Login

Thomas Radtke · 07-15-2009, 09:04 AM

It appears, the 20b acts partially like a RPL calculator: 1 ENTER 2 ENTER + results in 3 instead of 4. OTOH, a percentage operation a b % leaves a on the stack, as opposed to RPL calculators like the 48 series. Quite confusing, but I like it.

An interesting bug: Some error conditions apparently leave an invalid number on the stack. Try e.g.:

CE

TVM reset

N I/YR PV (ignore errors)

Now cycle through the stack, some entries are missing. These are maybe invalid numbers.

Gene Wright · 07-15-2009, 10:50 AM

Hi Thomas!

1) Yes, the 20b uses a "command line" approach as an input to the stack. Your addition example points out the effect quite well. Another result can be that if you have a stack full of 4 values and begin typing another number, if you press ON/CE, you still have all 4 values. On a "true RPN" machine, you would have lost the value in T. On the 20b, T is lost only if you do something with the value you are keying it. ON/CE is an "abort" button.

2) The % operation works as "RPN" does.

So, a bit of a hybrid, probably because it reuses the RPL approach to inputs on the stack.

3) If you're referring to an apparently blank display, then yes, this is a non-normalized number aka error result. I know HP is aware of this. Of course, a blank display should never be returned to the user, so this is a bug.

What ROM revision is on your 20b?

Thomas Radtke · 07-15-2009, 11:28 AM

Hi Gene,

Quote:
What ROM revision is on your 20b?

How to find it out?

Tim Wessman · 07-15-2009, 11:40 AM

On-PMT opens the hidden menu. Scroll up and down and you will see it in there.

Also, can you give the example that returns the blank screen so I can check it is fixed? Thanks.

TW

Thomas Radtke · 07-15-2009, 11:56 AM

Quote:
On-PMT opens the hidden menu. Scroll up and down and you will see it in there.

SW Version shows 8 01 2008

Quote:
Also, can you give the example that returns the blank screen so I can check it is fixed? Thanks.

As a more simple approach to get one blank entry, press

CE

N

I/YR

Arrow down (4 times)

Walter B · 07-15-2009, 03:44 PM

FWIW, SW Version 6 17 2008 shows the same bug.

Gene Wright · 07-15-2009, 03:53 PM

Of course, since that is an earlier firmware, it means it had not been fixed by August of last year. :-)

Hopefully, unlike some software companies, updated HP firmware does not break things that used to work!

Walter B · 07-15-2009, 04:07 PM

Quote:
Hopefully, unlike some software companies, updated HP firmware does not break things that used to work!

Hopefully, but not for sure ;)

Quote:
Of course, since that is an earlier firmware, it means it had not been fixed by August of last year. :-)

Oh, that was M DD YYYY :( Sorry I didn't recognize that.

Gene Wright · 07-15-2009, 05:56 PM

Oops. Tough to know when the digits don't indicate where the day is located, isn't it?

cyrille de Brébisson · 07-16-2009, 08:37 AM

hello,

it's not an 'invalid number', which would be a serious bug, it's a perfectly normalized NaN number...

the problem is that the UI does not realy know how to handle NaNs...

I am trying to remove all occurrences of such problems.

regards, cyrille

Thomas Radtke · 07-16-2009, 09:06 AM

Great, allthough this bug doesn't seem to affect normal operation. May I be so cheeky to ask whether someone is debugging the 35s? :^)

Marcus von Cube, Germany · 07-17-2009, 02:40 AM

I guess, the 20b software is a one man show and we are lucky he is here on the forum. :)

bill platt · 07-17-2009, 01:34 PM

Yes, that is true.

And of course never overestimate how man people are behind the curtain! (It is always amazing to realize how much one motivated and talented person can do).

I lurk most of the time on these issues but find it interesting to follow them.

Login
Username:
Password:	Lost Password?
	Remember me