2026-09-21 21:29 UTC
Replies (2)
-
@LyallMorrison@cloudisland.nz 2026-09-22 01:51
@grimalkina@mastodon.social For the sake of discussion, I experience it as "we cannot escape the effort of forming a mental model". Obviously the team who created the project has already done that. Other teams, whether they're clients, successor teams, outsourcers who are receiving delivery, or vibe coders who have willed a new project into being, all need to repeat the mental model construction. There's a long history (post Naur) of trying to remove that burden by simply documenting the project "properly". I'm thinking particularly about outsourcing in the early 2000s. It doesn't work though - documentation, or better live handover, assists the building of the new mental model but can't avoid it. Myself I find the paper a bit dry, but I've seen the issue all through my career and props to Naur for getting there first.
-
@joachim@drupal.community 2026-09-22 07:52
@grimalkina@mastodon.social I think both views are right in a way. We need to rebuild the mental model when faced with foreign code. That's why code documentation is so important. And rebuilding it with iterative experimentation is *possible*, but IMO it's very very wrong: code is not science or archaeology, and it should not have to be approached like that. Documentation matters!