Roman the Kat

Well, recent news mean that Stack Overflow Jobs gonna be discontinued on March 31, 2022. I liked this one. Kept looking from time to time what’s interesting out there, how do vacancies look like. For me it had kinda intuitive interface. Lots of things on the page, but most of things I would need are just right there, ~1 click away. Pretty straightforward to my taste. I also gathered ‘statistics’ of opened vacances per languages....

It’s a big thing for me to have an endless canvas for note taking. Sometimes it’s really useful. For example if you want to be able to easily look on the whole picture, just zooming in/out. Or if you have no idea how big the result will be, having no limits calms - you won’t run out of space, you can fit everything on one page, somewhere nearby. So I’ve being checking infinite/endless canvas apps available at the moment....

tl;dr It’s worth having ’notes’ and ’todos’ separately, probably in separate apps. They are too different. If you search for ‘productivity tools’ being used in a wild, you usually see apps for note taking/making. But it often comes in pair with some todo-lists app. For me it feels the reason is - vastly different lifecycles: there are ’notes’, made as a ~long-term thing; and ‘action items’, a ~short-term things to execute....

I rely on several apps for reading, mostly offline ones. Probably the main app for me is Pocket. There are a lot of great articles you can find, and most of them will be saved in there. Some will be tagged (top priority/first to read things, usually). On the other hand, I’m kinda curious in wallabag (self-hosted open source solution with mobile version). As a data source I often use RSS, for quite a set of blogs....

Fun thing about JSON - it doesn’t have dates or time support, as per RFC. So whereas you see something like 2021-02-18T21:54:42.123Z in JSON - it’s still just a string. Some languages firmly follow that, like nim stdlib parser. Some try to parse as if it were ISO 8601 time. Which is actually nice - it’s human readable, sortable, portable, well supported and precise. Though such support sometimes causes extra fun, like java’s jackson lib and JSR 310....

An interesting and predictable happens with note taking during scaling. Once you have about 400+ notes ‘plain’ structure becomes not efficient. Seemingly it’s become harder to keep the context of everything you are writing about. The whole idea of note writing is to not trying to keep everything in your mind all the time. That’s the place where outlining works - it adds structure. There are numerous similar terms - outliners, structure notes, index notes, hubs....

I like listening to podcasts. That’s a great way to use, to spend your time. There are tons of good and different podcasts on various themes. Some are useful, some are entertaining, some are both. Of course, with whole pandemic situation one of the main timeslot for listening podcasts (for me) is lost - commuting to work, but there are still options when to listen, like routine things, house chores, and so on....

There is an open-source programming language called Nim. It is a statically typed compiled language, looks a lot like python (small amount of syntax noise combined with spaces for identation/code blocks). Example from official site: import strformat type Person = object name: string age: Natural # Ensures the age is positive let people = [ Person(name: "John", age: 45), Person(name: "Kate", age: 30) ] for person in people: # Type-safe string interpolation, # evaluated at compile time....

As usual new major update for Ruby lang was announced on 25th of December. Cool thing about this year - that’s a release 3.0, aka Ruby 3x3. Main things are: performance (x3 in certain benchmarks) new concurrency primitives (actor-like abstraction called Ractor and Fiber#scheduler for async event pool) optional typing (RBS files for types hinting and typeProf tool for types inference) As for me, things I somewhat missed in Ruby were type hinting and more convenient concurrency primitives, especially once same ideas were added in Python (with async syntax and type hints)....

One of little details left behind with feature branches - the moment code reaches release. Popular branching models approach it mainly in two ways: once feature development is finished, and code review passed, one shall merge the feature branch; e.g. merge before feature validation feature branch deployed separately to conduct testing procedures, bugfixes if needed, and only after that you merge the feature branch; e.g. merge after feature validation So it’s a tradeoff whether we deploy features faster, or more reliably....

For a few days I’ve being reading zettelkasten related articles and previously mentioned notes of Andy Matuschak. Making totally about 235 articles. That was quite a long read. Tons of interesting stuff, resulting in some notes of/for myself. Speaking of notes - so far ‘web of interconnected notes’ idea works fine for me. Connections you make usually based on your thinking and associations, as a result navigation feels natural - easy to remember and use....

Seems like I mostly finished the migration from Org-mode to a set of markdown files. 337 so far. That took a while. Also that was quite a mess. Removed tons of entries which aren’t useful anymore. Haven’t migrated all the work/current project data yet, but didn’t really need that - gonna migrate as per necessity arise. I’m gonna keep the things this way, for a few months at least

From time to time I review the way I do my notes taking. I have a lot of notes on different themes, both personal and work stuff. And for last few years I’ve being emacs org-mode. I’ve tried quite a lot of different things and apps, from simple text files without any structure up to specialized tools. Yet still I come back to org-mode. For a quite some time I’ve being reading about [[202012051855 Zettelkasten]]....

I like it so far. Task 3 was basically about simple iterating. Task 4 can be solved using state machine, though data validation in hard part was a little bit hectic(missed java/kotlin spring validation thingy), so I wrote quite a number of unit tests. Task 5, as title ‘Binary boarding’ suggests, is about using binary search algorithm, a few times. Task 6 reminds me task 4, but was somewhat easier to my taste....

Well yeah, task 2 is even easier than the first, to my taste. Current IntelliJ IDEA Nim lang plugin doesn’t exactly work for me. It correctly highlights the code, understands declarations - but it has no idea about stdlib, and has no autosuggestion. Vim with plugin is more convenient as for now, but I like the idea of officially supported language plugin in such a popular tool - that could end up with something interesting

Long waited Advent of Code 2020 has started. First task was obviously simple (though hard part made the simple part code obsolete). Solved it in pretty bad Ruby. The second task looks simple, it won’t be much bother too. Gonna use Nim this time.

So the idea of using ML in food delivery notifier service did and did not work out. On the one hand - everything went relatively smooth. It was good to remember how to use python and ipython notebook, prepare and parse data (finite state machine for the win!), try out few regression models, make some ‘insights’. On the other hand - accuracy is really low. It does predict ‘scale’ right, hour-two, but makes severe mistakes....

There is a simple project of mine called food delivery notifier. It checks current food delivery status and shows it in system tray. As for me - a bit more convenient than constantly refresh the page. Currenlty supported service has a map and coordinates info, it shows courier location. So knowing how far away a courier is from restaurant or your place allows you to assess how long you’ll be waiting for delivery....

For me it’s always a pleasure to do some refactoring. The whole idea is about re-structuring, creating more clear, poilshed, sane, understandable code structure. More expressive, less obtrusive lines which just get the things done. In a more direct way. Tearing a problem apart, re-assemble pieces, solve a puzzle. Just like it

It’s been a while since I used Rust lang in any way or form. I wouldn’t say I’ve done anything noticable, or even useful it in, but still - kinda interested in this language, tons of little reasons. As a sort of warm-up before advent of code 2020 I would like to re-read The Book of Rust, take a closer look on 2018 edition of the language, and ideally try a few coding tasks using Rust specifically....