📚 node [[stpa handbook]]

[[Chapter 2]]

Good overall summary in the first page:

[[Pasted image 20210718144905.png]]

I wonder what the formalism is in this diagram. It is not an STPA diagram in the "main" sense as it doesn't model control loops explicitly, but is it part of the STPA framework?

[[Pasted image 20210718150828.png]]

First step: defining the purpose of the analysis.

[[Pasted image 20210718150742.png]]

[[Losses]]:

[[Pasted image 20210718150713.png]]

[[Pasted image 20210718155038.png]]

[[Stakes]] are [[values]]:

[[Pasted image 20210718151545.png]]

Definition of [[system]]:

[[Pasted image 20210718162616.png]]

[[hazards]] + [[environmental conditions]] = [[losses]]

[[Pasted image 20210718231941.png]]

The heart of STPA: the control loop.

[[Pasted image 20210718234343.png]]

Refining control structures:

[[Pasted image 20210719000525.png]]

[[Pasted image 20210719000516.png]]

[[Pasted image 20210719001721.png]]

[[Pasted image 20210719003446.png]]

[[Pasted image 20210719003520.png]]

On [[control]] proper:

[[Pasted image 20210719005502.png]]

tips to prevent common mistakes in a control structure:

[[Pasted image 20210719010243.png]]

Step three:

[[Pasted image 20210719010634.png]]

⥅ node [[hypothesis]] pulled by user

Hypothes.is

  • extend client:
    • read from multiple API-sources
    • read tags from API (not just from cache)
    • integrate with [[web-clipper]]
    • integrate with foam's wikilinks
    • retrofit discussion threads to follow [[kialo]]
  • clone and adapt API-server to support plugins for fetching personal notes for some url from:
    • twitter
    • slack
    • discord
    • reddit
    • gmail: fetch also tags
    • keep
    • ...
    • page flags for:
      • github issues (authored & commented)
      • chrome & firefox plugins (to know you have installed them)
    • AND/OR use [[promnesia]] browser extension (requires a lot of python configuration)
    • see also: [[Agora Bridge]]
    • see also HPI library
  • Hypothesis: what you see with your eyes closed, the lights you can see in the dark -- that is not random noise, it is computation.
    • The visual cortex may be used to compute arbitrary programs given the passage of time: an image can encode a spectrogram, for example, and pattern recognition applied to shapes in the spectrogram may yield at least some useful sound information (given the "right" neural networks)
    • Arbitrary programs could be stored and replayed in the visual cortex by e.g. evolving cellular automata or a neural network that can simulate them, which would seem to yield universal computation as per Wolfram et al
  • Hypothesis: there is much in the nature of [[light]] we still do not know.
⥅ node [[stpa]] pulled by user
📖 stoas
⥱ context