• Week 356

    28 October 2019

    Another week of writing: this time, writing a course on HTML and CSS for total beginners. This is slightly more practical than last week's: more about doing than thinking.

    But gosh, it's hard. It's a very short course, and so we're going at a lick to cover a reasonable tour of the environment. Very quickly, we get into jargon and complexity, despite my best attempts to avoid it. One reason for this is my avowal to avoid jargon when I can, but to never tell you things that aren't correct. So when I have to introduce a thing as a declaration or property or selector… I do exactly that.

    Fortunately, the browser, and markup, are forgiving environments: they do their best to render whatever you throw at them, regardless of mistakes (and in part for long historic reasons). And we're doing everything in the browser, without having to install software - CodePen makes for a great beginner IDE.

    I'm also relying a little on asking the learner to trust me in the course of a short article. I hope that's justified. What I'm really asking them is: “if things have briefly stopped making sense, can you trust me for a little? I promise to bring things into land in a little while - but we might have to briefly leave the ground." Or, in plain English: I can give you clarity paragraph-to-paragraph, but not always sentence-to-sentence. Perhaps that's a weakness in my writing style.

    Still, having just written a course about more imperative styles of code, it's nice to look at languages where I can show a learner the majority of the syntax in a few hundred words - and where the complexity is located in the implementation detail.

    Of course, there are still cans of worms that I'm finding and then having to write my way out of. Writing about CSS, I was talking about styling rules overriding one another, and used this as an example:

    p.intro {
      color: #ff0000;

    I am most familiar with writing CSS colors in the format #RRGGBB, where RR, GG, and BB are two-digit hexadecimal numbers; it's also what I've seen used most in the world.

    And yet this adds extra complexity for the absolute beginner. I have a limited word count, and do I all of a sudden want to lose 100 words to explaining hexadecimal numbers? Not really. So I replaced it with this, for the time being:

    p.intro {
      color: red;

    which lets us focus on the real idea I wanted to write about (namely, how p.intro overrides and extends p).

    In some down hours, I spent some time wrapping my head around KiCad. All my electronics work to date has been done with EAGLE. I've spent long enough with that piece of software that I've developed a kind of Stockholm Syndrome for it, despite its unintuitive nature (along with a large parts library). But a particular open-source project I'm working on is not ideally suited to EAGLE, especially if other people are to work on it, and I feel that future versions of it may be better supported in an open-source tool. Hence investigating KiCad.

    I've never really got my head around KiCad before. Partly that's time; partly that's it behaving much more like most schematic capture/EDA tools (where you do one, then the other), unlike EAGLE's merging of the two functions. It required a shifting of my mental model.

    What helped me overcome this hurdle was Chris Gemmell's Getting to Blinky 5.0, and I wanted to write about that quickly because it's helped me thinking about teaching new topics.

    Getting To Blinky teaches KiCad by going through the whole process of making a circuit board. It's a simple circuit: a timer that flashes an LED, with a coin cell battery. The course takes the viewer through making a schematic symbol, drawing up a schematic, annotating it and matching it to footprints, making a new footprint where one is missing, laying out a board, and getting their board fabricated.

    What I like about the course that it covers “one of everything”: at the end of it, you have done each major task that you'll ever need to do exactly once. You can then adapt what you've learned to your own work. I found it particularly helpful coming from another tool as, at each stage, I was mapping what Chris showed me to what I knew. I like the focus on end-to-end, with a simple (eight component) board - and in particular, teaching making symbols/footprints early on.

    I am not sure what it's like for the total beginner - it goes quite fast, for sure (although the pause and rewind buttons are underrated features of video content). But for me, it gave me exactly enough of the tour that with a little bit of time each day across the week, I had designed a development board for a potential project from to-to-bottom in it:

    Also, it has a surprisingly good 3D layout preview and raytracer in it.

    This project is to be filed away for November, but it was nice to pick up a new tool, and thinking about how somebody else was teaching it helped me think about my own teaching.

  • Week 354

    20 October 2019

    Weeknotes this week are dry, I'm afraid. I am still writing Longridge. I'd hoped to move onto writing my second course, but spent much of the week wrangling feedback on the first.

    On the flipside, the feedback was all useful, and easily sorted, by and large. It should also help to make the other two courses a bit faster to write, now I have a better idea of what's expected. But it's put me back a bit.

    Mainly, I rewrote exercises, edited an interview, drew up some quizzes, recorded one more voiceover, and then two more chunky screencasts that had me deep in Hitfilm's composite clip editor to tell the story correctly. Phew.

    I did get some time on Tuesday afternoon to spend thinking about an exercise for the second course - actually taking the plan in my head and walking through it myself. The good news is that at least half of it should be reasonably straightforward; however, I'm going to have to solve the second half of the exercise in the course of writing next week. Still, it's had me ruminating on the issues ahead, so that's a good thing.

  • Week 353

    11 October 2019

    A week on Longridge.

    I wrote an end-to-end draft of the first of the courses. Primarily, that was writing prose, but it's more than just long articles: I'm writing a mix of informative essays, the odd quiz, and exercises for learners.

    The exercises were versions of ones I've run before, but they prove particularly time-consuming - as well as writing them, I need to run them myself to check they work, and write scripts or articles to describe successful outcomes.

    But they're done, and now they're being reviewed by other members of the team.

    I also spent some time on my end of the videos for all three courses. That meant more assemblies of interviews to end the week, once the course content was shipped, and also submitting sample screencast videos recorded at home to see if they'd pass muster with the production team. It sounds like they will - so now I've just got to get on with the recording there.

    My home screencast recording setup goes like this: I record screen and audio simultaneously. I use Quicktime Player to capture the screen, and record audio from my external microphone as I do - a Red5 condenser mic going into my audio interface. At the same time as recording the audio+video in Quicktime, I also set the audio recording in Logic Pro, where a simple channel strip is setup on a mono channel. When I finish the recording, I open the video and both audio tracks in a video editor, align the Quicktime audio with the processed audio, and then disable the Quicktime-recorded audio. That way, I get the captured video with the processed audio. It sounds and looks pretty good to my eyes and ears.

    That made for a full week - not much time for much else. Next week, we move onto writing a full draft of course number two.

  • Week 352

    6 October 2019

    Back after a few weeks away, and straight to the studio. No time to build up the boards and stencils that arrived for Dent: instead, I went straight into writing and content review on Longridge.

    I got my head down on writing content so the team could review my style and give me some pointers. I managed a first draft of about a third of one course; I need to wrap up the rest by next Wednesday.

    I also reviewed the interview films that had been shot whilst I was away. The team kindly provided transcripts and timestamped footage to go over, and I had to work out what should go into the final film. It's hard to do this solely from a transcript - you end up finding that intonation is key, or that sometimes the way somebody's said something just flows more naturally in a different take. So I sat down with the footage and Hitfilm and hammered out a couple of ‘assembly’ edits - nothing final, and they wouldn't contain any of the cutaways or B-roll that the video team had for the final film - but enough to get a feel for if the final edit we were hoping for would flow. I know I could have just supplied a script or a cut-list, but it helped reassure me that what I hoped was possible would work.

    In a spare moment, I fixed a firmware bug on 16n that had been nagging a lot of people for quite a while. No great breakthrough on my part; instead, a well-framed bug report suddenly made the issue very clear, and half an hour in an editor got the bug fixed. Now I'm just finding people to help me confirm there's been no regression, and then I'll try to package that up into a release in the coming weeks.

  • Week 350-351

    22 September 2019

    Out of the studio, on vacation. Back at the beginning of October!

  • Week 349

    14 September 2019

    Head down on production for Longridge. That meant a few phone meetings with the production team, finalising a few more participants, and lots more pre-interviews with video subjects.

    That's all done, now, and the video shoots will be taking place over the next couple of weeks.

    I also spent some time preparing a few more scripts for non-interview films, and getting drafts of them out for review. One - just trying to describe in simple terms, the round-trip of a browser request - flummoxed me for a long time; it got lost in hundreds of words on DNS and ended up baggy, boring, and over-long.

    A few days away from it and I found a different approach. I sat down and worked out what the point of this film was - what I wanted learners to come away from it with. I knew this in my head, but needed to emphasise it to myself, so it was time to write out the big ‘beats’.

    Seeing them laid out gave me a slightly different direction on the same story, and allowed me to bring it in well within our time limit. Yes, it simplifies a few things, but it still tells the truth. The change is it now ‘spends’ its time on the key beats, and pulls focus on things that we don't need to know about (yet). If you write for a living, you likely know this lesson innately; for me, it's reminder of muscles I need to keep exercising.

    Finally, before I head off for a vacation, I generated a “revision B” of the boards for Dent and ordered them - along with a plastic solder stencil. They should be with me when I return, and perhaps we'll have more joy. I'll be head-down writing Longridge then, but a project for spare moments.

  • Week 348

    8 September 2019

    Head down on Longridge this week.

    I wrestled with a script for an animation early in the week, and just couldn't get it to anywhere I was happy with. I clearly need to zoom back out a level, and reminder myself what work this video is doing. Then I can work out what I want to show in it. I have a feeling the trick is, for what I want it to convey, to say far less. There's a follow-up piece of text that I can go into further detail if I want to.

    Another four-minute script went a little better, so that's in review now.

    Otherwise, I've mainly been thinking about production tasks: finding interview subjects, performing pre-interviews, writing them up as quasi-scripts for the production crew.

    I'm fortunate that I learned a bit about this during making Future Speak, where Kirsty - the excellent producer - ran a lot of these. “Pre-interviews” are where the interview subject gets phoned up ahead of the ‘real’ interview… and we discuss what we're going to ask them. This does a few useful things. Firstly, it prepares them for the questions they'll definitely get asked; for documentary-type work, there's nothing to be gained by anyone being surprised. Secondly, it allows us to help the interviewee frame their responses in the way we'd like - we can tell them a bit about what our target audience might or might not know, or ask them to be prepared to explain particular things in more detail. Finally, it also allows us to be surprised. When somebody says something interesting or exciting - great! We can dive deeper and add it into our script. Or if an expert in a field offers a better way of framing something than I might have asked - let's put that in, too.

    I take notes in writing, but ask if I can record them, just in case I miss anything or want to return to them. Piezo proves invaluable in that regard. There are lots of ways of recording audio from apps - such as VOIP calls or hangouts - but I've found Piezo to be reliable and remarkably simple; no messing with routing, just click some buttons and hit record. Really useful to have.

    Several interviewees all lined up and prepped, and a pile of short scripts drafted. Week 349 should see me sorting out the final interviews and submitting final scripts.

    But with Dent benched for a little while, that was my week.

  • Week 347

    1 September 2019

    On the projects front:

    • Longridge is getting off to a good start. I'm hunting down contributors, doing introductions, firing off emails. I'm also beginning to write some scripts just to get my hand in and to see if they line up with what the production crew are expecting. Not quite where I wanted to be by the end of Friday, but the team seem pleased where we are, and I had several good meetings with them during the week so I'll take that and carry that momentum into next week.
    • Dent is going back to the drawing board, but possibly not for a month or so, given Longridge combined with impending holiday. It turns out the current build just won't behave; I spent some time with a J-Link really prodding at the thing and no amount of work in gdb would solve my problems. So we'll start again in October.

    I spent a little time getting myself quickly up-to-speed with some new technologies this week - always good to keep my hand in, and at least one of them was an audition for Longridge:

    • I finally wrapped my head properly, clearly around CSS Grid; gosh, what a breath of fresh air. It makes two-dimensional grid layouts just a breeze to write. I also managed to understand its relationship with (and differences to) flexbox. After a long period of reading docs, I now had understanding in my hands. Good.
    • To that end, I managed to tear out all of Bootstrap from this site - which had used it for positioning. Replaced the lot with CSS Grid and Flexbox in under an hour. Very happy with that - and also happy that the trade-offs for users of older browsers are, to my mind, totally acceptable.
    • I continue to poke and prod at Javascript frameworks, and spent an evening (off-duty) kicking the tyres on Svelte. I'm interested in Svelte for how it approaches the reactive-UI pattern, but also how it removes the load on the client by just… being a compiler. Given the Javascript community's fondness for (or, perhaps, reliance on) transpilation, going the whole hog doesn't seem that much of a difference. I quickly ported my personal ‘hello world’ of reactive UIs to it, and enjoyed the process, the clarity, and the light page load. I'm still not sure how I feel about the ‘magical’ binding and slightly higher need for simple stores; I found myself using stores and subcription in places React would force me to explicitly pass functions and state down, endlessly, which to my mind is both good and bad., Still, I like its approach to reactivity and its single-file component structures, and especially its approach to performance and webbishness.

    And that was the technology I played with; not project-oriented, but useful for me in lots of ways, and worth sharing here, as a log of what I get up top.

    A good week - but gosh, the next two are going to be busy.

  • Week 346

    23 August 2019

    Longridge is now fairly positioned on the launchpad.

    The structure and broad content of the courses are now signed off; that means all that's left is writing them (and making a number of films). I had a call about content creation for the project, and that means I'm going to be kicking off equal amounts of writing and production tasks next week. A little bit nervewracking, but we've got a plan and a list of things to burn down, and so the job now is to work through it and make sure everybody's happy. I'm going to be making some spreadsheets to keep track of everything (or, maybe, some crazy walls)

    I also kicked some tires on an exercise I was thinking about using on Longridge, and everything seemed to work as I hoped, so that's a goer. One less thing to worry about.

    Over on backburner projects, my prototype PCBs for Dent arrived early in the week, along with the stencils for soldering it up. So I spent Monday bringing the board up: getting the paste applied, getting components on the board, baking them on, and then looking to flash the firmware.

    Which, not entirely unexpectedly, just didn't work at all. Not a complete write-off - I can see the board over SWD, but no joy flashing it, and the main thing that happens when you power it up is that the microcontroller gets really hot. That's… not the desired functionality.

    I wondered if it was my solderpaste, so I built up a second board largely by hand, with only hot-air used for the QFN MCU, but no dice. I went back to the drawing board - EAGLE - and decided that I'd definitely been having some strange ideas about how the schematic should come together. I drew up a Revision B, from some better reference material, and perhaps I'll send that off soon-ish and we'll try again. I also spent a while double-checking everything I could about flashing Atmel SAMD51 chips, and learned a fair bit: I might well use a different toolchain next time around.

    Anyhow, it's probably worth another iteration, so that's a thing to be going on with. It's going to be a few weeks before new revisions will arrive, which is a good thing, as I'm going to be busy! A long weekend coming up; Week 347, we begin some typing in earnest.

  • Week 345

    18 August 2019

    A quiet week spent exploring with my hands.

    I continued my explorations of Dokku as a hosting option, spending some time looking at best-practice for deploying Wordpress to it - and then working out how that lined up with my own preferences around deployment (notably, specifying Wordpress as a dependency in composer).

    That was a useful exploration - getting hand dirty with some servers, turning something I'd often end up doing by hand into something more automated - and I at least got to something sensible and working by the end of the week.

    I spent a day documenting what I'd done. Not ‘documentation’ that resides in the text files I use as a digital notebook; ‘documentation’ other people would find useful. That's a challenging thing to write clearly and unambiguously. The end result was perhaps too long, but I think it explains itself clearly enough, and includes appropriate code snippets and samples.

    I hope I'll publish that somewhere next week. If nothing else, it was good practice to write some real documentation for something technical, multi-stage, and complex. I've got more writing coming up this month, and it was good to get a handle on my pace of writing (and, for technical work, getting a feel for how often I need to bounce between edits and writing). And, who knows, it may still be useful for someone.

    On Friday, my circuit boards for Dent arrived from China and I set about assembly. These are the smallest component sets I've ever had to assemble - the 0603 is all fine, but the QFN microcontroller has been a pain. (QFN is probably the first of the surface-mount IC packages that is ultra-hard to do by hand, as it has no exposed pins. It's really designed for a robot and oven to do).

    I've been practicing with solder paste, stencils, and hot air, and whilst the first stage of my assembly seemed to go OK, the usb socket on the board has possibly not gone in correctly, leading to some powering issues, a very hot chip, and a frustrated me to end the week with.

    Still, I have an idea where to take it next week, and I solved a few problems at the firmware/bootloader end for once I've got the chip up and running.

    Quiet, but busy, and some good studio conversations with my colleagues throughout.