Re: [xacml] Question on Time Extensions

[Steven Legg <steven.legg@viewds.com>]

Hi Hal,

On 4/10/2019 5:41 am, Hal Lockhart wrote:
> Steven,
>
> I have finally taken a careful look at the profile. I understand the need for the timezone attribute to be in multiple categories.
>
> However, I am still stuck on the "equality rule for dayOfWeek" issue.
>
> I am not sure what functionality is desired here. Given two datetimes normalized to the same timezone, determining if they are both in the same day of week or not seems straight forward to me. What am I missing?
>
> It also seems like the same procedure would let you compare a normalized datetime to a dayOfWeek and determine if the former falls in the latter.

Comparing dateTime values to dayOfWeek values was already taken care of by the
dateTime-in-dayOfWeek-range function. The question was around how to compare two
dayOfWeek values to determine if they are the same, in the absence of any dateTime
value, though the dateTime-in-dayOfWeek-range function may have already implicitly
given the answer.

It currently seems that time zones range from -12:00 on Midway Island to +14:00
on Kiribati. There's a lot of overlap in the middle of the Pacific Ocean. To take
one example, Tuesday on Hawaii (2-10:00) is exactly the same span of time as
Wednesday on Kiribati (3+14:00). So the question arises, should we use that
equivalence as the equality function (so 2-10:00 equals 3+14:00) or is it too
weird to being saying that a Tuesday is sometimes equal to a Wednesday? Note that
2-10:00 and 2-09:00 are not exactly the same span of time so here we have a
Tuesday that is not even equal to another Tuesday.

Now consider the dateTime-in-dayOfWeek-range function. If the function returns
true for some dateTime value in the dayOfWeek range 2-10:00 to 3-10:00 then it
also returns true for the same dateTime value in the range 3+14:00 to 4+14:00
(and vice versa). So from the perspective of the dateTime-in-dayOfWeek-range
function 2-10:00 and 3+14:00 are equivalent and interchangeable.

Regards,
Steven

> Help me out here.
>
> Hal
>
> <http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> 	Virus-free. www.avg.com <http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
>
> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>