SourceCred

Trails, the Memex, and SourceCred

So I just posted yet another long ass comment. I then went on a walk and thought about it, can came back to add more details. I feel really good about it and think I had some really great things to say. If I could I would boost my own post! Wait… I can’t. Actually no one can. My “comment” is actually like 8-10 replies to various other comments in the thread. If someone “likes” the post how do I know which part they are liking? What if someone agrees with part of the comment, but disagrees with another? How do they show that in a way that can be transparently seen by the community and the SourceCred alg?

This made me think of the Memex. The idea was to curate “trails” through the archives of scientific databases so that researchers could credit and promote all the work that was part of their research for a project. This would be much more inclusive and dynamic than simple adding references at the end of each paper. They came up with this idea just after WW2 because there was too much damn data! Now look at us…

It would be really cool if there was a way to “tag” various comments (or parts of a comment) with topics like governance, economics, CredSperiment, etc… so that people could create, view, and contribute to trails of ideas. This could start on a chat forum like Discourse, but then expand to include code from repos, articles, live projects, etc… Then you could boost an entire trail vs just a post or comment.

If you’re thinking that this all sounds like a lot of work, you’re right. To address this the Memex (at the end of the post) also has the idea of historians and curators. It would be cool to see something like this integrated into SoureCred. People could blaze their own trails, but you could also contribute to others as well.

  • The more people use a trail the more cred flows through it.
  • If you want to support that trail you could boost it to boost everyone/everything that has contributed to it (where contribution = creating the trail or the content in the trail) would get cred.
  • (in an ideal world where data is immutable and verifiable and rules are enforceable) If someone does something awesome (breakthrough research! killer app! commercial product!), and it can clearly be proven that it depends on on stuff in a trail, then cred would flow back through that trail
1 Like

You can (and did) by making a new top-level post that points to it.

As SourceCred gets bigger (and gets more developer resources), I’d like to experiment with better UIs for cred. For example, being able to reference or like individual pieces of a post, or having more types of reactions than just :heart:. We already have some of this metadata because we can see which parts of a post were quoted in its replies.

On trails: I like the idea a lot. The challenge, as you note, is coming up with a UI. Right now, we don’t have the eng bandwidth to implement new UI paradigms, although since Discourse is open-source and supports plugins, it may be doable in the medium-term. However, we still have a lot of affordances with the Discourse UI as it exists, so we can get creative.

For example, I like the idea of having “summary” or “distillation” type posts. These posts would basically take a complicated thread that generated lots of ideas, and then distill out the most interesting concepts, in a way that allows a reader to get them without needing to slog through the whole thread. These would be new top-level posts, and would of course have citations back to the posts they draw content from. I think it would add a lot of value (thus meriting cred for the author) and would also serve to flow cred back to the (possibly obscure) posts where new ideas were put forth.

Concretely, I’d be happy to see a distillation of the original comment you linked, perhaps in the newly created governance category!

1 Like

I like this idea. Already, when a project puts together a newsletter, for instance, there is inherent filtering and ranking. If that newsletter is a Discourse post linking to a bunch of other posts, that’s a nice way to flow cred to them.

Oh that’s great! Didn’t do that on purpose, but I’m not exactly bummed out either lol

That would be so cool! Is there a product/fundraising roadmap/plan for this?

Huh… this would be a really interesting design challenge. My first thought is something like Medium/GoogleDocs where you can highlight stuff and leave comments, but I’m sure there’s more elegant solutions too… Will have a good think on this! Out of curiosity… is the bandwidth problem one of engineering or financing?

This would really improve the UX for community members a lot. Creating a TL;DR: of a long ass post is fairly intuitive, but making the citations user friendly might be challenging. ATM is there a way to link to a part of a comment rather than the whole thing? If not, then I guess we’d just link to the original post as a whole - or maybe create a summary of a “trail,” essentially summarizing a concept that spans multiple threads/comments. I dunno, we’ll figure it out, but I like the idea of higher and higher layers of abstraction on comments, distilling them down into their most essential components and then branching that out into all the contributing conversations and ideas.

I’m planning to write a roadmap that details a bunch of potential priorities, and then use cred-weighted voting to set the priorities. Comment on the plan here.

It’s engineering (and product+project management).

@protocol labs is willing to provide a bunch more financing to SourceCred. Internally we discussed just hiring more people to work on it, but I’d rather try to grow SourceCred via dogfooding through SourceCred itself, rather than having a separate company hire people. I don’t want any one company (even one as value-aligned and friendly as PL) to have undue control over the project, and it’s hard to avoid that control if a company is paying salaries for all the core contributors.

I’m hoping that the CredSperiment will provide a way to translate more financial resources into more engineering bandwidth without going through hiring. It’s definitely an experiment though; we’ll have to wait and see how well it works. :slight_smile:

1 Like

Yes! Very exciting. Much fun.

Incredibly interested in this. The entire goal of 1Hive is to help create tools and frameworks that make contributing to the commons and open source easier and more sustainable. We do that via Aragon DAOs and related tooling, but the challenge of making bottom-up community driven projects sustainable and scalable (ideally to a global level) is quite the pickle.

While SourceCred rewards and recognizes past contributions, how you you incentivize future contributions? For example: bounties are stakes to incentivize/reward specific actions in the future. Does SourceCred have a similar type of mechanism?

You beat me to it! I’m still catching up on unread posts, but it looks like there’s a bounties discussion. Great :slight_smile:

1 Like