Re: [DNSOP] draft-liman-tld-names-04

[Eric Brunner-Williams <[e-mail]>]

On 11/29/10 9:39 AM, Andrew Sullivan wrote:

> ... the Internet will not break if
> you register a domain name o'reilly.com,

please either stick to the ldh set and the issue of the validity of 
any string composed from this generating set, or explain why, other 
than for rhetorical excess, you are permitted to pick an arbitrary 
point in the us-ascii standard and wave it about as proof of some claim.

there exist rules, much earlier than 1123, about "-" as the initial, 
or terminal, byte in a sequence of bytes, and about sequences of two 
or more instances of "-" where neither byte is the initial, or 
terminal, byte in a sequence of bytes.

there exist rules, much earlier than 1123, about any digit as the 
initial byte in a sequence of bytes.

there exist rules, much earlier than 1123, about whether a sequence of 
bytes must contain more than one byte.

it is my understanding that a view associated with liman's draft is 
that some protocol restriction existed in fact, or should be 
generously imputed to have existed as a protocol restriction in the 
minds of some implementers, and that one or more of these rules are 
not completely explained as allocation policies by historic allocators 
of some zone of interest.

i don't share that view, and i don't think it prudent to abandon 
whatever merits may support that view by appealing to the absence of 
us-ascii values outside the ldh set, or phrasing like "old timey".

-e