Re: IRI length limits (ticket #37 => #46, #47, #48)

On Wed, Sep 29, 2010 at 3:13 AM, "Martin J. Dürst"
<duerst@it.aoyama.ac.jp> wrote:
<snipped>
>
> I agree in principle. But the 63 octet length limit for domain names does
> not directly affect IRIs, and therefore cannot be a normative limit, only an
> informative one. You give an example of that further below, but to quote
> from the URI spec (RFC 3986, see
> http://tools.ietf.org/html/rfc3986#section-3.2.2):
>
>   This specification does not mandate a particular registered name
>   lookup technology and therefore does not restrict the syntax of reg-
>   name beyond what is necessary for interoperability.  Instead, it
>   delegates the issue of registered name syntax conformance to the
>   operating system of each application performing URI resolution, and
>   that operating system decides what it will allow for the purpose of
>   host identification.  A URI resolution implementation might use DNS,
>   host tables, yellow pages, NetInfo, WINS, or any other system for
>   lookup of registered names.  However, a globally scoped naming
>   system, such as DNS fully qualified domain names, is necessary for
>   URIs intended to have global scope.  URI producers should use names
>   that conform to the DNS syntax, even when use of DNS is not
>   immediately apparent, and should limit these names to no more than
>   255 characters in length.
>

Perhaps the easiest way to resolve this, then is a set of informative
pointers to the suggested and normative limits.  RFC 3986 for the
suggestion to fit within DNS limits, RFC 1034 for the DNS limits,
and a note to check the scheme registration for specific schemes
when checking limits.

Does that work?

Ted

Received on Wednesday, 29 September 2010 17:22:58 UTC