6 minutes read | 1160 words by Ruben BerenguelSome links are affiliate links
This feels like a heavy engineering edition. A lot of Haskell, Rust, Python and
Scala. There’s still a bit of everything, but this will appeal hardcore
developers more than usual.
NOTE: The themes are varied, and some links below are affiliate links.
Haskell, Rust (and Spark), Python, formal methods, maps. Expect a similar wide
range in the future as well. You can check all my weekly readings by checking
the [tag here](https://www.mostlymaths.net/search/label/ReadingsOfTheWeek, Readings). You
can also get these as a kind-of-weekly newsletter by subscribing
I had a “I need a planner” project in my personal project list and was
considering a lame solution with Prolog (very 70s), but this is so much better,
and a great excuse to finally write something in Haskell.
An implementation of the RDD interface from Spark in Rust. The speed up is
promising, the code looks clean and clear enough. Has lots of potential, if I
have some time someday I’d like to help with some Arrow stuff here.
Caveat: won’t respect the python environment you start it on. Aside from what,
it offers jep for data interchange between cells written in Scala and cells
written in Python. How cool is that? It also offers auto-graphing (like
Zeppelin) and a very convenient autocomplete/symbol expansion system. All in all
seems to have huge potential, and is usable from moment 0.
Instagram has a very large Python monolith, and of course this means some
problems that are minor for others are huge for them. Here they share how they
handle the unconstrained “power” Python module loading “offers”. I’m looking
forward this being added to some Python version.
I’m not 100% convinced of the approach, since it feels overkill-ish (and Alloy
is pretty terse, so that feels like an additional downer), but I like seeing formal methods
approaching the data engineering space.
Writing git commits is hard. I’m trying to work on it, but sometimes I still
write TICKET-### fix bug. At least I have stopped naming my branches using
in_soviet_russia jokes. Although I kind of miss that.
I’m a kind of map junkie (and I don’t mean only Wardley
maps). Have several books on maps,
many maps (more than I can hang), I GIS-ed a custom map of my
would love to take some workshop on hand-drawn mapmaking at some point (does
this exist?). So this is just excellent.
I read this one last weekend. Had too much overlap with other books I have read,
so the new content was minimal. The only useful thing I got from it was learning
there is a form of prose known as flash fiction, ranging from 10-1000 words.
I’m kind of attracted to the idea of writing 100 word prose, so I have
investigated a bit more that one.
So, all told, a category is just a monad in the bi-category of spans.
This has taken me very long, and has sprouted several side readings of
category theory papers, a renewed interest in Haskell, and a list of exercises
from the book I want to try. Also a renewed interest in re-reading stuff about
schemes and sheaves (Hartshorne where are thou?). I’m not sure I recommend you
read it, it’s long.
Holly molly. I’m afraid of giving spoilers, but… So, here David
Beazley starts by writing a simple stack
machine in Python. Extends it a bit, not much. And then uses it to interpret
and play (through pygame) a game written in Rust and compiled into WASM. There
are like 4 instances where the public applauds in the middle of the talk, and I
felt like doing the same at home.
These weekly posts are also available as a newsletter. These days (since RSS
went into limbo) most of my regular information comes from several newsletters
I’m subscribed to, instead of me going directly to a blog. If this is also your
case, subscribe by clicking here.