Michael F. Lamb @datagrok@cybre.space

rich filmmaker debacle
cliché filmmaker beard
firmed clickable harem
clarified member chalk
childlike farm embrace
backfield mile marcher
half remember acid lick
crab milk remedial chef
cerebral milkmaid chef
bleached farm limerick
flaccid hiker embalmer

anagrams of my full name

Oh hey, keybase.io has native support for many Mastodon instances now.

https://keybase.io/datagrok/sigchain#1c2279e92adf150f042922f1774b677efa50c9b8f0395713e391312a3924ef1a0f

Feeling smart because I implemented a Ring List Zipper in #elixir to help me solve a programming puzzle 🤓👍

https://github.com/datagrok/adventofcode/blob/master/2018/adventofcode-elixir/lib/adventofcode/ZRing.ex

It's similar to the List Zipper from the "Zippy" library, but includes some unique functionality.
https://github.com/tchoutri/Zippy

Newbie adventures in trying to lazily tokenize an input stream with #elixir

https://gist.github.com/datagrok/1d13ceac855a324ced0c270e07b1ec51

With a link to someone's massive blog post on the same subject with a more thorough walkthrough: https://www.poeticoding.com/elixir-stream-and-large-http-responses-processing-text/

Well, that took a while but my changes to add support for the SH1106 OLED display driver IC are now merged upstream into QMK master! 🎉 https://github.com/qmk/qmk_firmware/pull/5787

Does this mean I can now put "C-language embedded driver development" on my resume? 😲

(no, my changes were tiny)

I got a little 128x64 I2C OLED display for $2.50. It said in the product listing that it used an SSD1306 driver IC, which made me think it would work with #QMK firmware.

I hooked it up but it only showed static.

So I read some datasheets and learned some stuff and did some hacking and now (or very soon) QMK will support this OLED display too! (Turns out it actually uses an SH1106.)

Pictured: the wireless receiver for my Mitosis #mechanicalkeyboards, with a speaker and this OLED display attached.

python zip(*rows) is matrix transpose

>>> [(1, 2, 3), (4, 5, 6)]
[(1, 2, 3), (4, 5, 6)]
>>> list(zip(*_))
[(1, 4), (2, 5), (3, 6)]
>>> list(zip(*_))
[(1, 2, 3), (4, 5, 6)]

>>> [[1, 2, 3], [4, 5, 6]]
[[1, 2, 3], [4, 5, 6]]
>>> numpy.transpose(_).tolist()
[[1, 4], [2, 5], [3, 6]]
>>> numpy.transpose(_).tolist()
[[1, 2, 3], [4, 5, 6]]

hey it rhymes too

🛎️ Website update:

- Still a bunch of CSS cleanup to do, but I've got marginalia working! 😍 Soon I want to get a "light theme" toggle going.

- Lots of revision to my article re: "purist free software systems shouldn't shun non-free programs"
https://datagrok.org/gpl/shun/

- Some revision to my "exec the arguments" draft, including implementation in bash, c, python, rust, and scheme
https://datagrok.org/src/exec-args/

Do you find that on some sites, emojis show up as outlines instead of color images? If so, try uninstalling DejaVu fonts and installing Bitstream Vera fonts instead.

https://github.com/eosrei/twemoji-color-font#why-bitstream-vera

… So I am excited to say that as soon as I get all the detritus swept up from having ham-handedly smashed their CSS together with mine, I'll have nice mobile-friendly marginalia and sidenotes on my Markdown-source-format webpages! It's already mostly-working! I've wanted this for so long! 😁

So I recently looked into how to get margin notes on my webpages again. In the time since my last attempt:

- Edward Tufte himself has published a pure-CSS open source mechanism for implementing them,
https://edwardtufte.github.io/tufte-css/

AND

- Someone published a plug-in to make Pandoc (the best markdown processor) (that I was already using to build my pages) take advantage of it,
https://jez.io/tufte-pandoc-css/

AND

- That tool has been packaged for Debian so getting it running was dead simple!

Through an accretion of heinous hacks the web has been turned into an application delivery platform. But ostensibly it's still about documents, right? So, it shouldn't be too hard to get some pretty Tufte-style marginalia on my web site.

… Is what I said the first time I attempted it, years ago, before rage quitting in disgust with CSS after burning days of my time.

(I was in good company: I later saw famous smart guy Gary Bernhardt tweet that he had the same experience! Ha.) 1/3

I've got a shell script with

while
<expression>
do
<statement>
<statement>
...
<increment>
done

But I want the body to always run at least once. So I want something like a do-while. But shell has no do-while.

But! The block after while can be a list of commands; only the last exit status determines its truth value.

So to transform the above into a do/while it's (non-intuitively):

while
<statement>
<statement>
...
<expression>
do
<increment>
done

TIL Two popular textbooks on mathematics for data science (one for math experts and one for everybody else) both offer complete downloadable PDFs at no cost!

An Introduction to Statistical Learning
http://www-bcf.usc.edu/~gareth/ISL/

The Elements of Statistical Learning
https://web.stanford.edu/~hastie/ElemStatLearn/

(Grumble grumble I wish they were epub though)

👀 Updated draft: "The 'run the arguments trick,'" up to 3k words now on this ode to these four little characters: "$@"

https://datagrok.org/src/exec-args

Added new Rust and Guile Scheme examples, plus lots of revision.

Just a few more improvements then I can rip the "draft" warning label off.

Attempting to implement this bit of shell script in #guile #scheme.

exec "$@"

My first working attempt used eval with quasiquoting and it looked very messy. Second attempt:

(let ((args (cdr (program-arguments))))
(apply execlp (car args) args))

It works, but is it idiomatic? Is this how you'd write that? Can it be shorter or clearer?