===============
Continuations
===============

No prominent events for me this month, mostly continuations of the
past themes.

I hoped that this month I will finally have no news (rants) on things
breaking. There were unpleasant but mostly expected changes, such as
increased taxes, increased utility (including Internet service) bills,
with general prices also going up, extension of mandatory genetic data
collection to administrative arrests (likely leaking soon after, as
government-collected data generally does, and as Rostelecom's user
data did recently), but those are not a sudden breakage. Ran into a
few more blocked websites, but that is a regular occurrence, not quite
news. But RKN keeps doing their thing, making use of its increased
budget, and this time they have blocked all the TLS connections (on
all the local ISPs, country-wide) for about half an hour, via TSPU
(actually it was just one out of the three recent large-scale
RKN-caused incidents: those happened on the 6th, 14th, and 24th of
January). Local news agencies reported major ISPs claiming that
everything is fine on their ends, and RKN later claimed to detect and
fix a connectivity issue. They still do not admit YouTube blocking,
either, with a court recently declaring that there is no evidence of
that, while someone tried to sue RKN (probably to express the protest
that way; though among millions of people there may be some who
actually believe in local courts functioning as they are supposed to:
people believe in all sorts of things).

At least nothing new broke for me personally. The newly set router
works fine, and I learned about the recently released OpenWrt One
router, which looks neat, though only has two Ethernet ports. There
are other Banana Pi and similar boards, and then one may consider use
of larger computers as routers, or simply to run some services on
those. I thought of moving some daemons from my primary desktop
computer to a small and quiet machine like that, to depend less on the
primary computer, but still unsure whether it would be worthwhile.

I had occasional (rather frequent, actually) unexpected naps following
insufficient night sleep. Naps are nice, but they harm the schedule
and perpetuate night sleep issues. And then it is wasteful to spend an
hour or more just trying to fall asleep.

Speaking of the lack of time, there is still plenty of work to do, but
we are now looking for somebody to help with my projects. Will have to
adjust the workflow then, but hopefully it will help with sorting out
the backlog. Unlike it is in the vacancies with oddly specific and
long lists of requirements I see sometimes, I listed just a few
primary technologies (mostly common and general ones, not some new and
niche libraries or services, with only Haskell being uncommon) and
familiarity with one broad area of expertise. I hoped for an influx of
enthusiastic applicants, but among the first ten, most only had two or
three matches out of five (usually Python, PostgreSQL, sometimes
Linux, just once Haskell). Most link their GitHub profiles (while some
seem to confuse git with GitHub), but with those only containing
synthetic projects: test tasks from job applications, practice or
study materials, not regular hobbyist FLOSS projects. Though those are
not required, but nice repositories, participation in others' FLOSS
projects, with properly written commit messages and documentation,
could have made up for the lack of matches in technologies. The search
is complicated by the position being on-site and in a smaller Russian
city (because a manager resides there).

Advertised it on Mastodon as well, and explored Fediverse a little
more. Many posts are political, even on technologies-related
instances. Given that the current situation here can be blamed on the
lack of political discussion and participation back when it was
relatively safe to engage in, perhaps I am not in a position to
criticize the excess of such discussions. But in addition to there
being much of that, it looks like an echo chamber, and though mostly
it advocates social equality, previously I spotted dehumanization of
religious people, recently saw prejudice against straight people;
rather like right-wing messages, but inverted, directed against other
groups. Thought to point it out in a reply, but not expecting a good
outcome from engaging in a political argument with
strangers. Considered reporting it, but while that server's rules
prohibit certain kinds of discrimination, they do not prohibit it in
general. But with some time and effort, it seems possible to compose a
nice personal feed there, filled with topics of interest.

Meantime, while working on a sub-project that includes a web
interface, I have been asked by a frontend developer to adjust the API
in a strange way that seemed to both complicate it and restrict the
possible configurations. In a conversation, found that it is caused by
the common tendency to pull an interface closer to the other models
one works with: to the UI in this case, making the API structure to
reflect that of the UI. I find it awkward, especially given that the
UI may change, and this project is supposed to provide a generic API,
but figured that this frontend and its developers are likely to be the
only users of this API, so it makes sense to target them, prioritizing
them over hypothetical and unlikely other users, and not to spend time
arguing over it. As a bonus, the potential future changes will not be
able to mess up a nice interface (and the corresponding model), with
it already being ad hoc.

In other news, I decided to restock perfume. Learned that the choice
is limited now, many brands are unavailable, including Hugo Boss,
which I used for the past two decades. Picked Eisenberg J'OSE and D&G
K instead; both are Italian, which made me to wonder whether they are
available because of the kind of government they have out there
now. Also noticed that perfume branding reminds me of computer
hardware branding: both are embarrassing, with the computer hardware
often aiming kids (including all the awful LEDs and main memory names
like "Vengeance", which always makes me to imagine a silly kid being
beaten in a game, and then acquiring that memory, hoping that it will
help them; or CPU cooler names like "Peerless Assassin"), while that
of perfume often aiming douchebags. But possibly in both cases they
figured that those who pay attention to specifications or scents would
mostly ignore the branding, at least consciously, and on the others it
will work. Maybe all the blockchain/cloud/AI branding works in a
similar way for certain kinds of software and online services.

Next, saw this nice toy OS implementation walkthrough, with QEMU and
RISC-V, and went through it:
<https://operating-system-in-1000-lines.vercel.app/en>. It reminded me
of the recently mentioned nandgame.com: this one is less interactive,
but still gives a toy hands-on implementation example of an important
technology, which is used widely, but is obscure even to many
programmers. Though there are quite a few things like that in CS, and
then in mathematics, physics, and plenty more outside of those: they
look important and neat, and as if it would be nice if more people
knew at least the basics of those, but unsure whether it is viable,
even if everyone was able to dedicate a chunk of time daily to studies
in varied areas.

Then I learned about the "Physically Based Rendering: From Theory To
Implementation" book, <https://pbr-book.org/>, which looks good. It is
much larger than those tutorials, the project it builds is not that
simplified, but apparently it still goes through the whole
implementation: laying out the theory for every aspect, then providing
an implementation. I did not read it yet, only skimmed, but that
should be a nice reading for when I will have more of spare time. And
possibly after reaching the part on optics in the physics textbook.

As for the physics textbook, I went through its chapter on equilibrium
and elasticity relatively quickly, since there were holidays, and
yesterday finished working through the chapter on fluid
mechanics. While I tend to rant about the annoying AI hype, found LLMs
to be occasionally useful with physics problems: they are not reliable
for getting correct solutions, but when you are stuck, they may help
by suggesting a step to take, helping to find where you have made a
silly mistake, or even simply working as a rubber duck for
debugging. Though people can be similarly helpful, including more
complex questions, and that can double as an occasion for
communication; the ##physics channel on Libera.Chat, for instance,
helped me recently by confirming a mistake in the textbook answers
that I was uncertain about. Now I have two chapters in the part on
classical mechanics left: gravitation and periodic motion. Maybe
before starting the next chapter, I will try to advance in (or finish)
some of the other books I started reading.

It is two years now since I started doing the same aerobic routine
daily. I noticed changes in the first year, but not in the second
year: perhaps just maintaining the shape now. Which is less exciting,
but also fine. As for other physical exercises, I am skipping the
third quick stretching (balancing) routine, though maybe will resume
it later. And heard of the reverse plank exercise recently, trying it
out: holding the regular plank for 3 minutes, and the reverse one for
2 minutes. As the regular plank, it looks like a nice, quick, and
simple exercise.

Today I mopped the floors, and it is still more tiring than those
exercises.


----

:Date: 2025-01-25