«

100% Remote and Autonomous Professional Learning and Development

Categories:  learning

I spent the first 10 years of my career teaching (various subjects) and training teachers. This training work included facilitating PLCs (which I’d say most public schools have), leading seminars and workshops, and coaching.

Since becoming a software developer, I’ve had the chance to participate in a lot of professional development sessions. And since taking my current job with Unabridged Software as director of development, I’ve been doing a lot of study into how software developers learn.

Software development is somewhat unique in terms of the amount of continued learning necessary to stay relevant (I spend 15-20 hours per week outside of work on my own learning, and that’s just with things that are immediately relevant to my work). Software development also has some interesting overlap with teaching, but that’s material for another post.

The biggest problems most teams face is the disjunct between “knowing” and “doing” – we know we need to change how we do things, but we haven’t done so yet and may never have the time or chance to do so. Most of my work lately is on closing that gap.

One technique we’ve arrived on is professional learning communities, ported over from my public school days, but with a bit of a twist.

Here is how we do it:

First, our software developers make a list of things they want to learn. Some of this is out of curiosity, and some if it is out of having problems they don’t currently know how to solve (or where the solution is unsatisfactory). Most of the time these lists tilt toward solving immediate problems.

In a group I’m working with on my own, a topic we’ve selected is “Building mobile applications” (for smartphones). None of us have done that before, though we have built other types of applications.

Then we divide into groups of 3-5 around topics. The group does a bit of research and comes up with a guiding question, in the style of Understanding by Design. They also do some conjecturing about what a successful answer might look like.

In our case, the question became “How do we build a mobile application that interacts with a Web API?”, with a successful answer being (unsurprisingly) a mobile application that interacts with a Web API. We already know how to build Web APIs and how to interact with them in other ways, so a mobile application is a relatively reasonable next step.

From there, we make our best guess (based on research) on the steps we might need to take: Our guess is something like this:

  1. Set up our development environments
  2. Build a mobile application that simply displays some static content
  3. Modify the application so that responds to user input, e.g. clicking a button and changing some static content
  4. Modify the application so that it fetches static content from an API
  5. Modify the application so it can also send content to the API
  6. Figure out authentication (how to handle login/logout and user accounts)
  7. Package it into an executable (app) that can be run on both Androids and iPhones.

Assuming we can complete these 7 steps, we’ll have the ability to start creating mobile applications for our clients.

But what does this look like in practice?

There are three of us in this particular group, and we have a Web API for which we’re building an integration. We meet weekly, for about 60-90 minutes (over Zoom, in this case), and we have a shared folder where we keep our code.

We take turns leading meetings. The facilitator for a meeting is responsible for:

  1. Determining with the group if the topic for the week is still appropriate (usually it is, but not always)
  2. Preparing an exercise to help the group meet the goal for the week
  3. Identifying pre-reading and other resources, if any
  4. Leading the group through the exercise (and presenting necessary content and context)

Thus, we write code together. We look at each other’s code (using screen-sharing). We try variations and experiments with our code. We ask questions and see if we can figure out the answers. We compare and contrast with other things we know.

After we’re done discussing, two of us write up our notes (paying particular attention to the metaphors we use and the connections we draw). The notes aren’t usually useful later (that’s what the pre-reading is for), but the act of writing them seems to make the learning last longer (we add meaning to facts, as Dewey might say).

Then the next week, we do it again.

At the end, we pause and reflect, then decide where to go next.

We typically spend 4-6 weeks on any given topic, and take at least a few weeks between (we still might meet, but with a much less formal agenda).

Sometimes the big topic is handed down from on-high or dictated by business needs. Usually it’s determined by developers simply knowing what problems they’re facing and imagining what they might do to solve those problems.

Going through this process does help to ease some of the frustration “I know there’s a better way, but we can’t do it yet” (and provides a series of interesting case studies to show to the bosses when trying to justify new practices). As a bonus, we often get to know each other a little bit better. And we get to feel like we are doing slightly better work.

We can also go through this process 100% remotely (and have been doing so for months).

Actually running a PLC Group presents a few more wrinkles than presented here. If you like, you can read the full PLC playbook I wrote for Unabridged Software and my work there. I’m also happy to talk more with you if you feel like this is something you want to do in your organization, whether you’re developing software or doing something else. See my About Page for contact information.

Share Comment on Twitter