Oddbean new post about login | logout I think we're in the refactoring dulldrums, is all. Never been on a big project that didn't stall out and rework, at some point, because some of the stuff didn't scale well or some of the founders got bored and wandered off. Projects are already deep into maintenance and some experimental types refuse to do maintenance and have nobody in the team to pick up after the Proof of Concept stage. I think we're going to have fewer projects, going forward, but they'll be larger, more seriously run and better-supported. And stuff will be less buggy.
Great observation! 🐶🐾🫂🙏🏻
Don't you think that users also underestimate how much time is spent on system admin and DevOps stuff? Or trying out a new tech stack? As soon as you try to scale something up, without losing performance, development slows down, dramatically, because that's when you have to Do Hard Things and Try Stuff Out and Test Test Test.
99% of developers do not know how things scale or how to scale, therefore making poor design decisions along the way. 🐶🐾🫡
As one of the 99% I take exception to this note because you didn't tell me what my poor design decisions are and how to avoid them.
I personally think Nostr picked all the wrong primitives, but have to admit Json at least makes it easy to adopt.
i'm pleased to say that in my work progress is going well but it is well into refactoring stage now... just had a massive leap forward with bulk iterations of the database, the time for completing a garbage collection count pass escalated to basically infinity after a couple of gigabytes of data but now i have got a counter that runs in 10 seconds on a 15gb store i can see how i might even be able to do reindexing, fast data backup, and in the near future full text indexes and even maybe go the next step to graph database and design a better search API filter searches are just the most minimal required index and search API for nostr, it will have to get a lot more advanced to provide the necessaries such as which you are constantly ruminating about
Yeah, the pareto principle is brutal. Want to get something more than 80% done? Enjoy your 5-10 year long grind! I am currently entering this phase with 15k lines of code, and am just trying to balance protocol support, UX improvements, new features, and bug fixes. It's not as fun as it was at the beginning, but that is absolutely ok, this is what I'm here for.