On 2015-03-04 1:32, Andrei Robachevsky wrote:
> Stephen Kent wrote on 03/03/15 17:29:
>> I worry that accommodating multiple signatures will cause confusion for
>> RPs. One would need to specify what to do if one sig fails, but other
>> succeed,
>> for example.
> 
> I think the draft is clear about that, requiring all signatures to be
> valid. And if we want to follow the RPSS/RFC2725 approach, then multiple
> signatures are needed.
> 
> But, it is not entirely clear to me why we need an "o" field and not
> just multiple "signature:" attributes in cases when signing by several
> parties is required.

Indeed, that is why we're dropping it. The o= field was suggested a long
time ago to make interdependent signatures. When thinking about the
implementability of it, it became clear that it has a *lot* of added
complexity, with not much benefit, if you compare to multiple, independent
signatures (which only make real sense for route objects, I think).

The draft already allows multiple signatures, therefore dropping the o=
field is the simplest and most forward looking step.

Cheers,
Robert

_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to