Re: [art] Question regarding RFC 8089

[John C Klensin <[e-mail]>]

--On Tuesday, December 18, 2018 11:24 -0600 Adam Roach
<[e-mail]> wrote:

> On 12/18/18 2:53 AM, Martin J. Dürst wrote:
>> The 'mailto:' scheme is about as non-standard an URI scheme
>> as there is,
> 
> Just for clarification, I'm not sure that's true. Off the top
> of my head, schemes like 'sip:' 'tel:', 'xmpp:', 'im:',
> 'iax:', 'jabber:', 'h323:', 'msrp:', 'pres:', 'sms:', 'mms:',
> and their various secure variants have very similar properties
> for the purposes of this conversation (and that's ignoring a
> host of provisional schemes like 'aim:' and 'facetime:').

Adam, you might have added "urn:", which has the added
distinction of being a name rather than a locator.   

In addition to the above, URNs add another complexity to the
discussion: RFC 2141 rather explicitly forbade fragments (and
specifically the "#" character) for use in "urn:" URIs,
reserving that character for future use.   Objections were not
raised at the time and 3986 does not update 2141 to change that
rule, so, while the default may be "yes", there is certainly
precedent for schemes banning the use of fragments.  (Of course,
for URNs, RFC 8141 removed that restriction but did it
explicitly.)

best,
   john