16

This is a proposal for an entirely free and open source rant like site/app.

devrant today has a couple of problems that I hate:
* Posts in the wrong categories (usually by new users)
* Low effort posts in the "recent" feed
* Good posts in the "algo" feed that are too old
* Longtime bugs
* No official code format in comments, ffs.
* Unimplemented features (like inability to search posts in android, or inability to mute posts in web desktop)
* Lack of admin involvement with the community

but it also has some aspects that I like a lot:
* Admins aren't trigger happy to suspend/ban you
* The avatars are awesome and help to associate users to faces
* The ++ system is good enough
* The community isn't too big so you know pretty much everyone
* There's a lot of variety in the roles and techonologies used by users
* Experienced ranters are usually smart
* Super simple UI
* The comments have only one level (as opposed to reddit comment trees)

This project should try to reimplement the good things while fixing the bad things.

I wrote two posts about a possible manifesto, and an implementation proposal and plan.
https://rantcourse.ddns.net/t/...
https://rantcourse.ddns.net/t/...

I think the ideas outlined there are very aligned to concerns of privacy and freedom users here vouch for.

This project is not meant to **purposefully** replace/kill/make users abandon devrant. People can continue using devrant as much as they want.

I'm hosting a discourse site on a 5$ linode machine to discuss these things. I don't know if it's better than just github.

If you feel that you would like to just use github issues, let me know. I'll create a github org tomorrow, and probably setup gitter for more dynamic discussion.

Comments
  • 0
    I'm also available from 9:00 to 18:00 EST
  • 4
    I have very mixed feelings.

    1) it's not devRant
    2) can you really fix the broken parts of devRant?
  • 0
    @C0D4
    1) agree
    2) yea

    depends on which parts you consider broken though.
  • 5
    @jesustricks broken:

    - low effort posts
    I don't think any community will ever solve that issue, posters will come from other social media platforms where "likes" or "upvotes" are treated almost like a currency, encouraging users to post frequently instead of contributing to the pool of knowledge and frustration.

    - image killer (unofficial name)
    Without the image compressor, you add to the running costs for storage, unless you utilise a 3rd party image hosts or something.

    - memes
    I think this falls into the low effort somehow.

    - no payments
    I get the need to prevent the need for dev financial dependency, but at the same time, you would leave this to the devs free time to generate new features and fixes, which is where devRant is currently at.
    I would seriously accept MORE subscription levels even without extra benefits if it provided additional hours into the app.
  • 4
    Idea: charge a few pennies to post to cover server expenses, prevent spambots, and discourage kids who just post shit memes
  • 1
    He should go @RantSomeWhere about it, banning users from downvoting >_>
  • 0
    @RantSomeWhere well there you go, free downvotes for you...
  • 2
    > Low effort posts in the "recent" feed

    > Good posts in the "algo" feed that are too old

    What do you expect? Good rants aren't growing on trees
  • 3
    @C0D4 That's an easy fix.

    Before posts become live they must first be approved by X amount of ranters.

    User posts. The post is sent to 5 online ranters selected randomly of whom at least three must approve before the post can be made public to the wider community.

    You can even track the reliability of the approver via tracking their actions. If a ranter continually votes 'yes', when others vote 'no' or vice versa the reliability of the ranter is decreased removing them from future post approvals.
  • 5
    @delegate212 this is not a terrible idea...
    Of course approvers should not know the identity of the ranter, and they should be at least 1000+. But I think freedom of speech is one of the most appealing aspect of DevRant and should not be limited.
    So instead of approving or blocking posts, approvers should only chose the category and tags based upon the ranter's suggestions. "Unapproved" or new rants will automatically go under "uncategorized" but will still be visible to everyone unless someone willingly filters out that category.
    Also, I think 5 make a too small subset of the community, maybe 10, 20, or a fix percentace of the whole user base is a better choice.
  • 6
    @michezio That's why I think this whole idea is retarded. Jesus christ let people post. What kind of elitist circlejerk is this
  • 2
    @12bitfloat
    > @michezio That's why I think this whole idea is retarded. Jesus christ let people post. What kind of elitist circlejerk is this

    woah relax, I never said anything about not letting new users posting, so don't associate the ideas those two users posted to the ideas I proposed.

    >> Low effort posts in the "recent" feed
    >
    >> Good posts in the "algo" feed that are too old
    >
    >What do you expect? Good rants aren't growing on trees

    that's true, I mean to say that there's a problem with algo where it recommends posts that sometimes are a month old. There should some intermediate algo, or being able to limit to a period, like last week.
  • 1
    You didnt mention the current hard delete when someone deletes their account. That i find the most annoying thing ever, because it breaks entire discussions.
  • 0
    @C0D4
    > - no payments
    > I get the need to prevent the need for dev financial dependency, but at the same time, you would leave this to the devs free time to generate new features and fixes,

    There are many devs who are willing to offer their time to write new features and fixes, a lot of the times for fun or the need to learn.

    >which is where devRant is currently at.

    I would argue that if devRant was open sourced, a lot of features and fixes would have been implemented by now.
    also devRant is a merely two man operation. While with proper delegation in an open source project, multiple people could take care of their specific expertise.

    > I would seriously accept MORE subscription levels even without extra benefits if it provided additional hours into the app.

    I agree. Although I guess offering the ability to pay more could be unpopular (depends on execution).
    Maybe they could have opened a patreon-like site and would have gotten more funding? but that's a diff business model.
  • 3
    @eval I agree in that's awful UX, to be fair though, it's very respectful of user privacy. That can be further examined.

    There's a couple of other points I forgot to include, like them giving out analytics information to google for free (which ISN'T respectful of user privacy). I think I'll just open up a github org where these things can be updated.
  • 2
    I honestly think users should be automatically banned after a certain threshold of downvotes in the negative range 😂 at least in a progressive state (I.e. 1 day ban, 1 week ban, 1 month ban, etc).

    I mean if you got downvoted to oblivion it’s pretty obvious that you’re a toxic user that does not contribute positively to the community.

    I really like the aspect of auto-modding and having some sort of toxic level threshold. To be honest, I like the way SO let’s the community drive the force behind a karma system, but at the same time it’s easily exploitable as we can all clearly see by the “welcoming” vibe when you ask a “stupid” question.

    I do agree about the effort in adding new features or fixing bugs that do exist. I work as a release coordinator for a FOSS project called Flarum and we solely rely on contributions that may or may not come in regularly and we use a lot of our personal time to dedicate to the project which can become tiring and time consuming.
  • 4
    I created a github org and repo, and a gitter room, and I'll probably deprecate the discourse instance.

    The reason for this is:
    * Easy updating and discussion of the drafts I wrote
    * Nearly everyone has a github account (rather than demanding users to create an account on a separate discourse)
    * Gitter is faster for discussion

    https://github.com/ranterix/drafts
    https://gitter.im/ranterix/drafts

    I'm tagging users that
    * Have over 500 karma or more than 6 months on devRant
    * Gave a ++ to the rant or one of my comments on the rant.
    @dudeking @RantSomeWhere @retnikt @M1sf3t @fishfork @saucyatom @thisizram @erandria @PrivateGER @12bitfloat @Wolle @Superviral
  • 2
    I don't really have a problem with devrant as I've not been here enough to experience them. But I still think it should be open-sourced. Why? Because we have a lot of good developers here and it would help with adding new features (like proper code formatting, maybe markdown support). I'm not keen on subscription (I'm already bleeding with my Spotify one (I don't make a lot of money)) but I would gladly support on patreon.
    As for low quality posts, I don't know how we could remediate that as I spent a lot of times on Reddit, Quora and dev.to and they still have that problems. The problem is that it's very subjective (I like see memes but can't stand seeing them twice).
  • 2
    @skydhash devrant should follow how dev.to open sourced their platform
  • 0
    @volttide dev.to is kinda trash though
  • 1
    @12bitfloat trash or not, well it's about our perspective, but the thing is that they are developer community that build their platform together so they can fix their platform's downsides.
Add Comment