Why does Chrome need this repository?

[coderbyheart]

I am curious to understand the rationale behind this repository. It is very much against the decentralized nature of the web to maintain the list of related domains to a website in a centralized repository.

I am fine with maintaining a /.well-known/related-website-set.json on my websites, but it is a big hassle to provide this information in a static form here.

The reason is that I am using serverless backends, that give me endpoint domains like https://<random string>.execute-api.eu-west-1.amazonaws.com/. So I would need to go through the extra effort to put these endpoints behind a static domain name in order to keep the information in this repo up to date.

Why can't chrome just fetch the /.well-known/related-website-set.json file on first visiting the website and discover related websites from there? You could even cache this file (while honoring it's caching headers) and serve it to other users.

[thegreatfatzby]

@coderbyheart think this might give you the short term answer WICG/first-party-sets#128

EDIT: more info WICG/first-party-sets#180