[Telepathy] How to name Ring.cx Telepathy Connection Manager

[Daniel Pocock]

On 12/06/16 09:17, George Kiagiadakis wrote:
> On 11.06.2016 10:51, Daniel Pocock wrote:
>> On 11/06/16 08:35, George Kiagiadakis wrote:
>>> On 06/10/2016 06:55 PM, Daniel Pocock wrote:
>>>>
>>>>
>>>> On 10/06/16 17:49, Stepan Salenikovich wrote:
> 
> [...]
> 
>>>
>>> #2 CM should not be named after the implementation because that can also
>>> lead to confusion between the two projects (the telepathy CM and the
>>> implementation software). See the example of telepathy-sofiasip that
>>> became telepathy-rakia (mentioned in the spec as well). This is also a
>>> factor against telepathy-ring (or telepathy-ring.cx perhaps) because
>>> ring.cx is another software. You don't want users to think that this
>>> telepathy CM is somehow part of ring.cx or it can be used to talk to
>>> people only if they are using ring.cx on the other side. Maybe another
>>> implementation will pop up in the future and then ring.cx will be
>>> irrelevant as a name.
>>>
>>
>> This is not so clear cut.
>>
>> I chose the name telepathy-resiprocate for the resiprocate-based CM
>> because telepathy-resiprocate is maintained in the reSIProcate
>> repository:
>>
>> https://github.com/resiprocate/resiprocate/tree/master/apps/telepathy
>>
>> Maybe an exception should be made so that when a CM is officially part
>> of a protocol implementation (reSIProcate is a SIP implementation), then
>> it can use the name of the implementation.
>>
> 
> Hmm, by the rule this was probably a bad choice, but I can see how you
> mean it. Being part of the same project as the implementation maybe
> makes those arguments less strong. I guess it remains to be seen whether
> this will have the same problem as tp-sofiasip in the future.
> 

For clarity, it looks like the tp-sofiasip code is not in the SofiaSIP
repository:

https://sourceforge.net/p/sofia-sip/git/ci/master/tree/

so they are independent projects and that rule makes sense in that case.

> I think that the spec is clear about that, though. Quote:
> "Connection manager names SHOULD NOT be the same as the name of a
> library on which they are based."
> 

Do you think the spec could be revised to accommodate the situation
where a CM is maintained by the same vendor as the protocol
implementation, most likely in the same repository?

Regards,

Daniel