Marcin Rataj

I like the idea of it being a CID! Small ones could be inlined, bigger ones could be fetched once and reused efficiently. Added it as **(B)**

- Note on cache control: DAG walk implemented by IPLD is deterministic, so we could indicate that response can be cached + (tdb if revalidated in the background). - Note...

> asking for bytes 1000MB-1001MB of a CAR file seems like a small request but in reality is very costly on the server Agree, there is dangerous resource usage asymmetry...

This turns out to be more involved, as we are lacking support for `dag-json` and `dag-cbor` in various places (e.g. https://github.com/ipfs/go-cid/pull/137, https://github.com/ipfs/go-ipfs/pull/8568). We can't ask users to provide selector CID...

In my mind, CAR resumption will not be sending the same request again. The idea is for the client to be smart to import as many blocks as possible, and...

Dropping some notes after [IPFS Thing 2022](https://discuss.ipfs.io/c/communities/ipfs-thing-2022/32): - feels like we may want to do more UX work before we pull the trigger on this one - subjective temperature check:...

Updated section related to `dist.ipfs.io` to reflect current state where DNSLink update is automated via CI https://github.com/ipfs/distributions/pull/745

@gitaaron what is the problem / use case you are trying to solve?

Triage note: - done: researched https://github.com/th-ch/yamux-js - identified gaps around working with streams - next: implementing gaps + migrating forked yamux-js to async/iter

Since https://github.com/ipfs-shipyard/ipfs-webui/pull/1563 we delegate API detection to ipfs-provider, which by default will try accessing API on the same origin: https://github.com/ipfs-shipyard/ipfs-provider/blob/3e275b3aa79788a3adce1e3fb3c0b1fbfff0d972/src/providers/http-client.js#L43-L53 This is a basic heuristic, more advanced logic could be...