Electron Adventures: Episode 100: Series Retrospective

#javascript #electron

113 days ago I decided to get back to tech blogging, and to play with Electron a bit.

I roughly estimated how many episode I'd need to explore what I wanted, and after reaching the number of about 100, I decided that the only way forward would be to post about once a day. And I did exactly that (including the planned two week holidays).

Blogging platforms

I have a long running blog at blogspot, but it's all HTML based, with dreadful editor, and it's actually somewhat hard to post code there. I needed to write some Markdown offline, then convert it to HTML, and then find a cat picture and post it, all really tedious and adding 15 minutes of pointless busywork per post.

There was also no good way to get much readership for my posts, since Big Tech jointly killed RSS, and forced itself to serve as intermediary between writers and readers, to sell ads and control what people read.

I wasn't sure which platform would be good, so I started two blogs - one on dev.to, and another on hashnode. They both accept Markdown, so the only extra work was posting screenshots and tagging.

Analytics

I didn't really promote this series other than on my Twitter account, with very little active following. I got about 90 views per post on dev.to and about 20 views per post on hashnode. There's no clear pattern why some of the posts got more views, other than the obvious first few.

Overall these aren't very high numbers, but considering how obscure the subject was, and total lack of promotion, I guess that's about what should be expected.

What I Learned

I already knew a bit about Electron, Svelte, and React starting the series, so most that stuff was just casual practice.

I had an opportunity to try out some new JavaScript frameworks like Vue, Malina, and Marko. I also had an opportunity to try out a lot of Electron alternatives like NW.js, NodeGui, Pywebview, and so on. To be honest I don't think any of them would become part of my regular toolbox.

How Much Work It Took

I didn't measure it, but I think it took me about 2h per episode. Sometimes I would have some free time during the day, often I'd fall behind the "one blog post per day" target during the week and write a bunch of posts in a row during the weekend.

I used Github Copilot while writing. I think it's actually better at tech writing than at coding, but even that saved me maybe 10% of work. For coding, much less than that - I was jumping from one obscure tech to another, and it usually had no idea what was going on.

About 20% of the time the idea I had couldn't be turned into a good episode. A few things I retried later. Of the ones I eventually posted, I think Opal Ruby took 3 tries. Imba took 4 tries, all without success.

I think it was important for the success of the series to be willing to drop an episode that's not going well, and try something else. If I kept forcing all the difficult ideas, it would take twice as much time.

What Did Not Work

I wrote about it back in episode 81, and actually covered a few of these (Opal Ruby, a few more Electron alternatives) in the final episodes; as well as gave Imba another unsuccessful try.

I think the most obvious still missing parts were FFI and WASM. Both are doable, but they'd take way more time than I was willing to spend on them on just boring parts.