[Cfp-interest 2193] Re: AI about new wording for unaccepted change

[Mike Cowlishaw]

OK, yes, that's clearer.
 
Mike


  _____  

From: Cfp-interest [mailto:cfp-interest-bounces at oakapple.net] On Behalf Of Jim Thomas
Sent: 04 October 2021 01:05
To: CFP
Subject: [Cfp-interest 2192] Re: AI about new wording for unaccepted change





On Oct 2, 2021, at 2:41 PM, Jim Thomas <jaswthomas at sbcglobal.net> wrote:




On Oct 1, 2021, at 6:09 AM, Fred J. Tydeman <tydeman at tybor.com> wrote:

On Thu, 30 Sep 2021 15:46:56 -0700 Jim Thomas wrote:



b) The results are not equivalent because they have different quantum exponents.




How about:

The results are equal, but have different quantum exponents, hence are not equivalent.



The statement suggests that if the results were equal (meaning compare equal) and had the same quantum exponent, then they would be equivalent, which is not true for signed zeros. The two result 12.34 and 12.340, on the preceding two lines, are obviously equal. So saying the results are equal doesn’t seem needed. 



Sorry. These last two sentences can only have confused the situation. The referenced preceding two lines 


1. × 12.34 = (+1, 1, 0) × (+1, 1234, −2) = (+1, 1234, −2) = 12.34 

1.0 × 12.34 = (+1, 10, −1) × (+1, 1234, −2) = (+1, 12340, −3) = 12.340


exemplify two different transformations, the first valid, the second not valid.

The current text and the suggested changes start with “The results” which would seem to refer to the two preceding lines, which is misleading. I believe WG14 was concerned about the use of “equal” here and not this second problem.

Here's another suggestion for the new words (to follow the two lines above):


c) In the second case, the factor 12.34 and the result 12.340 have different quantum exponents, demonstrating that 1.0 * x and x are not equivalent expressions. 


- Jim Thomas




(I believe the original idea behind the term "numerically equal” was to exclude different signed zeros and hence imply more than just equal or compare equal.)

- Jim Thomas




---
Fred J. Tydeman        Tydeman Consulting
tydeman at tybor.com      Testing, numerics, programming
+1 (702) 608-6093      Vice-chair of PL22.11 (ANSI "C")
Sample C99+FPCE tests: http://www.tybor.com
Savers sleep well, investors eat well, spenders work forever.

_______________________________________________
Cfp-interest mailing list
Cfp-interest at oakapple.net
http://mailman.oakapple.net/mailman/listinfo/cfp-interest




_______________________________________________
Cfp-interest mailing list
Cfp-interest at oakapple.net
http://mailman.oakapple.net/mailman/listinfo/cfp-interest


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.oakapple.net/pipermail/cfp-interest/attachments/20211004/248f0c10/attachment-0001.htm>