Re: Comments on draft-fielding-uri-rfc2396bis-07

Bruce Lilly wrote:
> Apparently the W3 archive implementor(s) also believe that
> '@' should be encoded as "%40"

It should be noted that in response to email harvesting, savvy HTML authors 
and especially those who publish arbitrary email addresses have, in recent 
years taken to obfuscating email addresses by various means, including the 
replacement of "@" with "%40" in mailto URIs, "@" with "@" or " (at) " in 
character data. This method works surprisingly well, according to research 
(harvesters generally only look for raw "@"), and is therefore unlikely to 
fall out of favor anytime soon.

In any update of the mailto scheme specification, I would hope that this 
practice is taken into consideration.

Besides, I don't think there is really any benefit to making "@" have a 
reserved purpose in the path component of a mailto URI. The path component 
need not be segmented in any manner. It will still adequately represent an 
email address. The significance of "@" is already handled by the syntax 
requirements of the address being represented by the path component; there is 
no need to enforce the syntax of the represented address within the 
representation too. And dereferencing is sufficiently specified, for the path 
component at least, by normalization of percent-encoding to determine the 
address being represented, and deferring the rest to the implementation, IMHO.


Received on Friday, 5 November 2004 19:36:44 UTC