Advent Of Code, Day 9:

Advent of Code Day 10

Advent of Code Day 11. Drama: Lamdu got stuck in my live stream 😱 but I’m also doing this to catch bugs like this.

Advent of Code Day 12:

Advent of Code Day 13. Lamdu: for the current lack of UI for defining nominal types the parsing is interleaved with the solution today..

Advent of Code Day 14: Ooh, looks like nobody's doing AoC with http://clerk.vision so far? Maybe not super FoC-y, but I got curious after hearing @Jack Rusher mention it. It's been super fun to play around with viewers and stuff, and it's extremely gratifying to have a presentable "notebook" to share with friends almost for free. My solutions are at http://zmb.cm/aoc/2022 with an auto-generated landing page that lists the individual days, so no spoilers from just clicking that link :)

The world is cranking through AoC and zooming toward a ChatGPT future, and I'm still here thinking about the right way to visualize past versions in a live-programming environment. And my note-taking app (https://spectra.video/w/051fef50-a121-40c0-bf3c-ee0730fb4a07). One thing I did recently: a

move

command that moves columns (analogous to browser tabs) to the left, while continually truncating columns on the right beyond some limit. The combination of these two has changed how I work, from messes all over the surface to much more localized access patterns where I live near the top left and move things over to it as I need them. (It's not a catastrophe if I lose a tab because on-disk search is working well.)

Advent of Code Day 15:

Advent of Code Day 16:

Clerk evaluated '2022-16.clj' in 1.1024717767137E7ms.

I know there’s a smarter solution, and I came close, but at 4am in the morning, leaving my laptop running as I go to sleep was the reasonable thing to do 😅

Advent of Code Day 17: Having visualizations close at had was super useful to me today, but even so I still got bogged down in false-start implementations. But the final solution works pretty well, I think. https://scl.zmb.cm/@T045T/109531068871567834

Aaand that's a wrap on Advent of Code! https://zmb.cm/aoc/2022/2022-25.html Today was nice and easy, so I had the time to write down some warm and fuzzy shout outs (shouts out?) at the end. I do plan to go back and make my solutions for the past week more presentable...

Been working on the foundation of my new editor. I set a goal this year to use a text editor I built my self for 10% of my development. So I'm starting with some new foundations. The most frustrating part of the my previous experience was the compile/run cycle for adding new UI. This time I'm starting with an extension mechanism that is powerful enough for me to prototype features live. But also, I want to be able to integrate extensions I made into the core of the program as things solidify. To that end, I've prototyped a little wasm extension mechanism. It is still early, but I can hot relaod code while keeping state. I can save state and rehydrate. I even have plans for being able to do some time travel debugging/snapshotting. Right now the extensions can just do simple drawings and respond to clicks, but the foundation is there and I'm excited for it.

it is not connected to programming yet, but I am trying to explore “less is more” principle and made typewriter on canvas prototype https://uprun.github.io/canvas-typewriter/canvas-typewriter.html it is all written in single html file

Kicking the tyres on my environment with a BF interpreter. This is inspired by https://tratt.net/laurie/blog/2023/compiled_and_interpreted_languages_two_ways_of_saying_tomato.html

Today I gave each app a default map for its code, so that people get a better initial experience when connecting the driver to it, rather than the blank canvas they've been faced with so far. Initialization is one of the categories of semantics that's hard to write tests for. What should happen when a config file is present. The driver maintains separate configs for each app it knows about, so it needs to handle the very first app vs first view of a new app. What if the app has no default map. Etc., etc. Anyways, I ended up touching https://git.sr.ht/~akkartik/driver.love/tree/main/item/Manual_tests.md after a while.

a snippet from my weekly blog this week (pretty niche tode/bret lore content) (please moderate if not appropriate)

Spending most of this week taking tree-structured explanations, and generating something approximating a series of paragraphs, then annotating them with cross references like a graph for navigation purposes. The result is persuasive for my target audience. An explanation that they can read and understand and navigated easily is a big deal for demonstration purposes. But I am hating it, because it is technologically trivial. Any tips for getting through the slog of actually doing the thing that is time consuming, valuable, but uninspiring?

Also, if anyone has any examples of an interface that uses an auto-fill to let you select a prototype statement, then let's you fill in the blanks in that prototype, I am looking for UI inspiration.

Before and after. Does the paragraph form seem "better"? Took a surprising amount of work, because I had to abandon the NLG being done by the reasoner and write my own version that knows how to deal with nested terms. I'm thinking the new one is probably better for validation, worse for debugging, but I'm not sure.

hi everyone, I have created another canvas experiment https://uprun.github.io/canvas-typewriter/canvas-documents.html and now I am not sure if this is the way to go for me … maybe next time I will start with pen&paper prototype first

Today I put together a little visualizer for Mastodon threads. The graph layout is nothing special, but: • it has structured keyboard shortcuts (up for parent, down for child, left/right for siblings) which might be more useful for understanding a complex thread • as always you can live-edit it so it's hopefully easy to improve One other little integration: opening links in browsers opens up all sorts of cross-platform cans of worms, so I ended up just copying a URL to the clipboard any time you click on a toot. https://git.sr.ht/~akkartik/mastodon-unfurl.love

Been making good steady progress on my new editor. I now have my wasm extensions automatically interrupt. So I can have a long running computation (or even an infinite loop) happening in an extension, on the same thread and not miss any frame times. Having that really sets the ground work for the live programming model I want for these "extensions". After some clean up I'm starting to do some dog fooding for another project I'm working on, generating an arm64 assembler from the xml specification of instructions. The plan is to write the coding utilizing my editor as the visualization environment to get fast feedback on if I have the instructions correct. Still a long way to go, but the fundamentals are taking shape.

Thinking about building some visual programming tools after taking a break for a while. One fun idea that popped up: hybrid text + node-wire environment where you can wire to the text. Like, a word in the text can act as a node. Probably been done somewhere, but I haven't seen it, so I'm kinda jamming on the thought. Like, what if Natto but you exploded the frames, brought some of Bret's Tangle in there, little bit of that style of annotating an image+paragraph by coloring a word the same as a corresponding thing in the image. Constants in the text live-update, etc etc.

FWIW - here’s how I think about it... You are programming a MACHINE. Every bit of syntax has to DO something (have a meaning). [The meaning of comments is “to be ignored by the machine”. A lot of other syntactic baubles have that same meaning - e.g. stick people and clouds]. What does it mean to connect a wire to a word? What happens if you connect that same wire to a different word?

Hi friends! This is my first time posting in the devlog, but here's today's progress on my structured editor -- autocomplete for object attribute access!

trying out "file attachment" as a first-class AST node type

today's progress, emoji support! it mostly consisted of

text => new Intl.Segmenter('en').segment(text).map(s => s.segment)

in a bunch of places, so that offset calculations made sense. I also had to add a hidden

<input/>

onto the page instead of just harvesting

keypress

, because emoji-pickers don't actually send valid keypresses, but they do trigger

oninput

in a focused

<input/>

🎉

Managed to release v1.5.0-alpha of Blawx this week, with better defeasibility features including defeated defeaters, and applicability checking. Only one major feature left to add to the language (that I am aware of): temporal reasoning.

Can’t believe I just found this channel after all this time in the server…been doing a bit of “building in public” on my Twitter, but it just goes into the void…but this community never fails to disappoint 🙂 Today’s update: Just got offline mode to work in Subset!