Four CAIP-350 profiles exist now (eip155, bip122, solana, starknet) and no two agree with the template on structure or frontmatter. Worth settling before the fifth, since ERC-7930 points implementers at these files.
type: — template says one thing, every profile says another. The template has type: Informational; all four profiles use type: Standard. Repo-wide the split is 82 Standard / 56 Informational, so my guess is the template is just wrong here, but somebody who knows the intent should pick.
## References is in every profile and not in the template. All four authors added it independently. It should probably be a template section — the two profiles missing link definitions (fixed in #195) are a symptom of it being ad hoc.
Optional sections are handled four different ways:
|
Error handling |
Implementation considerations |
Extra considerations |
_template |
✓ |
✓ |
✓ |
eip155 |
empty heading |
empty heading |
✓ |
solana |
empty heading |
empty heading |
✓ |
bip122 |
✓ |
✓ |
omitted |
starknet |
✓ |
✓ |
omitted |
Both readings are defensible — the template marks these _(Optional section)_ — but "empty heading" and "heading removed" shouldn't both be current practice. Suggest the template say explicitly: drop the heading if you have nothing to say.
Minor, adjacent: the template tells authors to check existing profiles and open PRs by hand for binary-key collisions. That's fine at four keys (0x0000–0x0003, currently clean) and won't be at twenty. A single table somewhere — even just in the template's comment — would make the check one lookup instead of a repo grep plus a PR sweep. Related to #141.
Can send a PR for whichever of these you want, once the type: question is answered.
Four CAIP-350 profiles exist now (
eip155,bip122,solana,starknet) and no two agree with the template on structure or frontmatter. Worth settling before the fifth, since ERC-7930 points implementers at these files.type:— template says one thing, every profile says another. The template hastype: Informational; all four profiles usetype: Standard. Repo-wide the split is 82Standard/ 56Informational, so my guess is the template is just wrong here, but somebody who knows the intent should pick.## Referencesis in every profile and not in the template. All four authors added it independently. It should probably be a template section — the two profiles missing link definitions (fixed in #195) are a symptom of it being ad hoc.Optional sections are handled four different ways:
_templateeip155solanabip122starknetBoth readings are defensible — the template marks these
_(Optional section)_— but "empty heading" and "heading removed" shouldn't both be current practice. Suggest the template say explicitly: drop the heading if you have nothing to say.Minor, adjacent: the template tells authors to check existing profiles and open PRs by hand for binary-key collisions. That's fine at four keys (
0x0000–0x0003, currently clean) and won't be at twenty. A single table somewhere — even just in the template's comment — would make the check one lookup instead of a repo grep plus a PR sweep. Related to #141.Can send a PR for whichever of these you want, once the
type:question is answered.