TL;DR: I'm not a programmer. I'm a sysadmin with a dream: a distributed forum database with no owner, no single point of failure, client-side filtering (like VLANs for content), and optional Nostr/ZeroNet/Atlas transport.

What exists: database replicated across nodes, each node chooses what to store and what filters to apply. Moderation = subscribing to trusted filters. No crypto required.

What I need: reality check, technical feasibility analysis, database schema advice, prototype devs (PHP/Go/Rust), testers.

What I offer: small budget ($100-200/mo for specific tasks), domain/hosting funding, endless gratitude.

Full article (detailed, 8000+ words, includes philosophy and use cases): https://write.as/zyhlc76pi2op3.md

Please be brutal. I can take it.

cross-posted from: https://lemmy.world/post/45047387

Title...

I'm kinda disgusted with Microsoft and Github has been declining into an AI-Centric hellhole, to the point my recommendations are almost exclusively AI related... And let's not forget, the new Copilot Training enabled by default (which honestly, how do you get rid of this thing, VSCode also feels intrusive with AI-First bullshittery)

I've been wondering about moving to Gitlab but.... "Finally, AI for the entire software lifecycle." is literally plastered in the landing page. So.. that feels like a no-go.

Codeberg is very decent, it's based on Forgejo so ActivityPub is also a thing (but is cross-instance contributions possible?) but it's exclusive for Source-Available and Free Projects, which, by all means, totally fine! Half of my "active" projects are for free, and are open source (does that make them FOSS even though I'm basically the only dev?)

And last but not least, Forgejo and Gitlab themselves are self-hostable, but...how expensive (price and storage) would it be to self host a Git Forge??

And maybe I'm being narrow-sighted... For FOSS projects in Github, sadly I'll have no choice but to contribute there, if that's the only place where the project resides, same for Gitlab, and Codeberg* (unless cross-instance contrib is a thing)

For now, I'm thinking of moving FOSS/OSS projects to Codeberg, but for personal projects? What are some good options?

cross-posted from: https://lemmy.ml/post/45214693

Also vm repo

I have a small Yocto Linux-based embedded device with a main application - a kind of scientific measurement system - written in C++.

I want to add a WebUI to it which shows status / errors, a few graphs, and some basic controls (like restart). That is a relatively new task for me.

What do you think are good / creative ways of making such a web UI?

C++ feels like the wrong tool here.

What I am thinking of is something like:

• storing some data in LMDB or SQLite
• storing status and controls from the main process (written in C++) in a mmap'ed shared memory file
• writing the UI in a lightweight script language which can read the file, and generate the output. Python is likely too heavy, and to difficult to build. What is possible is micropython, possibly also Rust. What I'd like most would be to use a program written in Guile, which has nice facilities to server web requests.

What ideas do you have? Which extra points might I want to consider?

The 6.11 release for Qt Framework is now available, with improved performance, newly supported techniques and capabilities on graphics, connectivity and languages, not to mention a whole new approach to asynchronous C++ coding.

• Hardware-Accelerated 2D Rendering: A new module, Qt Canvas Painter, based on the HTML Canvas 2D Context, provides performance & productivity gains.
• 3D Improvements: New rendering techniques Screen Space Global Illumination (SSGI) as an option for lightmap baking, and Screen Space Reflections (SSR). Also imrovements on the Temporal Anti-aliasing algorithm with motion vectors. New user-defined render passes for post-processing effects, color picking, layer masks, etc. directly in QML.
• Interactive Graphs: You can now implement custom graphs where a user-defined delegate renders each data point. There's a new Qt example, the Wind Turbine Dashboard, and many improvements, e.g. new ways to style line graphs, and multi-axis support on 3D graphs.
• Declarative Approach to C++: Qt Task Tree brings a whole new approach to asynchronous coding and C++ API design in Qt. In addition, various APIs have been unified to allow adapting any asynchronous task to work with the new module.
• Other Improvements: Improvements on vector graphics, controls, and accessibility. Connecting to web servicers is now easier with the new module, Qt OpenAPI. Navigating in an IDE between QML and C++, and making data available from C++ backend code to Qt Quick have gotten easier. A wealth of other improvements, such as for multimedia, Android, and API documentation.

Not sure if it’s just me, but I’ve been running into this quite a bit.

My client conversations are spread across different messaging platforms, and sometimes important or more detailed discussions just get buried or overlooked.

It’s not even about the number of messages, it’s the fragmentation that makes it hard to keep track of everything in one flow.

Anyone else dealing with this? How are you keeping track of conversations without things slipping through?