Re: motion addressing the need for consistency in distinguishing between numbers and NaNs

[r754@xxxxxxxxxxxxxx]

> Subject: RE: motion addressing the need for consistency in distinguishing 
> between numbers and NaNs
> Date: Tue, 5 Sep 2006 15:54:55 -0700
> From: "Ferguson, Warren E" <warren.e.ferguson@xxxxxxxxx>
> To: "Michel Hack \(1-914-784-7648\)" <hack@xxxxxxxxxxxxxx>,
>         "stds-754" <stds-754@xxxxxxxxxxxxxxxxx>, <stds-754@xxxxxxxxxxx>
>
> Dan, correct me if I'm wrong. 
>
> Dan proposes that we first settle on the name we give "Nancy", i.e., the
> collection of numbers and NaNs, before time is expended in style review
> on implementing any needed changes. I take it that Dan would support any
> motion that gives a name to "Nancy", not just the one that Jim and I put
> together and sent out earlier today.

        I guess I'll answer the questions as asked,

        First, you are wrong in thinking that you need my support for
        anything.  You can make any motion you please & I have no power
        to do anything about it.  (You need only to glance through the
        draft to see that.) :-)

        Second, you are correct that (while I have personal preferences)
        I am in favor of any consistent nomenclature that we can agree
        to before September 20th.

        Third, settling the definition of 'Nancy' (for those outside the
        rarified atmosphere of the style review, the name to be used for
        anything that can be stored in a floating-point variable, finite,
        infinite, or NaN) does not settle the nomenclature issue.  There
        will have to be a motion for that.
>  
> For example, Dan's email of last last month (8/31, 9pm) indicates that
> the following terms need their usage reviewed and modified/corrected as
> needed, especially in light of adopting whatever term we give to
> "Nancy".
>
>               grep -i number 754r.txt | wc
>               100    3553   23273
>
>               grep -i value 754r.txt | wc
>                86    3207   20558
>
>               grep -i entity 754r.txt | wc
>                 7     221    1475
>
>               grep -i entities 754r.txt | wc
>                12     457    3128
>
>               grep -i representation 754r.txt | wc
>                50    1874   12353
>
>               grep -i format 754r.txt | wc
>               319    8424   58966
>
> And yes, "value" should be "floating-point value". The intent was to
> answer Dan's question: "What can be stored in a floating-point
> variable". The proposed motion gives the answer: "a floating-point
> variable stores a floating-point value".
>
> I hope the above answers your last paragraph, i.e., that the proposed
> motion is more than wishful thinking. Let me know otherwise.
>
> Warren

        As to whether Michel is right in his 'wishful thinking' comment,
        I'd like to think he is wrong but only time & our behavior on
        this issue will tell.

        Perhaps we'll find out Thursday...

        Until then,

                                    Dan