Improve guidance for integer and floating-point numbers #398
Conversation
This list is currently missing 128-bit numbers, and we could get more integer types in the future. There's no reason to duplicate all of them here -- it will just get out of date again, and isn't needed as part of introducing the integer type.
- Per TR, correct the framing of floating-point as having a margin of error in the result of calculations, not as themselves being inexact. - Match the flow from the integer section, first introducing the floating-point types and then describing them (and contrasting their behavior to integer math). - Show an example of the change in value spacing (ulp) between large/small floating-point numbers. Co-authored-by: Steve Canon <scanon@apple.com>
amartini51
left a comment
amartini51
left a comment
There was a problem hiding this comment.
Thanks @stephentyrone for the corrections! Incorporated your feedback and re-ordered some prose to make it flow better.
Another piece of context I don't think I mentioned. I'm trying to give developers enough information about what makes floating-point math different that they can make an informed decision — so they can avoid the tripping hazard of assuming every number that includes a decimal point should be a Double. But also trying to scare folks away from using "weird floating-point math" when that actually is the right data type.
Co-authored-by: Steve Canon <scanon@apple.com>
colleenish
left a comment
colleenish
left a comment
There was a problem hiding this comment.
First pass edits. Let me know if you have any questions.
Edits from <rdar://162699805> Co-authored-by: Colleen Toporek <colleen_toporek@apple.com>
colleenish
left a comment
colleenish
left a comment
There was a problem hiding this comment.
Looks good; no additional changes from me.
amartini51
force-pushed
the
numbers_122436757
branch
from
aa04a2b to
0f4377d
Compare
4 participants
Summary of changes:
Fixes: rdar://122436757