Re: [apps-discuss] Fwd: FW: New Version Notification for draft-kerwin-file-scheme-13.txt

[Sean Leonard <[e-mail]>]

On 1/10/2015 11:21 AM, Sam Ruby wrote:
> On 1/10/15 1:15 PM, Sean Leonard wrote:
>> [SNIPS]
>>>
>>> I part company with you here.  IDNA processing can be avoided - 
>>> don't do
>>> it!
>>>
>>> Many of the IETF standards have no i18n in general, or IDNA in
>>> particular,
[SNIP]
>> Folks are confusing the "web browser text box at the top" with URIs. Web
>> browsers can display and accept whatever characters they want in the
>> address box. How they do that is not an RFC 3986 issue. It might be
>> governed by some de-jure or de-facto Web "standards" (or RFC 3987?), but
>> it's not directly relevant to URIs. The end.
>>
>> Nowadays it's not even a "URL bar" or "address bar". It's variously
>> called an "omnibox" and has a lot of code running behind it to do
>> context-sensitive things with lots of additional user experience
>> widgets. For example if you type "? ietf" in Chrome (and others...I
>> believe Internet Explorer 6 had this) it's going to do something that
>> has nothing to do with URIs.
>
> I call strawman.  Nobody here mentioned a quote "web browser text box 
> at the top" end-quote.

It's not really a strawman: there are different experiences at play. I 
do a lot of security work; in certificates, a URI (GeneralName -> 
uniformResourceIdentifier) is stored in an IA5String data type, which is 
US-ASCII only. Therefore unlike a blob of HTML, arbitrary Unicode 
characters cannot possibly be present. The example 
<file://０Ｘｃ０．０２５０．０１/foo> provided literally cannot exist in 
that format; it would be treated as a decoding error.

Under the (obsolete...but still widely used) HTML 4.01, it says <a 
href=""> is supposed to have only %URI characters, so the presence of 
non-URI characters is either an error (if the DTD is strictly enforced), 
or results in undefined/implementation-defined behavior (which is what 
actually happens, since the DTD almost never gets strictly enforced...by 
web browsers). I suppose that the newer standards are intended to clean 
that up...

Sean