WP34s: number display and rounding

[Dieter]

In the olden days I also used a TI58 here and there. As on most TIs, its standard display format shows all digits, similar to the original HP35 or the later ALL setting, while most HPs used a fixed decimal format - FIX 2 on earlier machines and FIX 4 on later models.

But there was another difference. HPs machines, set to a FIX format, usually switch to scientific notation if the value is so small that if would round to zero. Enter Pi, and in FIX 4 you will get 3,1416. Divide this by 10000 and you'll see 0,0003. Divide again by 10 and you will get 3,1416 E-5. While TI's machines showed the same results in the first two cases, the last value was handled differently: it was simply displayed as zero (0,0000). In other words: the HP way was a switch to scientific notation while the TIs showed the result like the one you would have got after an additional RND (which returns zero in this case as well).

I think both ways of handling this are valid options. There are cases where I actually prefer the "round to zero" way. So I would like to ask two questions:

1. Do you consider an option or a flag useful that sets either the one or the other display rounding mode?

2. Especially for Pauli: if (!) this option is considered useful, is there a chance this may find its way into the 34s?

Dieter

[Paul Dale]

I don't personally care for this. If you are in FIX mode and the display switches, the value would have displayed as zero. Sounds simple enough to me at least. We will still need the switch to exponential notation for large numbers -- I assume TI's do this even though it isn't entirely consistent to my mind.

Anyway, if enough people think it is worthwhile, we could dedicate the fabled last bit of RAM to supporting this mode. I'd have to see how much flash it consumes -- the display formatting code is horrible and making changes is more difficult & more space consuming than I'd like :-( Of course, we'd have to no think of a better use of the bit :-)

As mentioned, executing RND after the underflow to exponential notation does result in the TI answer. Is this a bug or a feature I wonder?

- Pauli

[Steve Simpkin]

I am very much against this mode. I always want to know the difference between zero and 3.1416 E-5.

I am hoping that last bit of RAM will be available for a USER mode :)

[M. Joury]

I *strongly* second this. Don't see much need for a round to zero automatically (I actually think that this is a bad idea) and am very much in favor of a user mode. Given one bit left I would *strongly* opt for a user mode if at all possible.

Just my 0.02

Cheers,

-Marwan

[Marcus von Cube, Germany]

Before we discuss any changes I need to squeeze the code a little bit more. The last version fills flash exactly. :-/