Yeah, it's very far from usable by other people at the moment =/.
The readme at https://github.com/kunalb/termdex/tree/main/markdown_files is probably the best bit.
HN user
https://explog.in (I keep forgetting my HN account; this is the latest copy.)
Yeah, it's very far from usable by other people at the moment =/.
The readme at https://github.com/kunalb/termdex/tree/main/markdown_files is probably the best bit.
I've built a couple for myself so far; the most recent is in zig (sqlite extension that treats markdown files / frontmatter as virtual tables) and it's lasted me. I plan to rewrite it soon to adapt to how I've been using it :)
I gave a small talk on how to really push using hooks for logging intermediate values (including capturing gradients from torch & fx scripted modules) that may be useful: https://static.sched.com/hosted_files/pytorch2023/40/Interme...
This was the first time I bought a font, and -- surprisingly -- I haven't regretted it at all. Every time I look at my terminal there's always a little bit of happiness at how elegant the font looks.
Happy to hear that!
Heh, definitely depends on the software itself but I generally find the contents /schema of tables used to save the data very illuminating: you can see the UX, whatever form it takes -- and then what is saved to the backend so it makes things slightly more understandable for me.
I've been writing here sporadically for more than 10 years at this point, at ~1 post a year. The more recent posts took months to write, and tend to cover things I find myself repeating frequently while working with other engineers.
- https://explog.in/notes/elephants/index.html: Tips for ramping up on large projects
- https://explog.in/notes/devtools/index.html: Building developer tools
Planning to overhaul it later this year.
I'll throw in my own too: https://github.com/kunalb/panopticon
Build my own project with it, ideally for fun -- if it's a tool or a programming language. Try to apply it immediately if I'm learning something new.
"you don't have to code all parts of your design".
That's an excellent articulation of what I think as "invisible seams" when I write code: they're soft points of extensions that don't need a separate interface/function/class _yet_. Sometimes I just mark them for myself with an extra newline within a function.
Huh, I wrote about this a few months ago too (with a lot more whimsy): Eating Elephants -- https://docs.google.com/document/d/1c07-Zj6bUbYPwx7Zttd1N74o...
The thrust of both our articles is similar, except I think I ended up using many more words.
Agreed, writing is necessary but not sufficient. But writing is still much better than nothing.
As much as I enjoy writing to think, simply writing doesn't lead to reality testing -- I've found even greater clarity by trying to add numbers or simulations and trying to program around what I'm writing. Writing with real data intertwined and numbers applied is much better.
I used this in AoC for the first time a couple of years ago, as a consequence of diving deeply into asyncio. https://explog.in/static/aoc2019/AoC23.html
Small Gods by Terry Pratchett, and the whole series on Discworld greatly affected how I think about life, and to look beyond first assumptions.
I have 3 suggestions! because that's how I tend to work too
- Scapple (mac/windows only) https://www.literatureandlatte.com/scapple/overview I love how minimal it is, but wish there was a linux/web version. I'll probably make my own some day.
- kinopio (web/mobile) kinopio.club Works well, but has a strong, quirky personality which sometimes gets annoying for me; I prefer the minimalism of scapple.
- draw.io / diagrams.net You can configure it to hide most of the toolbars and then just use it as mind mapping software, this is what I'm trying at the moment.
I haven't yet cleaned up the slipbox code enough to publish it, but I do have my site generation code public: https://explog.in/config.html -- which does something similar but doesn't generate the cross links.
The slipbox code is just a bunch of python to pre-process the org files and "fix" links, which then passes them on to emacs subprocesses to generate html.
I use logseq to take study notes in org mode, which are then published to my public slipbox by post-processing with emacs in a github action. The consistent part of this has been plain text notes -- the slipbox has stuck for the past year, and hopefully will continue.
I built https://github.com/kunalb/panopticon to export perfetto/chrome compatible traces and also draw arrows between async functions. (I think the arrows are only supported in about://tracing though).
https://explog.in -- a few posts over the past decade. I've tried to keep it fairly simple and straightforward.
Generated with emacs running in a Github action (https://explog.in/config.html); comes with a slipbox (https://explog.in/slipbox/index.html) that I fill in using LogSeq and export to a flat website using emacs again.