[RFC6265bis, Editorial]: Define "public suffix" in the terminology section. #1038
Conversation
|
Would you mind taking a look, @reschke? |
|
I think this document has the conceptual conceit that it might be possible to have distinct understandings of public suffixes. Browsers certainly have coalesced around https://publicsuffix.org/, but it's not clear to me that it's "normative" in any sense other than practice. A de facto standard, not de jure. I'm happy to just say "Use the PSL", as that's most likely to be most compatible? |
|
Well, if we have normative requirements attached to whether something is a public suffix, then we need to define how to decide that. If there's a risk that the list ever moves, we could introduce an indirection through a new IANA registry (just holding the location of the public list, not the list itself). |
|
The current normative requirements boil down to "Hey, user agent. Go figure out which suffixes you think are public. The PSL might be helpful. Have fun!" I'm happy to make that suggestion normative by bringing in the PSL in the same way URL does (https://url.spec.whatwg.org/#host-public-suffix). Would you be happy with that outcome? |
|
Yes, this seems to be needed (maybe not in the same words, but...). |
draft-ietf-httpbis-rfc6265bis.md
Outdated
| "a domain that is controlled by a public registry", and are also known as | ||
| "effective top-level domains" (eTLDs). For example, `site.example`'s public | ||
| A "public suffix" is a domain that is controlled by a public registry, such as | ||
| "com", "co.uk", and "pvt.k12.wy.us". For example, `site.example`'s public |
There was a problem hiding this comment.
Would it be worth making a comparison to "delegation-centric zone" from RFC 8499? It's certainly possible for a public suffix to not be delegation-centric (e.g., github.io) and there are some delegation-centric zones that are not public suffixes (most notably the DNS root zone itself), but there is nevertheless substantial overlap in both concept (different authority between parent and child domain) and contents.
There was a problem hiding this comment.
I'm not sure that bringing another concept in would clarify, especially given the discrepancies you note between "delegation-centric zone" and existing user agent understanding of "public suffix". I think I'd prefer to continue using the term that user agents have generally applied to cookies.
|
@reschke I've update the patch to more explicitly point to the PSL's algorithm, and switched from "registered domain" to "registrable domain" to match URL. WDYT? |
|
Works for me... |
Closes #1033.