On Wed, Apr 24, 2019 at 05:53:02PM +0000, Reshad Rahman (rrahman) wrote: > On 2019-04-19, 11:56 PM, "Benjamin Kaduk" <kaduk@xxxxxxx> wrote: > > On Fri, Apr 12, 2019 at 09:29:35PM +0000, Reshad Rahman (rrahman) wrote: > > Hi Aanchal, > > > > Thanks for the review. Please see inline. > > > > On 2019-04-11, 5:54 PM, "netconf on behalf of Aanchal Malhotra via Datatracker" <netconf-bounces@xxxxxxxx on behalf of noreply@xxxxxxxx> wrote: > > > > Reviewer: Aanchal Malhotra > > Review result: Ready > > > > The document is very clear and concise. I just have one minor clarification question. > > Section 3.4 Page 9 that says the following: > > "In addition to any required ........SHOULD only be allowed......". > > > > Is there a reason for using SHOULD instead of MUST? > > > > There may be reasons why an implementation decides not to enforce this restriction. Going by RFC2119 definitions, this is why we chose SHOULD instead of MUST. > > If you have some reasons in mind, it is often helpful to list them as > examples of when the recommended behavior would not be followed. > > What we had in mind is a "super-user" who could be given access to subscriptions of other users. Is this obvious or should I can add text to that effect at the end the bullet below? Something along the lines of "For example, a RESTCONF username with the required administrative permissions could be allowed to invoke RPCs modify-subscription, resync-subscription and delete-subscription on a subscription which was created by another username.". > > o In addition to any required access permissions (e.g., NACM), RPCs > modify-subscription, resync-subscription and delete-subscription > SHOULD only be allowed by the same RESTCONF username [RFC8040] > which invoked establish-subscription. I think it might help to have such text, though I would suggest a slightly pithier "Such a restriction generally serves to preserve users' privacy, but exceptions might be made for administrators that may need to modify or delete other users' subscriptions." Thanks, Ben
