 sloc) 4.23 KB

Best Practices for teaching with binder

Talking points:

  • Materials post-workshop (continued usability)
  • Beginner vs Advanced use cases
  • How much do they need to know about "magic"? - The plan is to get people interested and have them want to learn from there.
  • Four day series - 4th day teach them how to use Anaconda. The issue would be if you were trying to teach someone who didn't have the tools installed, then the issue is what do you do now? How do you move forward?
  • The idea that you could get really far and never leave the binderhub and not do anything locally might not be realistic but some students do it. Is that awful from a learning perspective that they're dependent on this crafted environment? Does it depend on the student - some just want to get their homework done, some will start it up from scratch. Do we want to create a world where you don't have to go through installation first? A lot of the jobs that students want are to be data scientists and they need to know how to start up projects which might be different from a sociologist student. Should we assume that we're teaching it at some point, maybe after we use the curated environment for a while, then have a specific module to create it on your own environment without having to upfront put the frustrating installation first when teaching.
  • We really just care about people learning how to problem solve
  • Some people show how other code has all sorts of problems. Students hate it, they get so angry because they don't want to know. Short talk on how to google and troubleshoot.
  • Having people work with real data and real problems and then find issues makes them motivated to learn how to work through them. People get frustrated when it's a predefined example sometimes.
  • Example - when teaching workshops for teenagers - they need to get a product out fast to feel like they accomplished something. When they leave the workshop, they can't figure out how to set up their own environment and start new projects. Often, the instructions from the workshop are now out of date because software changes so fast. How do we help people learn how to be able to deal with that and troubleshoot?
  • The art of advertising - prereq if they wanted to learn how to go under the hood. Know your audience - do they want to learn programming or do they just want to do their analysis easily? A lot of them don't want to learn how to the package system works. Even data scientists are interested in the data, not managing the packages otherwise they might be backend devs. Where is the point where people shut down? Attention plan for what they're interested in
  • People think that they're going to take a workshop and know everything. And they're not. People don't know what's going in the packages so they don't always know what's influencing the interpretation
  • Themes: generating interest vs sustainable use
  • Do you teach people the programming fundamentals first or do you try to show them something cool first?
  • If we do everything for them how does this compare to just point-click GUI instruction? One way of framing this is as an alternative to that, in which students are exposed to the concept of programming in an open source environment but in which there is really no attempt to teach programming at all.
  • Need for a training Module for teaching setting up your own environment (for non-CS students)?
  • Important to convey the fluidity of the expectations of the software developers; the crafted world is not the real world so don’t expect it to be the same tomorrow, next year, after your degree
  • Approach troubleshooting as “community problem solving” (don’t have students struggle in isolation)
  • Two approaches to teaching with Binder:
  • Binder for ease of instruction IN a specific discipline
  • Going under the hood (how binder works, troubleshooting) for using Binder and tools to work within that discipline
  • Ideal is for students to have a positive outcome that’s self sustainable
  • Undergrads are not used to failure - they expect all demos to work perfectly. Keep this in mind if student evals are important for your performance reviews!
  • Important to convey how decisions (dependencies, packages, etc.) impact the data and interpretation
  • Think of Binder as an interactive texbook!