Many users agreed spec-driven development fails for most projects because software is iterative and discovery-based rather than blueprint-driven, while others strongly rejected the claim as misguided or harmful to creativity.
Based on 21 visible X reactions from 139 accounts; directional sample.
Ask a question below.
Published answers will appear here.
I spec'ed this all 18 months ago. Specs are easy when you think only about people. That's very difficult for software engineers, they don't see people much. (I have a degree in psychology.) The only big feature I haven't delivered is a killer on that occurred to me when I nailed my security. The other one is the big money piece and I knew I needed everything else first. Everything else is practically done so I had Claude make me a PRD today to start my spec process, which will not be done by AI. http://Dartastic.io
@stevekrouse Completely agree. Software is fundamentally an iterative loop. If you try to write a massive, perfect spec upfront for a brand new product, you’re just guessing in the dark.
@stevekrouse Strong disagree. Agile was a response to slow dev cycles. Now that dev is fast, water is back to falling. https://x.com/jumploops/status/2076541216133730795/photo/1
@stevekrouse 💯
Too many people stop halfway. Ok so you have spec driven development. But building something is iterative and you tweak as you go along? You can still make those changes. Or rearchitect the whole thing on the fly. The cost of technical debt is 1/100 what it used to be. https://twitter.com/stevekrouse/status/2076383688753832382
I spec'ed this all 18 months ago. Specs are easy when you think only about people. That's very difficult for software engineers, they don't see people much. (I have a degree in psychology.) The only big feature I haven't delivered is a killer on that occurred to me when I nailed my security. The other one is the big money piece and I knew I needed everything else first. Everything else is practically done so I had Claude make me a PRD today to start my spec process, which will not be done by AI. http://Dartastic.io
@stevekrouse Completely agree. Software is fundamentally an iterative loop. If you try to write a massive, perfect spec upfront for a brand new product, you’re just guessing in the dark.
@stevekrouse Strong disagree. Agile was a response to slow dev cycles. Now that dev is fast, water is back to falling. https://x.com/jumploops/status/2076541216133730795/photo/1
Too many people stop halfway. Ok so you have spec driven development. But building something is iterative and you tweak as you go along? You can still make those changes. Or rearchitect the whole thing on the fly. The cost of technical debt is 1/100 what it used to be. https://twitter.com/stevekrouse/status/2076383688753832382
Many users agreed spec-driven development fails for most projects because software is iterative and discovery-based rather than blueprint-driven, while others strongly rejected the claim as misguided or harmful to creativity.
Based on 21 visible X reactions from 139 accounts; directional sample.
Ask a question below.
Published answers will appear here.