After uploading the video of my Rust/Yocto Project demo from the Yocto Project Summit last month I received a lot of positive feedback. This got me thinking that there was scope to record some more demos and some more free-form hacking sessions and share these via my YouTube Channel. But why spend the time and effort to record, edit and publish these videos? I'll try to explain some of the motivation here.
From discussions on the Yocto Project's mailing lists and IRC channels (#yocto and #oe on Freenode) it's clear that Josef's streams have been really valuable to new users and new members of our community. Not everyone learns well from written material, some people need to see how software can be used to get a feel for it. I've even learned a few things myself from these streams, especially the recent one on working with Windows Subsystem for Linux (WSL).
While I've been involved with the Yocto Project community for over 7 years now, I've only recently dipped my toes into contributing to the Linux Kernel. Kees Cook's streams have helped me understand a little better how maintainers work with and review contributed patches (or at least how one maintainer does). I've also picked up a few tricks for using git and other tools which I didn't know before.
What I want to achieve
It's probably easiest to start by explaining what I'm not trying to achieve here - the things that are explicitly outside my scope. I'm not trying to build a hugely popular channel with a large audience, pick up ad revenue or provide perfectly polished tutorial videos.
What I do want to do though is fill in the gap for intermediate and advanced level videos on Yocto Project and particularly for material that shows people how they can contribute to the project. I'm hoping to show how to hack on the openembedded-core metadata, on the project documentation and on bitbake itself. I also want to capture my experience of learning Rust as an embedded developer as I don't think this is as well covered in video content as other approaches to learning rust.
I also want to create videos which show the process of figuring things out as I go, of hacking and learning and debugging. Seeing only perfectly polished tutorials which show the direct route to a solution with no missteps inspires a bit of imposter syndrome and can make folks feel stupid if they can't figure things out straight away. While I plan to edit out long compile times and any uninteresting fights with my virtual machine setup or network, I will be leaving in mistakes and the process of discovery as I work on things.
My other goal here is to improve my ability to present what I'm working on and to explain things as I go. I don't think I'm terrible at this but I do think there's a lot of room for me to do better. With many conferences remaining virtual in the coming year I think this will be very valuable.
I'll be releasing edited videos onto YouTube rather than live streaming. Editing videos obviously adds to the effort required here but it saves me a lot of up-front effort as well. For the videos I've posted so far I've done very little advance preparation, I've just dived into recording and edited out any major dead ends or long compile times afterwards. I can also join a few shorter sessions from different days together into a single video where appropriate which helps me to fit this around other projects and responsibilities.
I'm using OBS to record my screen and webcam together so that I don't have to sync and overlay the videos during editing. For video editing I'm using kdenlive which looks sufficient for now but I still have my eyes open for other options.
Videos will normally be released at 11:00 UTC on Fridays. Depending on other commitments I'll likely be releasing a video every week or every fortnight but I don't give any hard guarantees. I recommend you subscribe to my channel and turn on notifications rather than checking each week.