Starting HP25 programs with numbers

[Dieter]

(11-24-2017 09:25 PM)teenix Wrote:  some 2^3 results
Classics/HP25
X = 08000000002000 (8.000000002)
HP65/HP67/HP22/HP27/HP25c/HP34c/HP19c
08000000000000 (8.00)
HP21
X=07999997770000 (7.99999777)

As already noted, the 25 and 25C should return the same result 8,000000002, as does the HP65. Probably all other pre-1976 calculators are off here. The HP67, 34C and 19/29C use 13-digit precision and thus correctly return a plain 8.

BTW I am surprised at the HP21 result. What's going on there?

Dieter

[Dieter]

(11-24-2017 11:18 PM)PANAMATIK Wrote:  I already integrated your wonderful patch for the HP-29C PAUSE interruption, and I think I should also use this patch for the HP-25 emulation, if you allow.

Please don't. One of the advantages of your emulators is their (almost) exact representation of the original device. If now the one or other "bugfix" or "improvement" is applied they no longer match the original calculator. Without your emulator I would never have found out the "special behaviour" of the HP25 which lead to this thread.

If you really want to do it, please set up a separate modified version that includes such fixes. And while we're at it: what about a "speed mode" and/or a "double memory mode" then? I recently wrote a ~400 step program for the HP67 that has to be run on another emulator. If you don't like the idea: this may have to get activated by a "secret" key sequence. ;-)

Dieter

[PANAMATIK]

(11-25-2017 01:10 PM)Dieter Wrote:  Please don't. One of the advantages of your emulators is their (almost) exact representation of the original device....

Dieter

You are right. Don't worry, the PC emulators will remain original. I was thinking about the LP circuit and "new ACT".

Bernhard

[rprosperi]

(11-25-2017 02:21 PM)PANAMATIK Wrote:  You are right. Don't worry, the PC emulators will remain original. I was thinking about the LP circuit and "new ACT".

I agree with Dieter and feel that even the WLP machines should exactly match original machine behavior, even if they have bugs (or at least, unexpected behavior). If there is some easy way for you to provide alternate versions (without having to include another nearly identical set of f/w) that could be desirable, but IMHO not justified.

I believe running any existing programs (e.g. from the Applications books, PPC Journals, etc.) just as the original machines did is more important than correcting the f/w for the odd case such as that which prompted this thread (which btw, thanks to Dieter for noticing and documenting).

Other folks may have a different view; if so, please comment here, as I may be in the minority about this.

[PANAMATIK]

(11-25-2017 02:34 PM)rprosperi Wrote:  I agree with Dieter and feel that even the WLP machines should exactly match original machine behavior ...

OK, I think I will not change anything. Only the HP-29C firmware was changed, to allow interrupting a program during PAUSE with a short key press. I agree that relying on the original behavior is a valuable feature of an emulation.

Although the "Woodstock Low Power" will display program steps as alphanumeric text, and has many new key sequences and new functions, most of them can be disabled and it is easy to see the original behavior.

Bernhard