Welcome to my little experiment: Regenerative Software
Hi everyone,
Today has been a special day for me. I finally got to share a little experiment I’ve been working on for the past few months.
That experiment is modelARch. But behind the product are two connected ideas I’d like to explore with you here: Evolutionary Models and Regenerative Software.
An Evolutionary Model is a living, structured representation of a product’s intent that serves as the authoritative description of what the system should be. It evolves together with the product, its users and its environment.
Regenerative Software is the software that emerges from that model. Rather than evolving the implementation directly, we evolve the model and deterministically regenerate, validate and deploy the implementation from it.
Put simply: The model evolves. The software regenerates.
This forum is part of the experiment. I’ll use it to share what I’m learning as I build (ideas, decisions, doubts and failures) and to exchange perspectives with the community.
My hope is that, through these conversations, Regenerative Software can gradually grow from an early hypothesis into a clearer and more robust paradigm.
Replies
the bit i keep turning over is that regenerative only stays true while regenerating is cheap and frequent. the moment a team regenerates rarely, the model stops being authoritative and quietly becomes documentation, and then you are maintaining both it and the code instead of one thing
so the number i would watch across your users is time since last regenerate, per project. if that starts growing the paradigm is failing for that team, whatever the generation quality is. it is also the earliest signal you will get, long before anyone writes in to say so
and the thing that stretches that number is the emergency patch nobody folded back into the model. once an unmerged patch exists in the tree, regenerating becomes frightening, so people put it off, and the gap widens on its own
@jernej_jan_kocica Apologies for the late reply! I genuinely hadn’t seen this message.
Coincidentally, I was planning to open a new thread tomorrow specifically about these operational scenarios and the questions that both you and @abdullah_javaid3 raised: emergency changes, regeneration, migrations, external services, scaling, and rollback.
I think these questions deserve a more precise answer than I could fit into a nested reply, so I’ll address them properly there and tag you both.
Thanks again for pushing on this. It helped me see where the paradigm needs to be explained much more clearly.