Re: Draft: P1788.1 Standard for Interval Arithmetic (Simplified)

[Walter Mascarenhas <walter.mascarenhas@xxxxxxxxx>]

I believe things would be clarified if the standard

stated explicitly the expected result for

   ceil(  [0.5, 1.0]_com )

 and

   ceil( [0.5, 0.75]_com )

All definitions have pros and cons and we have

passed the time to argue about "the best one",

but there should be no ambiguity on the standard

regarding the committee's choice.

            

On Tue, Sep 8, 2015 at 1:34 AM, Kreinovich, Vladik <vladik@xxxxxxxx> wrote:

So here is when we have a confusion: what should we look for:

* the function is continuous on all the points from the interval

or

* the restriction of the function to the interval is continuous

ceil is a good example when these two things are different

________________________________________
From: stds-1788@xxxxxxxx [stds-1788@xxxxxxxx] on behalf of Oliver Heimlich [oliver@xxxxxxxxxxxxxxxxxxx]
Sent: Monday, September 07, 2015 10:24 PM
To: Walter Mascarenhas
Cc: Ned Nedialkov; IEEEP1788a
Subject: Re: Draft: P1788.1 Standard for Interval Arithmetic (Simplified)

On 07.09.2015 23:33, Walter Mascarenhas wrote:
> Are you sure about the following statement in your message?
>
>  >> For example: ceil( [0.5, 1]_com ) = [1, 1]_dac, because ceil is not
> continuous in 0.5, but the restriction of ceil is.

Of course ceil is continuous in 0.5. I was thinking of the function
roundTiesToAway. However, ceil is not continuous in 1, so the example
still works.