 https://spectrum.ieee.org/bionic-eye-obsolete I"m convinced that any software deployed for medical reasons - implants, tools, analytics, doesn't matter - needs toolchains that generate reproducible builds stored in escrow that are made free and public the moment device support ends. 7 comments
 I'm convinced it needs to be open source in the first place. Because any software deployed on devices placed in human bodies or otherwise used in ways that affect human health needs to be reviewed by as many experts as possible. @jackwilliambell That is not the FDA-Way, but you are correct in the way that implanted devices need to be open sourced latest when they reach EOL or the manufacturer ceased to exist, at least to give patients a chance to get support from a third party.@mhoye I'm saying it should be open source – and reviewable by third parties – as of the time it is first used in human trials. We need to start treating software that affects human health and lives the same way we treat drugs and medical devices. Software involved in any high risk situation should be: 1. Open source and reviewable 2. Certified by either the government or an NGO funded by insurance companies (think 'UL' certification) We do these things with drugs and hardware now. @mhoye there's plenty of non-profit orgs in the pharmaceutical space, it feels like they should be joined by one or more implant specialists. And part of any approval should be ensuring that such an org has all the information required to maintain those devices for the lifetime of the patient. @mhoye (including the OS, lest someone upload a toolchain that works perfectly on one specific build of Windows XP and is useless elsewhere) @mhoye Exactly! SAAS-centered, licensed IP approaches are wholly inappropriate for embodied technology. |
Gregory's Server
@mhoye 💯