Back from vacation! Turns out I’m bad at vacationing. Anyway, quick update on projects:
Crossover Project: I recently gave a 10 minute lightning talk draft on this. It went pretty well! I’m transcribing all of the interviews I did, and hope to have a new draft two of the talk by end of April. Yes, I’m also going to write essays, but writing talks is a bit easier for me. A talk is only 5-6k words, the essays are looking to be 10k+.
Alloy Docs: I was hoping to meet with Daniel Jackson in a couple of weeks, but coronavirus is going to make that probably not happen. I hope to dedicate a big chunk of the next two weeks, though, to finishing draft one of the docs. We’ll see if I actually hold myself to that, lol
This Is How Science Happens: Oh Hey, This Is Done! the final version is just under three times as long as the first draft I shared back in November. It covers the whole story in much more detail, as well as additional info about the importance of replication and why P-values are terrible. You can read it here. Thanks to everybody who helped out!
tlacli: work on this continues. The last time we got to the point where you can install the entire thing in a single Python package! It’s almost at the place where it satisfies all of my requirements; I’m not sure what the needs of the broader community are. So if you have any requests please let me know!
For the most part I’ve done corporate workshops, because they’re easier to organize and logistically much simpler. But I’ve also done some remote workshops for companies and it occurred to me, hey, why not try that publicly? So I like to gauge interest in people attending a remote workshop. Nothing certain yet, just trying to figure out interest. If you think this is something you would like to do, please fill out this form here. Thanks!
(I promise this wasn’t COVID-19 related. Been stewing on this for a while!)
Leo Trip Report
A couple of posts ago I said I was trying out the Leo text editor, which is a structured programming editor. Instead of everything just being flat text, you’re also able to work on the node structure itself. It seemed like an interesting idea, but wasn’t sure how would work out in practice.
After exclusively using it for a few days, I’m still not sure. I was able to write scripts that directly manipulated the nodes and found a few cool uses for that. But there was one killer problem: input latency. Very occasionally there would be a 40 to 50 ms delay between each keystroke. This made it infuriating to use. So I’m temporarily tabling that. I want to go back to it, but probably not immediately. I still want to prototype some ideas of how manipulating nodes directly could be incredibly powerful.
New workflow experiment: voice to text. I was inspired by Emily Shea’s Voice-Driven Development at Deconstruct. That version’s not online yet, but her StrangeLoop talk is. So I bought a temporary trial of Dragon and am trying it out for writing prose. I don’t think I’d use it for programming, but I can talk a lot faster than I can type. I used it to write out this newsletter, though I switched to typing for editing. Typing is very good for editing but not so good for just getting a lot of stuff out very very fast. And then I edit the word vomit. Circle of Life!
Also, it’s less stressful. I wrote this paragraph out in text and it felt like more effort than just spitting out the words. Maybe something to do with typos?
(Update on editing: wow there were a lot of wrong words. And it’s harder to find them than typoes since the red squiggly doesn’t appear. This might be less of a problem when I get more practice, though.)