A certain stage of development has ended - a prototype was created that allowed me to verify whether such a form makes sense in the #fediverse. Until now, it was a project that I developed in my free time. Now, thanks to the great people from Nlnet, I will be able to devote much more time to it. This time, we have already outlined a plan for the next few months officially: https://nlnet.nl/project/Kbin/
The progress of the work can be followed here: https://dev.karab.in - this is the version set up on the cheapest server, and it will remain so until all stages are completed, because one of the goals is to create a lightweight platform that will have very low requirements in its basic version, and therefore low maintenance costs for instances. With each passing day, more subpages and features will work.
Many of you are working on cool projects, it's worth calling on the Open call for funding ;-) https://nlnet.nl/news/2023/20230201-call.html
Soon, the kbin.social instance will also officially launch, now work on the server is still underway.
Stage 1 Frontend - refreshed, modern and accessible layout
To save time, the /kbin prototype was created using tools such as Bootstrap. Due to frequent iterations, good interface development practices were often overlooked. The goal is to create a clean, modern, responsive and accessible layout based on feedback from the community, which will work on any device, also without the need to use JS.
Stage 2 Backend - upgrade
Updating the environment, PHP, Symfony, API-Platform, etc.
Stage 3 Search module and alpha release
Currently, the search engine is based on Elasticsearch, which significantly increases the minimum server requirements. The goal is to eliminate Elasticsearch and create a search module with the ability of contextual filters and prepare the first alpha release.
Stage 4 Backend - refactor existing services.
Update of frameworks and tools used in the project. Refactor existing services. Refinement of existing services in order to achieve efficiency. Writing documentation for other developers, admins and visitors. Preparing a repository for contributors.
Stage 5 Federation improvements
Optimization of communication using ActivityPub. Completion of integration, creation of documentation for fediverse developers.
Stage 6 Admin section and user UX improvements
The goal is to provide tools for Kbin instance admins and to improve instance community sections.
Stage 7 Prepare for stable release
The goal is to release a stable version of the platform, solve problems reported by users and, using appropriate tools, develop the highest scalability and efficiency.
Stage 8 Implementing Events (incl. ‘stretch goals’)
At this stage, kbin will also become an events aggregator.
Stage 9 ActivityPub federation of Events (incl. ‘stretch goals’)
The events module will communicate with other platforms via ActivityPub (e.g. Mobilizon)
Currently, I am at stage 1/2/3.
- Project site
- Repository && Issues
- Weblate translations
- Devlog in Polish
- Devlog in English (soon)
- Kbin on NLnet
Other instances:
I went to see Alexis Ohanian talk back when reddit was still small. He said they were doing it for fun, and meant it to be a small fun website. If memory serves, he called it tacky. I didn't like reddit back then, but over the years a big chunk of answers that were supposed to be on stackoverflow would show up on reddit. Google on the other hand started messing with search results to serve more ads, so appending reddit at the end of a query helped in many cases.
Making a better aggregator / search engine for the fediverse would be interesting. But its utility for the current content of microblogging is negligible. I wouldn't mind cutting out mastodon to be honest. It's just a bunch [british word for cigarettes] and hentai watching nazis arguing about which group is more retarded. It's entertaining, but it gets boring pretty much. You're guaranteed to offend both of those sides, so it's better to avoid the whole microblogging sphere. I guess Poland is a magical place where far right politicians go to anime conventions and pose for pictures with femboy attendants.
LBRY is more of a protocol like ActivityPub, but it's also a cryptocurrency. You can click download and get the file (unlike YouTube). It can distribute content in a bittorrent fashion, and that's a lot more scalable in the context of the fediverse. You need to ensure that the project gets properly funded, or it's not gonna be used. I understand your caution about crypto, but that could be a viable source of funding for this. You can buy compute with crypto, so it should cover most of your expenses.
I'd recommend you working on it alone, on the core at least. You don't need help if you're good at using LLMs. They can write most of the code for you, and even tests. Software teams of humans are slow and fragile. Single dev projects are a lot faster. Obsession moves mountains :)
I don't care much about open internet. I think it's open enough already. And most people want moderation (censorship). There is too much stuff out there, and most people are offended by a big chunk of it. It seems like the same people who want moderation also don't want the fediverse to searchable. Is that right?
I get your point of view :D But I still see it a bit differently. Maybe because from the beginning, I started consuming content from the fediverse my own way, without relying on any of the apps. And it was actually mainly Polish-language content, where it was a fresh and enjoyable experience. Even now, on the main instance, there are a lot of interesting discussions and content that I wouldn't want to miss. I think building small, local communities is incredibly important. And in such a magical place, it's good to have a little bit of independence ;)
It's true that there is a lot of content on the publicly available internet, and among all the noise, it's harder to catch the valuable content than ever before.
What you're talking about, as you've noticed, is an idea for a separate project. But you're absolutely right, there are also real-world problems and you need to have different options in reserve if you want to keep developing continuously. I'm taking note of all of this, but one thing I've learned is to bring things to the end and focus on the current goal right now.