Good point. I've created a kbinMeta magazine for topics related to the platform. I'll add some important information there later. Reputation points are a simple counter that boosts your posts and comments. https://kbin.social/m/kbinMeta
I actually tried to go to /m/kbin, and since it wasn't there, I created it. Can I redirect it somehow, or should I add moderators from the dev team (you, for example)?
That's right. The entire traffic on the website generates around ~2400 people whom I follow along with their reactions and followed communities. Kbin allows to track both (+ specific domains). I'm not a big fan of relays; I prefer when an instance grows naturally and is built by people. Relays will probably be implemented into /kbin someday, but first, better moderation tools are needed.
and I've noticed that there is quite a lot of activity of those hashtags from general mastodon instances.
Yes, that's a feature of /kbin - based on the magazine's name, posts are assigned to the microblog (soon there will be an option to disable this feature ;)). You can also add additional related tags in the magazine panel that you would like to gather.
based on the magazine’s name, posts are assigned to the microblog (soon there will be an option to disable this feature
Can’t wait to play around with kbin!
I’m not a big fan of relays; I prefer when an instance grows naturally and is built by people
💯 Great explanation on why one should avoid using relays over here.
Snippets:
99% of the content are things you don’t want to see. They may be posts in a foreign language, advertisements, pornography, bot posting, and spam.
Over time, your database size will grow and grow. Before you know it, you’ve stored 200 GB of Japanese shitposts and furry porn for no apparent reason, and your community is no better off.
What I was really asking there was about how you have so much non kbin/lemmy content to pull in to those magazines. I understand how the magazine filters them once they federate with kbin.social, but I wasn't quite sure how you got so much content to you in the first place given that most kbin.social users will be following group actors rather than individuals
Currently, the "Add New" link works contextually. If you are in a specific magazine and click on it, the magazine will be automatically set in the form. Maybe @cody will have a better idea on how to make it even more intuitive.
Yes, you can use it to mark content (links, articles, photos) that you are the author of. After completing the guides, I will add them to the FAQ section.
The preview was done quickly and without thinking, so it's not the most logical and beautiful. But it is, and it seems to me that it illustrates how the component should behave.
It's more of a layout issue, really. Add article is in an odd place - I'd expect it to be on the right side infobox, or maybe above the articles. If I didn't know from earlier experimentation how it worked, I would never post articles.
Another issue, in the same vein, is that in the magazines articles are under 'Threads,' but they're called articles. There's pretty much nothing associating "Add new article" to 'Threads'. Again, something that you'll have to just try out.
Other topic altogether. Why is there an empty space at the beginning of edit area?
This item is partially designed. In addition to the "general" button in the navbar, in the general list and microblog as well as in the content belonging to the selected channel, a form is visible at the top. When clicked, it should go to the full, expanded form. You can also do as on #reddit, i.e. move to a new tab after clicking, but in my opinion this is a bad solution and unintuitive.
I think the /new form itself is fine as implemented today. What's missing in the current implementation – but present in this comp – is the "Add content" button at the top of the right rail. I'd expect that to be present when I'm logged in and looking at a specific magazine, and I'd expect it to link to /m/<magazine slug>/new
A similar button is present as part of the instance-wide intro box that atop the right rail in a logged-out view, but it disappears once I log in and there's no equivalent button in the magazine section.
I also think this comp's stronger treatment of the "+ Add" button in the global nav bar helps by making the button stand out a little more, but it's still grouped on the right side of the global nav. The problem is that everything else in that area of the nav is a global, instance-wide action, so intuitively it's not the place we look for a magazine-specific action like adding a new article.
It may happen if there was an update, then for security reasons, the session is invalidated. I will try to ensure that it happens as rarely as possible, or alternatively, I will change the storage, but it will take some time :)
This is something I'm currently experimenting with. I understand your point of view, but boosts are the only way to increase the range in the fediverse, and they are extremely important. I am definitely open to discussing this topic. I will come back to it later.
I'm left to wonder if having both upvotes and boosts might be seen as novel and interesting to the Reddit crowd. Something that makes the fedi groupoverse feel like Reddit++.
I get this too, having up and downvotes is like "speaking the language" of Reddit and would enhance adoption through familiar UX. That's a big positive for something that aims to serve as a Reddit replacement.
At the same time, it's a headache to translate into ActivityPub, because as OP said, upvoting and boosting are very different concepts. They both aim to increase visibility, but an upvote differs from a boost in that 1) an upvote produces a gradual increase in visibility; 2) you can opt to ignore vote counts entirely, for example, by sorting by new posts; and 3) the source of an upvote is (generally) not public, whereas you're always told which user has boosted a post.
Maybe Lemmy/Kbin/both could implement a system with upvotes working as in Lemmy, but also the option to boost a post or comment? I honestly think that'd be the best of both worlds.
Instead of just upvotes/downvotes, it'd be upvotes/downvotes/boosts
Thanks, this is a really helpful roundup of links – a lot to consider.
This comment is especially interesting. It makes me think kbin should stop translating "upvotes" and "downvotes" into Fediverse concepts, and should instead embrace the Fediverse concepts and present them directly: instead of upvote and downvote buttons we could have 🚀 (boost) and ⭐️ (favorite) buttons, perhaps with a combined score for sorting. It would mean abandoning downvotes, but it sounds like downvotes are a problem anyway.
That would give us a reputation system that makes it clear what you're doing when you boost a post or comment, but also one that won't be jarring to users of other apps and instances.
I've been thinking about it lately, and I even wanted to change it quickly, but it turns out it's not easy, and I wouldn't want to make such a big change hastily. It will have to wait a bit because I have a queue of critical tasks, and more external matters are diverting my attention. I've also decided to make the placement of the boost/fav entirely configurable by the instance admin.
Honestly, I love the current state, I'm very used to it, and I would like to keep it. At least on my smaller instance. I also have plans for a short interactive tutorial that will briefly explain how the most important features on the website work.
However, I completely understand the arguments that some change is needed. But making it configurable won't take much more time than making it fixed. Downvotes won't be federated.
I do like the idea of both likes and boosts adding to the score (but not double) which I think (I used translataion) the last comment in the codeberg issue is making.
And even at the risk of making things more complicated I wonder if it would make sense to have separate logic for threads and replies. I think for threads, upvoting to increase relevance/have it be shown to more people in the community makes way more sense, while within the conversation, I feel it correlates much more strongly to simpe approval/disapproval. And if you already are IN the thread thread, you probably want to see upvoted comments, but if you aren't, getting random comments from random convos boosted into your timeline is just confusing.
So I think there's some arguments for separating thread/comment logic, which also opens up new avenues for UI.
Main threads could even get rid of the upvotes/downvote buttons and have one single button "increase reach" button that corelates with fedi-boosting and also is used as the scoring mechanism.
Replies could have the classic up and downvotes with the upvote correlating to likes.
And I guess both main threads and replies could have a separate smaller button that correlates to likes for main threads and boosts for replies.
We expect all users to treat each other with respect and kindness. Harassment, hate speech, or any other form of harmful behavior will not be tolerated. We reserve the right to remove any content or user that violates these guidelines.
edit: problem is I'm not sure how to search for old discussions/posts to import them here. I tried with a few posts in the last few days but it didn't work like it does on masto/misskey. Searching the user sometimes seems to work.
The third result on that page is @linux@lemmy.ml – do you not see the same?
I should be able to merge instances for viewing
As long as it's optional and you're able to choose which ones you merge.
I totally get how that could be useful, but it's not something I'd want. There are federated instances that I mostly avoid, like lemmy.ml, which which welcomes and encourages CCP propaganda and disinformation.
@kjr I mean that the Hebrew language will be added to this list. Marking a post with a language allows filtering on Mastodon and other supporting platforms later on. kbin will soon also allow filtering content by language. I noticed that there are a few people on the instance who speak Hebrew :)
@ernest
מעולה, שמח לשמוע !!
It sounds excellent!!
Then (a little unrelated here maybe) makes it sense to add additional languages in the translations?
And is possible to publish here in other languages than English, using the language mark? (I thought it was an English-only instance)
@kjr That was the assumption. But hey... why not, the fediverse is so diverse that a lot of content in other languages ends up here anyway. Since there's an option for tagging, feel free to use it. I'll do what I can to ensure the filtering option is available as soon as possible.
@ernest@kjr A quick question: Can individual users select which language to display the UI in? I can't seem to find a way to change the UI language for kbin.
I see that the server .env settings has a default language setting, but how can a client change that to show a different UI translation? Is this possible or does the admin have to change the entire instance's language to a different one for all clients/users?
Currently, the language is automatically selected based on the system/browser settings. However, this will change after the mentioned update is released. There will be an option to choose the language in the footer.
@ernest Hmm. Ok! I must be doing something wrong then since I couldn't get the Japanese language translation to show. I downloaded the .yaml and put it with the other translations, but maybe I'm missing a piece.
I use a browser plugin that overrides headers for testing purposes. On my local kbin installation, I set the Accept-Language header to "ja," and it seems to be working ;)
@ernest Do I have to re-build anything? My header is also set to ja, but I must be missing something. Does the language need to be added to a list somewhere as well as the .yaml file?
I simply uploaded the translations to the folder. APP_ENV=dev. You can try clearing the cache by running the following command: docker-compose exec php bin/console cache:clear
@dannekrose Wow, I see that you managed to set up your own instance. Are you familiar with PHP/Symfony? How would you rate the difficulty of this process?
I am not very familiar with PHP nor Symfony. I set up a PixelFed instance from scratch as well and that uses PHP Larval and that experience did help some, but I'm not a PHP developer at all.
As for the experience, after I got everything set up, I can see where the lack of documentation caused me to get stuck, but if I were to try again right now while using my experience as a guide, I can imagine a "relatively" straightforward set of steps to set up.
The only piece that I'm still not 100% clear on are what are the optimal configuration values for the JWT tokens (.env only mentions one token, but mercure's Caddyfile mentions both a publisher and subscriber), and which daemons should be run under which accounts. There are a lot of pieces:
php8.2-fpm
nginx
postgresql
mercure
redis
...etc
Also, setting up the worker threads via supervisord was unexpected and only mentioned in that other issue raised on codeberg.
I've done my share of documentation in the past and have thought about editing the wiki, but due to work at the moment I didn't have the time to record all my steps down in detail for this deployment.
I do hope to have some time soon, though and might try to take a look at the wiki and do a "test" deploy again.
My personal preference, though, is to not use Docker so I also understand that can add a LOT of complications.
I'm happy to share anything specific if you like as well about the experience. I used it as a learning experience and ... I did learn a lot for sure.
As for difficulty - with the state of the documentation as it was when I tried, it was quite difficult. However, once you have all the correct pieces and software parts installed, the configuration isn't any more complicated I think that other platforms (Mobilizon, Pixelfed, etc). There are a lot of pieces to get running correctly, though and any documentation should probably focus on making the installation of those different pieces as easy as possible as well.
On a scale from 1-10 with 10 being basically imposssible, I'd say without documentation completed - 8
With documentation completed - 4?
These assume at least a basic level of comfort working with Linux and installing software like this, though.
@ernest I'll give that a try! If it works, I can then tune the translation better being able to see where the different elements are actually used! Thank you!
/kbin meta
Oldest