Editorial: adopt Infra number definitions (and redo IPv6 pieces) #838
Conversation
| <p>An <dfn export id=concept-ipv6>IPv6 address</dfn> is a <a for=/>128-bit unsigned integer</a> that | ||
| identifies a network address. This integer is composed of a <a for=/>list</a> of 8 | ||
| <a for=/>16-bit unsigned integers</a>, also known as an <a for=/>IPv6 address</a>'s | ||
| <dfn export for="IPv6 address" id=concept-ipv6-piece>pieces</dfn>. |
There was a problem hiding this comment.
This definition is obscure: it is not clear how to convert a 128-bit integer to a list of pieces and vice versa. Also, all algorithms use an IPv6 address as a list of pieces, not as a 128-bit integer. By the way, RFC4291 defines an IPv6 address slightly differently: "IPv6 addresses are 128-bit identifiers...". So I think it is better to use the RFC4291 definition - replace "unsigned integer" with "identifier":
An IPv6 address is a 128-bit identifier that identifies a network address. This identifier is composed of ...
There was a problem hiding this comment.
All algorithms inside URL, yes, but outside? This we would have to change separately though as changing it from integer to identifier would not be an editorial change.
There was a problem hiding this comment.
I do agree that this is a bit confusing. I can see how it's "conceptually" a u128, or "corresponds to" a u128, but in practice every part of the spec treats it as a list of 8 u16s.
| <p>An <dfn export id=concept-ipv6>IPv6 address</dfn> is a <a for=/>128-bit unsigned integer</a> that | ||
| identifies a network address. This integer is composed of a <a for=/>list</a> of 8 | ||
| <a for=/>16-bit unsigned integers</a>, also known as an <a for=/>IPv6 address</a>'s | ||
| <dfn export for="IPv6 address" id=concept-ipv6-piece>pieces</dfn>. |
There was a problem hiding this comment.
I do agree that this is a bit confusing. I can see how it's "conceptually" a u128, or "corresponds to" a u128, but in practice every part of the spec treats it as a list of 8 u16s.
Preview | Diff