ethcatherders / EIPIP

EIP Improvement Process

77 stars 36 forks source link

Call for Input: Allow Links to CAIP #320

Closed SamWilsn closed 3 months ago

SamWilsn commented 4 months ago

Call for Input

Decision	Do we merge https://github.com/ethereum/EIPs/pull/8247 ?
If Affirmed	Links to [CAIP](https://github.com/ChainAgnostic/CAIPs) proposals are allowed.
If Rejected	No change.
Method	Rough Consensus
Deadline	March 21th, 2024

Background

See https://github.com/ethereum/ERCs/pull/99

g11tech commented 4 months ago

i am of the opinion that we allow as much links as possible to the sources which we deem to be safe to refer (immutability and reputation wise)

bumblefudge commented 4 months ago

^ I would be glad to open a PR that consolidates the allowlist into a simple two-column table with domains (reputation) and regexs (CI-able immutability-check), since each source has a different URL/URN scheme for its immutable docs (and non-immutable docs we presumably don't want to allow links to).

In the case of CAIPs, it might be easier to do allow rendered CAIPs (domain chainagnostic.org) and check a <meta> tag I can rejigger jekyll to include in each rendered page containing that document's status from the jekyll frontmatter? or maybe i can create a separate subdomain or path just for docs with final status so that a simpler regex can be created, like https://chainagnostic.org/CAIPs/final/caip-25?

SamWilsn commented 3 months ago

I wanted to avoid commenting on this one since I wrote EIP-5757, but since no one has discussed whether CAIPs meet the requirements there, I guess I probably should:

Versioned: Yes; git hashes. Availability: Questionable; only existed since August of 2019. No Cost: Yes; repository is CC0-1.0.

My two concerns are:

CAIPs haven't been around that long. 5757 mentions ten years as a reasonable benchmark.
Can't we just copy the CAIP into the assets directory and link to it from there? CC0 would allow that.

Since we have a reasonable workaround (copying to assets), and the only availability benchmark we have hasn't been met, I am going to oppose this CFI.

SamWilsn commented 3 months ago

There's also the question of legitimacy, which I neglected to bring up earlier. While it isn't an official requirement, do we want to recognize CAIPs as a "legitimate" standards body? If we do, we should link to them. If not, put them in assets/.

xinbenlv commented 3 months ago

In strong favor to allow link to CAIP

SamWilsn commented 3 months ago

The general consensus is to affirm this CFI.