all comments

[]go2dfish -3 points

Thanks for the feedback. I’m considering alternative approaches to the way collapsing currently works because I’m not happy with it either. We don’t have enough people signing in to make comment whitelisting viable.

[] -0 points

you would have more people signing in if you didn't tell them that their comments would be visible, but without mentioning that it takes someone else to approve their posts, minutes to upvote anything that isn't a one line comment, and even then you can get downvoted back into the spam.

[]go2dfish -0 points

I'm having trouble understanding exactly what you are trying to say here, but I think you are saying there should be a more descriptive explanation of how curation works and possibly adjust the collapse thresholds?

[] -0 points

As far as I can see, indexers are naturally evolving into spaces, this is good because the cpu power required to handle them isn't free, someone has to pay for it. Coming from that point of view, I'd say yes, default collapsed thresholds should be managed by indexers, instead of the's ui. You shouldn't have to direct noobs to a specific instance on the notabug network just to get a good ui default. It's a bit of a different story with indexers as their users can just hop from instance to instance without any headaches.

But that's not the point I was making there, I wanted to put this into a comment, but the post ended up growing a lot bigger than I originally expected:

[]Anonymous -0 points

It sounds like he’s saying that you’re micromanaging.

I just created an account but I don’t see any reason to stay because it’s a ghost town.

[] -0 points

That's not my intention; we had a user show up and aggressively (and manually) spam random useless content to an extent it was making the site unusual so I and others came up with an approach to mitigate it without rate limiting contributors; but it works best when at least a few people are willing to log in and comment on submissions that aren't totally worthless/malicious in nature.

The approach we took (for posts) only affects t/front and not t/all

The comments threshold is a bit more annoying and I hope to have an improved to this problem this weekend.

[]go2dfish -0 points


[]Jackwagon -0 points

Anything people can do to help, besides signing in?

[]go2dfish -0 points

Running other access peers helps, should be easier and lower resource now esp if you don't want to run your own indexer.

Running an indexer can be done like FrontPageGuardian has done without exposing your IP or requiring you to run an access peer. This lets you play with moderation.

But mainly signing in and commenting on posts that contribute marks that content as being not denial of service spam.

[] -0 points

Signing in can fuck off. I'm literally most of the content here that's not politics and I'll keep posting anon.

[]go2dfish -0 points

That's fine, people signing in and commenting is a way to raise up legitimate content from the noise of denial of service spam. I don't want to ever eliminate anonymous contributions but they do provide challenges for preventing spam in a way that does not place undue burden on contributors.

[]🐌 -0 points

Is there anywhere I can go to get a high-level, architectural overview of NAB? Perhaps some more documentation about how it works? So far, all I've found is the github, but I'd rather not dig through the source just yet.

[] -0 points

At a very high level the model is a cloud of public content peers replicate amongst each other (GUN). Peers can enforce restrictions on that data (like the structure or content).

Cryptography is used to provide accounts that can sign content, and oracles are clients that generate and sign data on request so you know it comes from a trusted authority.

Those oracles generate signed, sorted lists of identifiers to content in the public cloud of data.

This approach allows individual peers to avoid storing/redistributing content they are morally opposed to or legally enjoined from spreading while focusing the bulk of curation on highlighting content within a shared content network

Fundamentally nab is a pretty thin layer on top of with a react UI based on open source reddit.

[] -0 points

what about captchas?

[]go2dfish -1 points

This is something I'm open to, but not sure exactly when or how I want to apply them yet. I think if I use them I want to use them as an optional label the tabulator applies to make visible someone solved a captcha for the thing and make it possible to filter on that.

[] -0 points

captchas are cancer, but a necessary evil for anon to be a thing; captcha notarized comment wouldn't really be a problem from the centralization perspective since it's a one off thing per post

using it alongside user whitelisting, even for comments, sounds nice

[]go2dfish -0 points

captcha notarized comment wouldn't really be a problem from the centralization perspective since it's a one off thing per post

It can be if you end up using it for anything, it becomes another oracle or part of an existing oracle like the tabulator if you use it for filtering you have to implicitly trust that the tabulator isn't manipulating it in a way you disapprove of.

But the approach of using oracles for everything allows people to pick who they trust easily hopefully without much friction.

[] -0 points

I was thinking about the ip notarization while writing that where the more oracles you use, the less secure your indexer is

[]go2dfish -0 points

So oracles is a pretty muddy concept, I speak of tabulator/indexer as different oracles because they can be split; but functionally right now they are the same account. So in that sense an oracle is a schema of handling a type of data; but I think the way you use it here (equally valid) is in reference to oracle as an account and trusting multiple different oracles?

[] -0 points

Yes, the more ip notarizing oracles you have, the more keys you can get on the same ip. There's no way around this unless the oracles work together.

For captcha notarizing oracles, they're all validating a single piece of content at a time. Therefore the more good oracles, the better.

[] -0 points

Captchas don't work well in an open environment though, each oracle verifying humans with captchas has to present its own challenge so both reader and contributor have to agree on a arbiter for this approach to have use.

[] -0 points

Why would you need to present humans with multiple challanges? A single one, picked by their indexer should be good enough. The indexer is the one that has to listen to multiple different validations.

[]go2dfish -0 points

Hmm yeah that might be an interesting approach.

[]go2dfish -0 points

Also captchas are at best a rate limit for attacks like notabug has seen recently.