• [[spaces to tabs conversion]]

  • I didn't like the result so I took the liberty to revert. Some rationale in the node above :)
  • Reach out if you think it's the wrong call! But I'd rather go with the Python community standard by default and not impose an extra burden on contributors (to override their editor defaults).

• DONE fix links in /journals

  --[2022-04-09 Sat 19:32:08] => 00:00:01

  • was higher priority than expected, see below :)

• LATER perhaps tell crawling bots to take it easy while I work

• LATER [[sqlite]] experiment

• LATER I should update to Python 3.8 so I can use f-strings with = at the end to print variable name and value in one swoop.

  • Surely there are also other improvements :)

• NOW agora load balancing

  • it is time.

    • update: actually it is not quite :) squashing the 500s bug mentioned below made the Agora a lot more efficient. this is still high priority but not burning.

  • once the next item is ready, this unlocks arbitrary [[horizontal scaling]].

  • I think I'll start with the simplest setup possible, solving balancing first and reliability later -- that is, add balancing to the [[nginx]] running in [[thecla]].

    --[2022-04-09 Sat 19:32:28] => 00:00:02

• LATER experiment with [[podman]] as [[docker]] replacement

  --[2022-04-09 Sat 19:32:23] => 03:46:32

  • continue going through [[please contain yourself]] with [[podman]]?

    --[2022-04-09 Sat 19:32:29] => 00:00:03

  • nice, got it running just fine; I said I was not going to default to [[rootless containers]] but they seem to work fine for simple examples, and the

    --[2022-04-09 Sat 19:32:29] => 00:00:02

• crawlers/bots are hammering anagora.org quite a bit, might need to actually write a [[robots.txt]] file to tell them to take it easy while we work on better performance :)

[[scaling]] must happen

• but we still have time thanks to the block below :)

DONE hmm, but there is [[low hanging fruit]]: the per-worker cache should not all expire in unison (!)

--[2022-04-09 Sat 19:10:28] => 01:17:51

• also I wasn't caching calls to G.node() (?).
• AND, much more importantly, the Agora was restarting all the time due to 500s in URLs hit by bots -- so none of the performance work I was doing was taking effect. Now that that's fixed it feels much snappier! I am happy about this development.

• [[almanac mode]]

• [[Businesses]] did call on violence providers to resolve [[disputes]] and provide other services, but once called, they often found themselves in a permanent relationship with the provider that they may not have wanted.
  • A dummy [[business]] partnered with a [[violence]] [[provider]] might setup and con customers. When customers found the dummy business, they would be stopped from punishing the dummy business by the violence provider. However, if violence providers from another area requested that the dummy business returned the unrealized investment, the request would be honored. On both ends, violence providers would gain between 30 and 50 percent as a [[commission]]. The prevalence of [[scams]] and recuperation by violence providers led to a sort of [[insurance]] industry, where credit providers and local businesses would end up paying a violence provider for insurance against scams and [[theft]]. The violence providers observed a [[norm]] of never scamming each other. In the cases where that happened, [[punishment]] was especially harsh.
    • [[Violence]] providers would 'create a [[problem]]' and offer to solve that problem.
    • The [[transitionary]] post-[[Soviet]] [[state]] was trying to [[privatize]] as quickly as possible. This rushing was an opportunity for [[violence]] providers, who came in to deal with [[debt]] [[defaults]], [[cash]] flow, [[contract]] [[enforcement]], and [[secure]] [[delivery]] of [[goods]].
    • For a [[business]] in [[Russia]] in the 1990s, a lack of [[transactional]] [[stability]] was regarded as the biggest [[problem]]. To deal with this, businesses relied mostly on informal [[relationships]], a little on the [[state]], and more on [[private]] [[enforcers]]. Many [[arbitration]] [[courts]] took more than three or four months to process [[disputes]], which gave violence providers an opportunity to step in. Some charged far less in transaction costs than the state.