@marick Those two aspects of debt - borrowing as an intentional act and paying back a defined amount - are what make me avoid the term. I find the term technical debt is often used to launder short-sighted development and management, underresourcing, and basic neglect.
I saw the term used recently in a slide deck describing our project - we are designing a small nuclear reactor to be licensed, built, and operated at a site we own and are actively preparing (i.e. this is very much not a "paper" reactor). We do a lot of work in parallel and our QA process requires we track Open Items and Assumptions Requiring Verification. Stuff like "We used software <X> which has not yet been qualified for use", "we don't know the dimension of this room because it hasn't been designed yet so we estimated how big it might be", or "we don't have a reference for the data and correlations used in this software and can't confirm they're applicable or correct*"
We qualify software and issue calculations with Open Items, each of which are tracked, assigned an owner, given a description, and conditions for closure. We have a formal tracking system and accountability. Between design cycles we burn down the Open Items list by resolving each issue.
In this case I'd say the debt metaphor is accurate and justified. We need to meet a schedule, we are doing work in parallel that in a perfect world would be done sequentially, and we cannot afford to quietly sweep problems under the rug**. Schedule-driven engineering sucks but it's the world we live in so we do our best to do it responsibly. Open Item tracking has been done for decades in big engineering projects like this. We don't have bots that close issues as WONTFIX just because nobody commented on it in a month. We don't have engineers closing OIs to meet metrics. We address each issue or ensure it is no longer relevant before closing it (i.e. we redid the analysis with software <Y> which is qualified for use and revised our calculation). We track our debts and pay them back.
* Tell me again how LLMs and generative AI help me do my job vs actively making everything worse
** I was laid off in 2017 from a division of Westinghouse unrelated to the project that basically lost $6-12 billion and cooked the books to hide it. Now I'm on a similar project and I'll be damned if I let anyone pull that shit near me and sabotage this project.
