Minutes 07 Jul 2023

Host: Paul Albertella

Participants: Pete Brink, Leonard Hubner, Igor Stoppa, Raffaele Giannessi, Kate Stewart

Agenda: Workshop review & Safety Analysis of Telltale use case

• Paul talked about ELISA workshop and discussion of safety analysis approach Presentation
  • In particular, using the approach as a basis for collaboration with other WGs e.g. Learning from experience using STPA in other groups (Medical - OpenAPS, Safety Arch - inside the kernel)
• Medical - used STPA to drill down to find how Linux was being used to support the safety-relevant parts of the system
• Safety Arch - finding it hard because focus at a lower level meant very complex picture

Pete: Did STPA on OpenAPS find any flaws?

• Yes - had to backtrack a few times to get the full picture
• Used to direct analysis of kernel interactions
• Recorded in spreadsheets initially - converting to JSON

Change management / publication review

• Pete: Need to establish a set of criteria that we apply to review
• Quality and/or safety
• Paul: Planning to start with a lightweight peer review for submissions to GitHub repos, then a more formal review for completed documents
• Barriers to contribution - how to use GitHub, permissions
• Kate: Don’t want this to block progress
  • e.g. If we require all review comments to be addressed, could be open to issues if an uninformed reviewer’s comments blocked merge

Use of outputs of STPA / Safety Analysis approach

• Can help to inform design
• Can be used to derive safety requirements for existing systems
• Could use Linux-based systems in the role of ‘prototype’ to analyse and derive requirements, which would be input to concrete system design
• Paul: I am currently working on something similar - a Linux-based OS as SEooC

Where do we get the requirements that are the foundation of the analysis?

• Can we do this for the Automotive WG use case?

Gab is interested in picking this up from the Safety Arch WG perspective as well.