Multiple Queues Impact and Feature Mapping

a. You don’t have to “wait for me to paste”, I just ask claude to read the feedback directly if there’s anything actionable, and do it. I have a feeling you have unbased opinions against the tool, maybe you have not tried using it yourself, but things are not that complicated, plus that is on me.

b. You can very cheaply ($0) add a tool like Qodo or Code Rabbit on Github to check the code first and only pass it on for human review after I’ve ironed out most of the kinks.

c. “Or I have to read pages and pages of AI slop plans” - I don’t see why you should, if I won’t be committing them in my PR, unless you ask me to. I too hate reading all the verbiage (which has a very good purpose, it’s called Spec Driven Design, and it was invented years before LLM coding existed), but it’s on me as the developer to go over it before I deliver you a PR because I read some Java but I speak English better, so if I can program well enough in English, the tool will be able to “compile it to Java” more accurately.

d. As for UX design - don’t understand anything about it, I do distributed cloud design, batched jobs queued for workers on queue managers, microservices, things very different from frontend, aka UI. for me “UX” means fewer clicks and a good UI design where I see just enough info at a glance but not too much (I have ADHD and hate clutter, too much data and a bad mix of colours and fonts just hurts my brain), and above all UX means as much logic as possible is hidden from the eye, behaviour is configurable, but comes with sensible defaults. All that I have in my fork.

e. You are right, I didn’t start from the UX design. I had a general good idea about it from the get-go since I used such a system once or twice before, but since it’s my fortè, I started from the backend, designed the migration to enhance the existing tables and made new tables where it made sense. I sat for a long time interviewing Claude about the existing infrastructure and class hierarchy before I directed it where to change what, or it would go rouge and duplicate work instead of reusing existing code. So I did my home work, it was not “Vibe coding” slop, it was meticulous and intentional. It was just written in English rather than Java or Kotlin or whatnot.

If you guys want, I’ll just record you a session on the current fork I have and you can get the hang of my implementation, even if I don’t speak the lingo of a UX designer, I know how to make my usage experience more convenient to operate, aka better user experience for my not very low standards.

In addition, as I have offered before, I can chop it down to several stages, backend/storage functions separate from UI implementation, that way you can keep the backend improvements but someone else can implement the UI that uses them. Mine would be the POC and someone else will take that and implement a different UX over my backend work. sounds better? I’ll do it your way, just tell me how to better coordinate and cooperate, that’s all I ask.

One more thing I can donate, if you want, and you are overwhelmed with AI PRs, is the scaffolding for agent coding. that is, look at the current style and project design, and not just document it, but write it out as sets of rules and “skills” for coding agents that will cause those PRs to be written more to your liking in the future. it will tell them how to be your team mates and save you time and bother by generating code that is more to your standards.

Yes, I’m well aware of the problem, I work in engineering myself and maintain a few FOSS projects as well, and I am painfully aware there’s a paradigm shift and the load moved from developing to reviewing. it’s painful for me because I have 35 years of career under my belt and I feel like a boomer every time I complain I miss the way we did things in the 90s (I don’t really. I hated CVS and remember when subversion became the new breath of fresh air).

So I’m only saying, i feel you, I want to cooperate, just give me a hint where you prefer I start and how.

Sorry to hear that, I can’t bring back those contributors, what I can do is give it a shot. I’ll draft a proper QA about what I gathered of information on the GH repo and here on the forum, stablish a baseline for us to move forward with the main definitions for the feature, I feel like what @igor did and what I see from @seefood as well work to our advantage on this and hopefully won’t take long for us to revisit the issue with these docs in hand.