bug#27765: 26.0.50; Calc b% and relch are inconsistent in algebraic and rpn modes

[Lars Ingebrigtsen <larsi@gnus.org>]

eythanweg@gmail.com (Eythan Weg) writes:

>  >> Agreed.  However, the displays are different. Won’t expect b% and relch
>  >> behaves exactly the same?
>
>  >That's called convenience, probably.
>
> A little more than that.  E.g. the org-mode speadsheet uses calc.
> Summing cells which display percentages like 10% and 90%, one expects
> the result to be 100% rather than 1.  Here the uniform display of
> percentages is significant, I would say.
>
> Curiosity lead me to check the same operation on libreoffice and gnumeric.
> They both adhere to the expected.

As far as I can tell, relch isn't documented to display the percentage
in one format or another (i.e., "50%" or "0.5"), while `b %' is
documented to do "50%".

I don't use calc much, but I think the current behaviour makes sense --
a function like relch might be supposed to use the format "0.5" since
that's what more math-ey types are used to seeing relative change
represented as.

But perhaps others have other opinions; I've added Mattias to the CCs --
perhaps he has some comments.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no