Re: [admin] Security/Privacy normative text

(this is tagged as [admin] — is that really so?)

Hi,

On jeu., 2014-03-27 at 13:49 +0000, Frederick.Hirsch@nokia.com wrote:

> A concern appears to be the normative requirements in the security and
> privacy considerations. Personally I believe it is very important to
> retain requirements for these topics on implementations despite
> difference in nature.



As I discussed during the call, I think we should limit normative
requirements in our specs to requirements that have an impact on
interoperability. Since the security/privacy considerations of HTML
Media Capture don't (they're about how to communicate to the user the
possible associated risks with media sharing), I think they should not
be normative requirements.

A logical consequence is that they should not have test cases in our
test suites (and thus the pull request should be amended as Tobie
suggested).

If there is consensus on this, we should:
* update the spec to remove the normative requirements
* comment on the pull request to confirm the removal of the relevant
test cases


> I also note (based on discussion during last teleconference) that the
> Geolocation Recommendation has almost the same requirements, that they
> are normative, and included in the test plan - and it is a
> Recommendation. Thus we are not the first and such material has been
> accepted in another instance.



As noted also during the call, Geolocation is probably not the best
example of spec to follow (given it was the first device API to speak
of). 
In general, I think we should align with what most specs do (as long as
it is sensible), rather than what one or two specs do.

Dom