roman-kashitsyn

What version of Chrome do you use? The images are encoded in .webp for compactness; I hoped that most of the browsers in use support this format, but I might have to go multi-source after all.

> I always underline ш and write a line over т (which looks like m) to distinguish them

I used to do it as well! But I gave up on it for the same reasons I described in the article: impaired working memory.

> лишили лилии

That’s a tough one, but I think it’s possible to make it legible with careful choice of spacing to indicate the groups: the distance between the strokes within a letter should be noticeably smaller than the distance between letters.

> I also normally write д as ∂, which breaks the flow.

I also like this variant of ∂, though I eventually switched to the version with a descender. It doesn’t annoy me because it introduces only a short pen lift.

> In text, as in code, I prefer to optimize for easy reading rather than faster writing.

Author here!

It really depends on what you write.

When I write something for others to read, I type it or (on special occasions) write it one letter at a time with a calligraphy pen.

I’m the only intended reader for almost all of my handwriting (personal journal, article drafts, exercises, and notes). In this setting, it’s much more natural to optimize the script for enjoyment, convenience, and speed. But even then, I think my script is still quite legible.

> you shouldn't make decisions on how to develop aerospace software based on how "authentic" the development process feels.

You don’t just land in an aerospace company with an unbiased, rational mind and make an engineering tradeoff. You join it because you find building high-integrity software fulfilling (I am!), and the job enables you to exercise choices that feel right to you. Alternatively, you are so passionate about space that you can endure whatever it takes to get there, including writing Ada/Spark.

The best swimmers don’t grow high and develop long limbs and a big chest due to their practice and rational decisions. They become best by already having a matching physique and working hard to capitalize on it.

I don’t promote a fixed mindset, saying that you can’t get better if you weren’t born with a natural talent. My point is that some things will resonate more with you and bring you the most satisfaction. You should try different things, listen to your inner response, and not judge others because they made different choices.

Author here.

> because whoever is different from, and possibly smarter than, you is obviously autistic

I’m not sure how you came to this conclusion from the text I wrote.

Autism has nothing to do with intelligence level.

BTW, I identify as autistic. And I find Haskell very soothing.

This article explores the possibility that our technology choices depend more on our upbringing than rational choice.

Those are good points; thanks for raising them!

> Not for writing. But you'll have fun reading this file.

Regular file reads aren’t much more complicated; they require only one extra call to stat and a few extra file seeks.

> Read the file from STDIN or from a socket, where you only get one pass over the data without buffering.

That would be a problem for files you want to keep in memory. My assumption with tensor files is that they are so large that you will always save them to disk before using them. That’s how HuggingFace tooling works anyway.

> Read the file, but the file is truncated by one byte (i.e. the last byte is missing).

That might be an argument for using the floating header! However, I don’t expect it to be a real issue with tensor files specifically. If they aren’t intact, they aren’t useable, and you should throw them away. Error correction should be one level higher.

I believe Sam is a predecessor of Acme[1][2], which still attracts new users today. One of its notable users is Russ Cox, the tech lead of the Golang team. I haven’t used Acme yet, but I want to try it out one day.

If you’re interested in learning more about ed, I highly recommend “Ed Mastery” by Michael W Lucas [3]

[1] https://doc.cat-v.org/plan_9/4th_edition/papers/acme/ [2] https://www.youtube.com/watch?v=dP1xVpMPn8M [3] https://www.amazon.de/gp/product/B07BVBSDNZ

Nice!

I used https://wiki.systemcrafters.net/emacs/org-roam/ for a while but switched to LogSeq (https://logseq.com/) because org-roam was buggy.

I like working with LogSeq, but even after a couple of years of using it, I’m not convinced by the Zettelkasten method. Maybe I’m doing it wrong!

I worked with John Wiegley for a couple of years and discovered one of his projects, Ledger (https://ledger-cli.org/), during our conversations. This tool taught me double-entry accounting and helped me understand finance and blockchains on a deeper level.

Unfortunately, I’m too lazy to use the tool on a daily basis, so it’s the second most insightful piece of software I’ve never used :D

I was thinking about including Mathematica as enlightenmentware. Mathematica 6 (https://www.wolfram.com/mathematica/newin6/) was the first truly interactive system I used (we happen to have a box with a license at the university). It impressed me so much that I still have a lot of warm fuzzy feelings toward Stephen Wolfram and his work.

Unfortunately, my relationship with Mathematica didn’t go anywhere: It was too expensive back then, and I never found a good use for it except for double-checking my homework.

I tried other computer algebra systems, but they didn’t impress me as much.

If you own a Mathematica license and found a good application for it, please let me know!

> But once in a w̶h̶i̶t̶e̶ while

Thanks, fixed! It's “occasionally” now. Even Grammarly didn’t find it :(

Haskell was one of my most important discoveries; it affected my thinking and approach to software engineering the most.

However, as I mentioned in the opening section, I deliberately removed programming languages from candidates for several reasons: 1. They received enough praise already. 2. My presentation would be very biased. 3. The article would be way too long.

I tried God Mode and Evil several times, but it never ended well. Mostly because of the muscle memory I accumulated over a decade.

Evil is fantastic, but it doesn’t play well with all the packages in the ecosystem. There are adapters for Magit, Compile, etc., but most packages define key bindings that require Ctrl-Meta cords, which causes cognitive dissonance.

God mode is a neat idea, but I never committed to it, and I found it confusing at times.

Overall, I decided to accept Emacs as it is, with all its quirks, and embrace its way of doing things. I’m not married to it, so I can occasionally cheat by switching to Vim or VS Code without feeling guilty.

Bazel confused the hell out of me at first, and I think the two-phase execution model (the “plan-execute pattern” as I called it) is to blame.

My favorite thing about Bazel is how easy it is to get stuff done if somebody sets up the rules for you. Copy-paste a code snippet and fiddle with the dependency list until it works.

But as soon as you go deeper, you get overwhelmed with new concepts, and the documentation doesn’t explain them well enough. I think this huge spike in complexity makes people hate Bazel, especially if their colleagues force it on them, breaking the usual workflows.

I don’t love Bazel, but it’s the build system I hate the least. And it taught me a lot.

Fixed, thanks for reporting!