[Ietf-calsify] Issue 80: security considerations section -- proposal

[Aki Niemi]

ext Reinhold Kainhofer wrote:
>> This is a point that has been discussed a lot. It is true that RFC2119
>> text should only be used when there is an actual implementation choice
>> that is observable from the outside. Here, I agree that a "MUST" is just
>> hand-waiving, so I'm ok making it a "must" instead.
> 
> I think it's rather a SHOULD (or a should?). It'a really up to the groupware 
> server developers/designers and out of scope for the data format definition. 
> If a server team has good reasons against implementing some security measure, 
> it's their business, not ours.

Yes, agree. And this is why it says capabilities. It's always in the 
admin's or user's discretion whether to turn them on.

Cheers,
Aki