On 1/20/15, 04:13, "Chris Dent" <chd...@redhat.com> wrote:
>On Mon, 19 Jan 2015, Dean Troyer wrote: > >> Independent of actual implementations in OpenStack, I prefer always >> including null/empty properties here because it is slightly more >> self-documenting. Having spent the morning chasing down attributes for >>an >> API to be named at a later date by looking at server code, we do not >>help >> ourselves or the users of our APIs by omitting this sort of thing. > >+1 I’m in much the same position as Kevin. I’ve tried coming up with reasoning for both positions but I don’t see a really compelling reason for either side. On the one hand though, having strict schema about what is returned can be valuable, so not allowing something to be omitted **may** catch a bug somewhere else. Allowing things to be nullable is perfectly reasonable to me. It seems like a few of us are in agreement with this direction. Perhaps one of us should write a proposal for the API-WG to review about this. Cheers, Ian __________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
