Anyone else think a decentralized banner/link exchange standard is a good idea?
If you want to encourage your visitors to also visit somebody else's site, you should be able to do that by copying a single URL they provide. No third party needed.
That URL would point to a JSON document or something with the info you need to show link banners of various sizes. It could also list other sites you might want to support.
We could replace ads with a free and easy way to support each other.
@zaclittleberry Yeah, I'm coming at it from the perspective of webcomic creators since I've been running a centralized banner exchange for that community for a few years now, but I think lots of indie creators and social good orgs could use help with cross-promotion.
I think the first step is a WordPress plugin, probably, as proof of concept. That should be "simple" but unfortunately I'm not nearly comfortable enough with the WordPress plugin API...
@jamey Cool! I definitely think it should be open source and support smaller instances that are topic focused. That way different communities could run and join what makes sense for them. My immediate thought is that it could run off of ActivityPub (small independent instance subscription) or maybe something with .well-known directory (that feels very manual, and like a discovery mechanism. so, not sure). A WP plugin to make it easy is great, but I also think the arch. is first. @alana thoughts?
and yeah Zac, i agree about arch — sketching it is always such a useful discussion-prompt, and going from ludicrously over-simplified to refined seems to do half the work of the subsequent steps (bucketing things into their own features/functions; flagging areas that smell potentially annoying; prioritizing… blah blah… 😊)
@jamey @alana I think the diagram/concept I proposed would fit your first three items under 'Use Cases' with account following, community booster following, and boosting an individual account's content (which I didn't explicitly diagram, but smaller scale concept of a community booster, which would probably be automated). The 4 use cases after those however are more algorithmic/analytics based and I'm not quite as sure about. Also your discovery section aligns with activitypub as well I think.
@jamey @alana Using ActivityPub was to get de-centralization, and federation/interoperability, and I think some existing Instances of activity pub (such as Pixelfed) could be forked/modified to get this working quickly. I also lean towards standardized models for non-centralization/federation vs yet-another-independent-spec.
That process helped clarify some of my hidden goals, like: if possible, I would like one or both sides of the protocol to be implementable on a purely static web site.
Also now I need to read papers like https://crypto.stanford.edu/adnostic/ and https://eprint.iacr.org/2018/323.pdf to see if there are usable privacy-preserving ways to choose the best banner to show…
On the internet, everyone knows you're a cat — and that's totally okay.