I would be supportive of removing non-mailto contact examples.

I don't think it's accurate to say that "mailto" is the only contact scheme
currently supported by the ACME standard. That isn't the case for the spec
as-written today - it is the only mandatory contact scheme. I believe its
been left intentionally open ended what other schemes may be supported by a
server.

In my mind ACME servers wishing to support `tel` should come up with their
own understanding of what parts of the relevant RFCs they support and use
the `unsupportedContact` and `invalidContact` errors as appropriate based
on the `tel://` contacts received. I don't think there is value trying to
create a subprofile of `tel` for ACME use that is defined in the ACME draft.

- cpu



On Fri, Nov 17, 2017 at 11:12 AM, Logan Widick <[email protected]>
wrote:

> Or, would it be best to remove the non-mailto contacts from the examples,
> and add a note indicating that "mailto" is the only contact URI scheme
> currently supported by the ACME standards?
>
> Sincerely,
>
> Logan Widick
>
> On Wed, Oct 25, 2017 at 3:18 PM, Logan Widick <[email protected]>
> wrote:
>
>> This is related to my previous "Allowable mailto contacts" thread earlier
>> this month. However, this is about a different type of contact, so I
>> decided to start a new thread.
>>
>> Although ACME servers don't have to include support for "tel" (RFC 3966)
>> contacts, "tel" contacts appear in multiple examples in the ACME
>> specification. Thus, some may try to include "tel" contact support in their
>> ACME server implementations.
>>
>> Although the example "tel" contacts in the ACME specification seem to
>> indicate a simple, narrow subset (the "global-number" with zero "par"
>> instances), the "tel" specification has many more possible productions. For
>> example, both "global-number" and "local-number" may have "par" instances
>> (like ";name1=value1;name2=value2") after the phone numbers.
>>
>> Assuming I understand the "local-number" specification correctly, a
>> "local-number" could potentially contain a pound sign ("#") for reasons
>> related to DTMF. If I understand the current and past URI specifications
>> correctly, this could be problematic. Specifically, a URI parser may put
>> the parts of the "local-number" after a "#" in the fragment.
>>
>> If someone wanted to implement "tel" contact support for an ACME server,
>> what subset of "tel" would be required? Would both "global-number" and
>> "local-number" instances be allowed? What restrictions would apply to
>> "global-number" and "local-number" instances? Which types of "par"
>> instances would be allowed? Extension? ISDN-subaddress? Others?
>>
>> On a more general note, I think that as ACME becomes more widely used,
>> the subject of what subsets of what contact URI schemes would be allowed
>> under ACME (as opposed to what schemes are supported by a specific ACME
>> server, which was a separate issue already discussed in March on the
>> mailing list under "Allowing clients to understand the account creation
>> functionality supported by a server") may come up more often. Would it be a
>> good idea to request the creation of an "ACME Contact Schemes Registry"
>> like in the ASCII art figures below?
>>
>>
>>
>>
>> +--------+-----------+-----------+--------------------------
>> -------------------+
>> | Scheme | Status    | Reference | ACME Restrictions
>>      |
>> +--------+-----------+-----------+--------------------------
>> -------------------+
>> | mailto | Mandatory | RFC 6068  | Exactly one "addr-spec" in the "to"
>> portion |
>> |        |           |           | No "hfields" are allowed
>>       |
>> |        |           |           | [Source: "Allowable mailto contacts"
>> thread]|
>> +--------+-----------+-----------+--------------------------
>> -------------------+
>> | tel    | Optional  | RFC 3966  | [insert restrictions picked in this
>> thread] |
>> +--------+-----------+-----------+--------------------------
>> -------------------+
>>
>> Figure 1: A possible "ACME Contact Schemes Registry" table. In this
>> version, the ACME restrictions on the original contact scheme
>> specifications are stated in the table, and the original contact scheme
>> specification is listed as the "Reference". These relationships are
>> depicted graphically in Figure 2 (below).
>>
>>
>>
>>
>> +-----------------------+        +--------------------+
>> | RFC 6068 (mailto RFC) |        | RFC 3966 (tel RFC) |
>> +-----------------------+        +--------------------+
>>
>> Figure 2: RFC relationships for Figure 1
>>
>>
>>
>>
>> +--------+-----------+---------------------------+
>> | Scheme | Status    | Reference                 |
>> +--------+-----------+---------------------------+
>> | mailto | Mandatory | RFC XXXX                  |
>> |        |           | (ACME RFC)                |
>> |        |           |                           |
>> +--------+-----------+---------------------------+
>> | tel    | Optional  | RFC YYYY                  |
>> |        |           | (Some RFC after RFC XXXX) |
>> +--------+-----------+---------------------------+
>>
>> Figure 3: Another possible "ACME Contact Schemes Registry" table. In
>> this version, the "Reference" RFCs discuss the ACME-specific
>> restrictions for the contact schemes, and refer to the original
>> specifications for the contact schemes. These relationships are depicted
>> graphically in Figure 4 (below).
>>
>>
>>
>>
>> +-----------------------------+
>> | ACME restrictions on mailto |
>> +----------+------------------+
>>            ^
>>            |
>>            | Contains
>>            |
>> +----------+----------+                  +--------------+
>> | RFC XXXX (ACME RFC) |    References    | RFC 6068     |
>> +                     +----------------> + (mailto RFC) |
>> |                     |                  |              |
>> +----------+----------+                  +--------------+
>>            ^
>>            |
>>            | References
>>            |
>> +----------+-------------------+                  +-----------+
>> | RFC YYYY                     |    References    | RFC 3966  |
>> + (An RFC after RFC XXXX)      +----------------> + (tel RFC) |
>> |                              |                  |           |
>> +----------+-------------------+                  +-----------+
>>            ^
>>            |
>>            | Contained in
>>            |
>> +----------+------------------+
>> | ACME Restrictions on tel    |
>> +-----------------------------+
>>
>> Figure 4: RFC relationships for Figure 3 version of "ACME Contact Schemes
>> Registry"
>>
>
>
> _______________________________________________
> Acme mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/acme
>
>
_______________________________________________
Acme mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/acme

Reply via email to