Re: [Uri-review] URI scheme for postal addresses?

[Christian Amsüss <[e-mail]>]

Hello Nicholas,

(hoping this is not in conflict with the liaison manegement topic Eliot
has started)

While the format is URI compliant in the sense that its ABNF describes a
subset of the strings the URI ABNF allows (unless I missed something),
it will behave quite weirdly when used as a URI. In particular, URI
normalization will turn upper case letters in the third segment into
lower case letters if the first two segments are empty. (I'm also told
that common but non-conformant implementations of URI processors will
turn integers in the third segment into IPv4 addresses if the first two
segments are empty, but I personally don't care about those).

The case normalizations would be highly unexpected by users (to whom the
third segment has no special meaning). To avoid such issues, I'd suggest
that an empty first segment should be disallowed; the corresponding ABNF
would be:

    addressdata = segment-nz *("/" segment)
    segment-nz = 1*urlchar

Best Regards
Christian

-- 
To use raw power is to make yourself infinitely vulnerable to greater powers.
  -- Bene Gesserit axiom