Re: Fw: Comments on LDP-Paging - 5.1.5 whole != sum(parts)

On Thu, Jul 31, 2014 at 1:31 PM, John Arwe <johnarwe@us.ibm.com> wrote:
>
> >  MUST NOT is hard to test compliance to anyway.
>
> 1: I don't think it's any easier nor harder to test in this case than
Should Not.  It won't be automatable either way.
>
I'm not sure what automation has to do with it (I never brought up
automation), I was just talking about someone (or something) being able to
test conformance to this clause.  MUST NOT is much stronger than SHOULD NOT
+ adding the fact it is near impossible for someone to test (manually or
automated), not to mention I think MUST NOT is too strong, that is why I
was suggesting SHOULD NOT.  Yes, still hard to test the should not.

> 2:   That's a distraction.  Automated tests are desireable+helpful, not
required.  Sandro has a numbered rant for that, if you think back to March.
 5.1.6 isn't automated-testable either, but it's still the right behavior
and WE get to define the compliance criteria for LDP Paging clients.  Ditto
5.1.2, 5.1.3, 5.1.4... so 2/3 of the existing client constraints are not
easily testable using automation, by my count.
>
See above

- Steve

> Best Regards, John
>
> Voice US 845-435-9470  BluePages
> Cloud and Smarter Infrastructure OSLC Lead
>