@neuroexception I find that you need to realize that some day your implementation is going to be deprecated, and you should find solace in that!
Your objective right now is to implement your roles as correct as is possible with your current resources (and no more than that!)
Having some other "more-incorrect" implementation remain the dominant one *is a failure case* just because your perfectionism prevented you from releasing your solution.
correct me if i'm wrong, but that inclination may be the result of the absence in #ansible of a good structure to provide reusability. there's a lot of friction in reusing ansible roles and building on each other's work.
the best effort i've seen towards structure and some kind of standardization is debops (https://debops.org/)
@zeh @neuroexception Perhaps to some extent, but at some point the entire stack your implementation is built upon will be "the old way of doing it". I find this very useful to limit how "perfect" I feel I need to be in my implementation.
That said, that's my experience as a perfectionist.
@jmarion
right, i agree. and when confronted with those decisions of "good enough" vs "naack let's code for these use cases too" we get skewed because we know it is hard to iterate on what we have now, that it is hard for others to pick up this code we have and make it better. so we go for another round of grind instead
@neuroexception
@zeh @neuroexception I feel like keeping "what we have now" as minimal as possible helps in allowing yourself to pivot to an unknown direction in the future.
In addition, being able to help others get up to speed with your problem in the future should require as little as possible in order to drive momentum in your effort.
@jmarion @zeh @neuroexception The good old "do one thing and do it well" instead of hammering everything in just one role 😁
@marco_m_aus_f @zeh @neuroexception The *nix philosophies will resonate for generations in computing. We owe all of what we have today to those early pioneers 🙏
