Re: FastCode RoundToEx
- From: "John Herbster" <herb-sci1_AT_sbcglobal.net>
- Date: Sat, 29 Dec 2007 15:46:53 -0600
"Dennis" <marianndkc@xxxxxxxxxxxxxxx> wrote
Now with all the misunderstandings cleared up we can focus on the real question.
Is the RoundToEx a healthy improvement to RoundTo?
Who would use RoundToEx and for what purpose?
I certainly would *not* use it nor its predecessor, nor recommend their use.
My objection is mainly because the D7 Help documentation about what its predecessor, RoundTo, is supposed to be doing is not clear. If its goal is bankers rounding and the documentation has been or is changed to say so, then seems clear to me (from these recent discussions) that the latest RoundTo does not satisfy that criteria and should be fixed. After that, it might be worthwhile to try to extend the precision of the returned results and the allowed input range with RoundToEx.
Can we make RoundToEx even better?
Is the TRoundToEXRangeExtended = -20..20 proper?
etc.
All this discussion has reinforced my belief that *most* applications programmers do not know how to properly code decimal fraction arithmetic. It has caused me to collect my thoughts on the problem of rounding such numbers. I have started work on a DecimalRounding_JFH2 module that I think will help these programmers better than anything that I have seen so far.
--JohnH
.
- Follow-Ups:
- Re: FastCode RoundToEx
- From: Q Correll
- Re: FastCode RoundToEx
- References:
- FastCode RoundToEx
- From: Dennis
- Re: FastCode RoundToEx
- From: Dennis
- Re: FastCode RoundToEx
- From: John Herbster
- Re: FastCode RoundToEx
- From: Dennis
- Re: FastCode RoundToEx
- From: John Herbster
- Re: FastCode RoundToEx
- From: Dennis
- FastCode RoundToEx
- Prev by Date: Re: FastCode Reference Compilers
- Next by Date: Re: Nexus versus FastMM4 Round 2
- Previous by thread: Re: FastCode RoundToEx
- Next by thread: Re: FastCode RoundToEx
- Index(es):
Relevant Pages
|