I have finished (re-)reading Aristotle's "Nicomachean Ethics", this
time more slowly and thoroughly, and with more context, than on the
first reading (or skimming) a while ago. As expected, much of it is
about "good" and "evil", but there are many interesting speculations,
and it is simply interesting to consider older perspectives, to
observe both how the prevailing views change, and how people do not
seem to change much. As with Plato, there were ideas that I imagined
to be more modern: e.g., "I think, therefore I am" (though noticed
that Wikipedia lists Aristotle in predecessors to the Descartes's
"first principle"), and references to (in the form of arguments
against) Protagoras's relativism ("Good man, in that he is such, is
the measure of everything"). And a worrying proposal for states to be
more involved in citizens' personal lives towards the end. All in all,
as with many older texts, it looks like a decent effort to build a
model, using the information and resources at hand: coming up with
something usable and a fine approximation, even if it is not the final
one. And it is a useful book for discovery of connections between
major materials and ideas, for the context.
I have also recently read or re-read chunks of a few other major
philosophical works from different times; such an overview makes it
all the more apparent how those lean on their contemporary scientific
theories and models, not to mention the evolution of views. Even
though I was somewhat familiar with most of that before, as I guess it
is hard to miss major ideas completely these days, adding the details
and checking out the primary sources (well, mostly translations) is
still interesting: like reading a book or watching a movie, rather
than its plot summary.
In addition to those, I decided to check out another introductory
book, "Introduction to Philosophy" from OpenStax. Unlike some other
introductions, it is not a teaser or a summary, but a textbook, with
good suggestions on how to approach texts: explaining cognitive
biases, suggesting to "steelman" arguments, to keep in mind that many
philosophy books were written by smart people, but living in a
different world, even including a section on information
literacy. Those I would include into an introductory text as well.
Following my mention of a history book in the previous post, agk of
sdf.org shared a list of history-related books, one of which I have
read then. A fictional one, which was easier and quicker to read, as a
break between more involved reading: Aitmatov's "The day lasts for
more than a century" (but in Russian: "И дольше века длится
день"). Enjoyed its mixture of a rural life with drama in a historic
context, fictional mythology, and science fiction. The science fiction
bits apparently had an issue: mentioning sentient beings from another
galaxy, who can travel at the speed of light, but somehow the flight
to Earth taking them mere 26 to 27 hours, and the communication seems
to take no more than that. While Andromeda is millions of light-years
away, even Proxima Centauri is a few light-years away. But the
accuracy in that was not the point of the book, since as the spoiler
placed in the introduction said, it was included for the sake of a
metaphor; it is not a book to help with daydreaming of interstellar
travel.
Other news
==========
- Acquired an OpenWrt One router. Checked that it works, but did not
play with it yet. Planning to use it primarily as a small
network-connected computer, to move some services there, and as a
backup router.
- I have set biboumi (an XMPP-to-IRC gateway), which seems to work
well, though did not use it much. Also have set XMPP and IRC
servers, along with biboumi, at work, to have a backup communication
channel, as well as a more secure one, but they are unused. And
finally tried out poezio (picked it as a client to have on the
server, so that an SSH connection alone is sufficient to join a
chat, over either IRC or XMPP), which looks like a fine XMPP client
(standalone TUI, akin to irssi or WeeChat).
- Other work adventures included a bureaucratic push for security,
complete with many weird documents, translated CVE and CWE
databases, and bastardized standards, which threatened to be a waste
of time, but allowed to do actually useful things: to update the
systems on servers, and to work on other system hardening, following
some suggestions produced by a scan with lynis. Though I guess that
it is common in such systems to have a few major custom-made
vulnerabilities, which by far outweigh any hardening: the latter is
like reinforcing sections of a wall that has a few large holes in
it.
- LLMs reached my day job: we are hiring a junior developer to work on
making an LLM-based interface, as if the regular broken and JS-heavy
web interface does not make it hard enough to retrieve
information. Both as a user and as a developer I usually wish that
reliability and accessibility were prioritized over a stream of new
features, especially over incorporation of hyped technologies. While
struggling to think which projects actually benefit from addition of
LLMs: it seems that those are simply bolted on everywhere only to be
there, as distributed ledgers and SPAs were before. And then there
also are the technologies that developers themselves like to
incorporate for no good reason, such as external message queue
systems, unnecessary databases, microservices, excessive
containerization and orchestration, and just about anything shiny
and new. Meantime, the accumulated accidental complexity already
leads to loss of needed functionality: in the past few days I had to
direct one person to a direct interaction with a database instead of
a JS-based web UI, and another one to either an API or finding
somebody who can debug an SPA that does not work for some.
- Worked a little more with Python and GraphQL (FastAPI and Graphene,
to use packages from Debian repositories). Python does feel wonky,
compared to Haskell, even while I think I am quite open to liking
it. I would prefer SPARQL over GraphQL myself, if anything like that
was needed at all in the first place, but used it here for an
integration of a small service into a website that uses it.
- Looked more for languages usable and practical for implementing an
efficient and portable proxy server using their standard libraries
alone, went through the D language's tour. It looks more like a
"better Java" to me, rather than a "better C": a pile of language
features, far from minimalistic, but a neat pile at that. I think it
can be nice to write in, though did not try it yet. But its "fibers"
are not quite like lightweight threads, and its functions are
blocking, the polling it has for sockets uses plain select(), so it
is not easily usable for an efficient proxy server without
dependencies, either. Then I looked into Python, which does support
asynchronous and efficient networking out of the box. Quite a
contrast between the language itself being slower than all the
others I have considered, and the tools it provides out of the box
being more efficient.
- Local Internet censorship keeps worsening. Mobile Internet had been
temporarily shut down a few times, before and during the military
parade (for the first time, without a warning), unsurprisingly
leading to numerous disruptions. All websites behind Cloudflare are
blocked if ECH is enabled (which is the default in Firefox). SSH
connections from a remote continuous deployment runner to one of the
work servers fail during authentication now, possibly also
blocked. While the censorship is just a part of the overall
situation, which exhibits similar tendencies.
- I heard of Collapse OS before, and recently learned about Dusk
OS. Those are interesting projects, which I plan to look more
into. I cannot help but to think that for actual post-apocalyptic
scenarios, similarly to state-imposed isolation ones, it would be
more useful to focus on gathering and making freely available
documentation, textbooks, and information generally, rather than
systems and code, but those projects still look fun, and it
generally seems to be a good approach to educational projects to
make something simple, imagining such a hypothetical scenario.