To control my micro Sawppy rover, I want to present a HTML-based control pad that works even on older phones like my Nokia Luma 920 running Windows Phone 8.1. This will require that I revisit the world of HTML development and this time I’m trying to avoid client-side frameworks like jQuery. Eventually, these HTML files will be served from an ESP32 on board the rover, but I didn’t want to use an ESP32 while I’m focused on client-side development. The most obvious reason is that I didn’t want to upload a new ESP32 image every time I made a change in my client-side code. Fortunately, the nature of Jamstack-style web development meant that my HTML (& associated files) sent to the browser are all static files. So I could use any static file web server to act as a placeholder for the ESP32 while I work on the HTML side of things.
But given my recent experience installing development frameworks, I was squeamish about installing another one on my desktop without some way of keeping it isolated from the rest of my system. Again I turned to solutions already developed for the web world and decided to use Docker containers. I’ve dabbled with Docker on a few prior occasions, and this project is a chance for more experience.
Once Docker Desktop for Windows was installed, the opening screen invites me to run the
docker/getting-started container. This is a Docker tutorial on multiple levels. Not only does it take me through an increasingly complex set of scenarios on how Docker can be used, it is itself a container that has a web server hosting the tutorial content. (For those that want to take a look before installing Docker, a substantially similar tutorial is on docker.com.) So after running through the tutorial, I understood enough to come back and poke around inside the
getting-started container to see how it was done. Which was very useful for me, because a container acting as a web server is exactly what I want to do right now.
As neat as Docker looks, there are a few problems preventing me from using it for everything. The major limitation for me, at least on Windows, is the lack of hardware access. This rules out all the projects that need access to a USB port. So while I could run ESP-IDF inside a container, I wouldn’t be able to flash the resulting image to a ESP32 nor would I be able to use JTAG debugging. Docker is most closely associated with network applications, not hardware, so it’s best to stick with its strengths as I use it to solve my project problems.