Re: [Maxima-discuss] float(exp(10000)) works, but float(exp(10000/3)) doesn't

[Stavros M. (Σ. Μ. <mac...@al...>]

Is auto-promotion of floats to bfloats a good idea?

Are there practical applications, where a user will get a useful answer
rather than an error to a substantive problem
​? I doubt it; I suspect this only happens with

​​
artificial problem
​s​
designed to test Maxima's limit
​s.)​


Are there cases where it makes it easier to understand what Maxima is
doing? (I would say the opposite.)

Are there cases which give a less accurate or less useful result? (Yes, I
showed some.)

If it is useful, shouldn't underflow be handled the same way?
​(Otherwise
float(cot(10^-400)
​) overflows.) ​
​
Should bfloats be used instead of IEEE gradual underflow?

Does it work the same way for compiled code with float declarations
(presumably not)? To fix that, should float declarations be taken into
account in interpreted code?

ee(ff):=(modedeclare(ff,float),exp(ff))$
ee(1000.0) => 1.970071114017047b434
translate(ee)$
ee(1000.0d0) => arithmetic error FLOATING-POINT-OVERFLOW signalled


​Some additional banal fail-to-promote bugs:

1000.0!
expintegral_ei(1000.0)
product(1000.0d0,i,1,200)

product(1000.0,i,1,40)    <<< single float

sum(1e308,i,1,2)


​PS Shouldn't all Maxima floats be double-floats and print with "e"
exponents? These single floats are annoying. Is this a bug in the SBCL
implementation?
​


On Wed, Jan 27, 2016 at 2:45 PM, Raymond Toy <toy...@gm...> wrote:

> >>>>> "Jaime" == Jaime Villate <vi...@fe...> writes:
>
>     Jaime> On 27-01-2016 16:21, Raymond Toy wrote:
>     >> This is why I was opposed to this change. It is no longer possible
> to
>     >> reason on what a program will do without knowing all the possible
>     >> values.
>     >>
>     >> I lost that battle.
>     Jaime> Let's restart the battle then in view of this new thread :)
>
> I'm not bitter or anything about it; I'm sure I'll lose many others
> and maybe win some too.
>
> I just didn't like the change, but many others did. I bowed to the
> will of the people. :-)  And the will of the users does matter.
>
> --
> Ray
>
>
>
> ------------------------------------------------------------------------------
> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> Monitor end-to-end web transactions and take corrective actions now
> Troubleshoot faster and improve end-user experience. Signup Now!
> http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
> _______________________________________________
> Maxima-discuss mailing list
> Max...@li...
> https://lists.sourceforge.net/lists/listinfo/maxima-discuss
>