News for hackers. Stuff that matters.
Re:The cybernetics scare and the origins of the Internet (2009) by Terretta on Monday September 27, 2021 @11:52AM (#28672231)
> "Cybernetics" as a descriptor went out of style by the 2000's I think.

A bit off topic, but in the ‘science of automatic control systems’ sense -- one could argue today it could be pronounced ‘kubernetes’.

But imagine the consternation when we first collectively realized k8s had no cybernetics, no auto-governance (another related word*) of tech running on it. This is why a few of us met with CoreOS in the mid-2010s to talk about the idea that would turn into Operators: put the cybernetics back in kubernetes.

Disappointed even today that so much mindshare and effort is around standing up and running k8s itself, or ‘declaring’ what runs in it (tbc, this part is good), and much less on ‘Operators’ so k8s could run your SaaS for you (which would be better).

The original public introduction of the concept:

- Brandon Philips @ CoreOS, Nov 2016: https://web.archive.org/web/20170129131616/http://coreos.com...

A more recent take:

- RedHat, 2020: https://www.redhat.com/sysadmin/kubernetes-operators

---

* Note: “The word ‘governance’ came from the Latin verb ‘gubernare’, or more originally from the Greek word ‘kubernaein’, which means ‘to steer’. Basing on its etymology, governance refers to the manner of steering or governing, or of directing and controlling.”

Re:The cybernetics scare and the origins of the Internet (2009) by _jal on Monday September 27, 2021 @12:40PM (#28672900)
> ‘science of automatic control systems’ sense -- one could argue today it could be pronounced ‘kubernetes’.

Only if one's view of the world is very constrained.

Re:The cybernetics scare and the origins of the Internet (2009) by Terretta on Monday September 27, 2021 @01:14PM (#28673404)
Well, I agree.

And yet, are there any other generalized cyber control systems even remotely usable to manage arbitrarily shaped workloads other than kubernetes (which isn’t that either, not yet)?

Why not?

Re:The cybernetics scare and the origins of the Internet (2009) by kragen on Wednesday September 29, 2021 @12:52AM (#28690411)
Today "κυβερνητική" is pronounced "kivernitiki", not "kubernetes" (which at any rate means "steersman", while "κυβερνητική" means "government") but I'm not sure if it's in common usage.
Re:The cybernetics scare and the origins of the Internet (2009) by dogman144 on Monday September 27, 2021 @12:36PM (#28672842)
Actually on topic!

Taking that hypothesis you have on kubernetes, check this out...

The Cybernetic Hypothesis book discusses a philosophical/systems-thinking/governance obsession with the importance of "steering" or "piloting." Foucault taught a class on in it in the 80's prior to his death. The train of logic is interest in steering all systems that could be steered for <various reasons> led to a natural interest in cybernetics, as the automation offered a way to codify or "steer" this systems.

Kubernetes <- derived from the greek "kubernesis," "the action of piloting a vessel."

You're highlighting, I think to the extent I'm still grokking this all myself, a common counterpoint to cybernetic thinking - it offers a way to steer things, but often doesn't address core questions like "well who is steering this, why do we want to steer this, what other answer is there beyond efficiency for efficiency's sake?"

Re:The cybernetics scare and the origins of the Internet (2009) by paulpangaro on Tuesday September 28, 2021 @06:41AM (#28681219)
The discussion here, so far as I've followed it (only beginning) is focusing on what is called 'first-order Cybernetics' about the role of feedback and information in systems that steer some variable toward a goal (think: pilot of a ship; thermostat). 'Second-order Cybernetics' (see von Foerster, Glanville) does ask, 'who is steering' and brings in both subjectivity and responsibility in an instance.
Re:The cybernetics scare and the origins of the Internet (2009) by dogman144 on Tuesday September 28, 2021 @07:18PM (#28688586)
Nice distinction. Based on what I’m reading lately, sounds like I need to chase down more of the latter.
Re:The cybernetics scare and the origins of the Internet (2009) by convolvatron on Monday September 27, 2021 @12:52PM (#28673077)
I read the slides you posted - they seem quite reasonable in intent.

why though do you take a jab at declaration as being a worse model - it seems like a valid implementation of the general operator idea.

Re:The cybernetics scare and the origins of the Internet (2009) by Terretta on Monday September 27, 2021 @01:12PM (#28673374)
Sorry, declaration is fantastic. But the blueprint or playbook of what to stand up (say, Ansible style) is very different from cybernetic control of operating things.
Re:The cybernetics scare and the origins of the Internet (2009) by convolvatron on Monday September 27, 2021 @05:19PM (#28676163)
what about declaration for operation?
Re:The cybernetics scare and the origins of the Internet (2009) by Terretta on Wednesday September 29, 2021 @10:12AM (#28693906)
I have not seen a DSL allowing specification of a SLO.

Or any kind of nominal behavior convergence tooling or automation.